Date: Sun, 1 Jul 2001 02:03:01 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: [lang] [BF] Emacs mode Hi, I worked up a little emacs mode for brainfuck including an interpreter and odd debugging functionality. Please look at my comments for more details, though. They comprise more than a third of the file, after all :-) Markus -- Attached file included as plaintext by Listar -- -- File: bf.el ;;; bf.el --- Major mode for editing, running and debugging brainfuck ;; Copyright (C) 2001 Markus Kliegl ;; Author: Markus Kliegl ;; Created: 24 Jun 2001 ;; Version: 0.1 ;; Keywords: bf brainfuck ;; This software is free software; you can redistribute it and/or modify ;; it under the terms of the GNU General Public License as published by ;; the Free Software Foundation; either version 2, or (at your option) ;; any later version. ;; This software is distributed in the hope that it will be useful, ;; but WITHOUT ANY WARRANTY; without even the implied warranty of ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ;; GNU General Public License for more details. ;; You should have received a copy of the GNU General Public License ;; along with this software; see the file COPYING. If not, write to the ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330, ;; Boston, MA 02111-1307, USA. ;;; Commentary: ;;; Using ;; Editing commands: ;; C-c % -- jumps to the matching bracket. ;; Running and debugging commands: ;; C-c C-s p -- sets the value of the pointer ;; C-c C-s a -- set the values of the elements of the array ;; (val-of-elt-0 val-of-elt-1... val-of-elt-n) ;; (val-of-elt-0 (10 val-of-elt-10) val-of-elt-11... val-of-elt-n) ;; ex: (5 (10 12) 6 (4 2) 4) will set element 0 to 5, element 10 to ;; 12, element 11 to 6, element 4 to 2 and element 5 to 4. ;; C-c C-p -- Display the value of the pointer ;; C-c C-a -- Display the value of the nth element of the array ;; (defaults to a[p]) ;; C-c C-b a -- Add a breakpoint at current cursor position ;; C-c C-b d -- Remove breakpoint at current cursor position ;; C-c C-b r -- Remove all breakpoints ;; C-c C-t -- Reset pointer and all array elements to 0. ;; Reset current position (in case of being after a breakpoint) ;; C-c C-r -- Interpret the buffer ;; C-c C-c -- Continue interpreting after a breakpoint ;; The Interpreter ;; It should work similar to most other interpreters, though ;; at the moment it is very slow (be sure to byte-compile this file ;; or it will be *way* too slow!) ;; When reading input, pressing ESC will put 0 into a[p], similar ;; to the EOF->0 of other interpreters. ;; I tested the interpreter with a hello world and cat ( +[,.] ) ;; program and they seem to be working. 99 bottles of beer, however, ;; doesn't seem to be working. Oddly enough, using the debugger, it ;; appears that it manages to print out the text of the first 9 bottles ;; (i.e. down to 90 bottles...) and then runs into an infinite loop with ;; weird negative numbers in the array. The program runs fine on Panu's ;; bfi, though. ;;; Installing ;; Emacs must be able to find this file, so there are two options: ;; (1) Copy this file to your local site-lisp directory. This will ;; usually be something like /usr/local/share/emacs/site-lisp ;; To see a list of all directories that emacs searches on ;; startup, start emacs and in the *scratch* buffer type ;; 'load-path' (without the quotes, obviously) and press C-j. ;; (2) Add the following to your ~/.emacs file (where "/some/dir" ;; is the directory which contains this file): ;; (setq load-path (cons "/some/dir" load-path)) ;; In order for emacs to automatically put you in brainfuck-mode when ;; editing a file, whose name ends in either .b or .bf, add the following ;; to your ~/.emacs: ;; (autoload 'bf-mode "bf" "Brainfuck major mode" t) ;; (setq auto-mode-alist ;; (append '(("\\.b$" . bf-mode) ;; ("\\.bf$" . bf-mode)) ;; auto-mode-alist)) ;; For efficiency reasons it is recommended to run the following at your ;; shell in order to obtain a byte-compiled version, bf.elc, of this file. ;; $ emacs -batch -f batch-byte-compile bf.el ;; In particular the interpreter and startup are considerably faster. ;;; Customizing ;; I don't quite get the differences between the following two methods, in ;; particular I'm having problems understanding all this 'customization' ;; stuff and only included it in this version to get some hints on how to ;; use it correctly. ;; You can customize several things. There should be two ways to do this: ;; (1) type M-x customize-group RET bf RET. There you should be able to ;; change the values of the customizable variables. ;; (2) Set the bf-mode-hook in your ~/.emacs. For example, you could use ;; something like the following: ;; (setq bf-mode-hook ;; '(lambda () ;; (setq bf-arr-size 1000))) ;;; Todo: ;; * optional running of external bf interpreter/compiler ;; * editing commands ;; + generate-string ;; + set-current-elt-to-val ;; + copy n times to left/right ;; + ... ;; * fix the customization stuff ;; * clean up matching-bracket (especially to reuse common code) ;; * speed up interpreter ;; * don't print ^@ when ESC is entered, etc. ;; * more debugging capabilities (stepping through the program, etc.) ;;; Change Log: ;; 0.1 ;; - initial version ;;; Bugs: ;; * the temporary buffer doesn't display until the program ;; finished running or is otherwise stopped and is apparently ;; reopened when continuing after a break ;; * why doesn't 99 bottles of beer work? ;;; Code: (eval-when-compile (require 'cl) (require 'custom)) ;;; Customization (defgroup brainfuck nil "Major mode for editing, running and debugging brainfuck." :group 'languages) (defcustom bf-arr-size 30000 "*Size of the array." :type 'int :group 'bf) (defcustom bf-mode-hook nil "*Hook that is run on startup of brainfuck mode" :type 'hook :group 'bf) ;;; Editing (defun matching-bracket-open () (interactive) (let ((cur (point)) (end (point-max)) (depth 1) (not-found t)) (while (and (< cur end) not-found) (incf cur) (case (char-after cur) (?\[ (incf depth)) (?\] (decf depth) (when (= depth 0) (setq not-found nil))))) (goto-char cur))) (defun matching-bracket-close () (interactive) (let ((cur (point)) (end (point-min)) (depth 1) (not-found t)) (while (and (> cur end) not-found) (decf cur) (case (char-after cur) (?\] (incf depth)) (?\[ (decf depth) (when (= depth 0) (setq not-found nil))))) (goto-char cur))) (defun matching-bracket () "Jump to matching bracket." (interactive) (case (char-after (point)) (?\[ (matching-bracket-open)) (?\] (matching-bracket-close)) (t (message "Not on a bracket!")))) ;;; Running and debugging (defvar bf-ptr 0 "Pointer") (defvar bf-arr (make-vector bf-arr-size 0) "Array") (defvar bf-break-points nil "Breakpoints for debugging purposes") (defvar bf-cur-pos (point-min) "Current position") (defun set-ptr (n) "Set the initial value of the pointer." (interactive "NInitial value of pointer: ") (setq bf-ptr n) (message "Pointer set!")) (defun set-arr-aux (lst n) (let ((x (car lst))) (cond ((consp x) (let ((idx (car x))) (aset bf-arr idx (cadr x)) (set-arr-aux (cdr lst) (1+ idx)))) ((null lst) nil) (t (aset bf-arr n x) (set-arr-aux (cdr lst) (1+ n)))))) (defun set-arr (lst) "Set the initial values of the elements of the array." (interactive "xArray initialization: ") (set-arr-aux lst 0) (message "Array set!")) (defun val-of-ptr () "Display the value of the pointer." (interactive) (message "%d" bf-ptr)) (defun val-of-arr-elt (&optional n) "Display the element of the array denoted by the numeric prefix argument. Default value is the pointer." (interactive "P") (let ((x (if (not n) bf-ptr (prefix-numeric-value n)))) (message "a[%d]: %d" x (aref bf-arr x)))) (defun reset-for-new-run () "Reset the pointer, array and current position for a new run." (interactive) (setq bf-arr (map 'vector #'(lambda (foo) 0) bf-arr)) (setq bf-ptr 0) (setq bf-cur-pos (point-min)) (message "Reset!")) (defun add-break-point () "Add a break point." (interactive) (push (point) bf-break-points) (message "Breakpoint added at %d" (point))) (defun remove-break-point () "Remove a break point." (interactive) (delete (point) bf-break-points) (message "Breakpoint at %d removed" (point))) (defun remove-all-break-points () "Remove all break points." (interactive) (setq bf-break-points ()) (message "All breakpoints removed!")) (defun bf-interpret () (let ((end (point-max))) (while (< (point) end) (when (member (point) bf-break-points) (message "Breakpoint at %d" (point)) (setq end -1)) (case (char-after (point)) (?+ (incf (aref bf-arr bf-ptr))) (?- (decf (aref bf-arr bf-ptr))) (?> (incf bf-ptr)) (?< (decf bf-ptr)) (?. (write-char (aref bf-arr bf-ptr))) (?, (aset bf-arr bf-ptr (let ((c (read-char))) (if (= c ?\e) 0 c)))) (?\[ (when (= (aref bf-arr bf-ptr) 0) (matching-bracket-open))) (?\] (unless (= (aref bf-arr bf-ptr) 0) (matching-bracket-close)))) (forward-char 1)))) (defun run-bf (&optional cont) "Interpret the current buffer." (interactive) (message "Running...") (save-excursion (unless cont (setq bf-cur-pos 0)) (goto-char bf-cur-pos) (with-output-to-temp-buffer "*bf-run*" (bf-interpret)) (setq bf-cur-pos (point)))) (defun bf-continue () "Continue running after stopping at a breakpoint." (interactive) (message "Continuing at %d..." bf-cur-pos) (run-bf t)) ;;; General (defconst bf-mode-version "0.1") (defvar bf-mode-map nil) (unless bf-mode-map (setq bf-mode-map (make-sparse-keymap)) (define-key bf-mode-map "\C-c%" 'matching-bracket) (define-key bf-mode-map "\C-c\C-p" 'val-of-ptr) (define-key bf-mode-map "\C-c\C-a" 'val-of-arr-elt) (define-key bf-mode-map "\C-c\C-sp" 'set-ptr) (define-key bf-mode-map "\C-c\C-sa" 'set-arr) (define-key bf-mode-map "\C-c\C-ba" 'add-break-point) (define-key bf-mode-map "\C-c\C-bd" 'remove-break-point) (define-key bf-mode-map "\C-C\C-br" 'remove-all-break-points) (define-key bf-mode-map "\C-c\C-t" 'reset-for-new-run) (define-key bf-mode-map "\C-c\C-r" 'run-bf) (define-key bf-mode-map "\C-c\C-c" 'bf-continue)) (defvar bf-mode-syntax-table nil) (unless bf-mode-syntax-table (setq bf-mode-syntax-table (make-syntax-table)) (modify-syntax-entry ?+ "." bf-mode-syntax-table) (modify-syntax-entry ?- "." bf-mode-syntax-table) (modify-syntax-entry ?< "." bf-mode-syntax-table) (modify-syntax-entry ?> "." bf-mode-syntax-table) (modify-syntax-entry ?. "." bf-mode-syntax-table) (modify-syntax-entry ?, "." bf-mode-syntax-table) (modify-syntax-entry ?\[ "(]" bf-mode-syntax-table) (modify-syntax-entry ?\] ")[" bf-mode-syntax-table)) (defvar bf-mode-abbrev-table nil) (define-abbrev-table 'bf-mode-abbrev-table ()) (defun bf-mode () "Major mode for editing and running brainfuck. Please see the comments at the beginning of the file bf.el for instructions on how to setup bf-mode and general information and documentation on what bf-mode provides and how to use it. Key bindings: \\{bf-mode-map}" (interactive) (message "Brainfuck mode v%s" bf-mode-version) (kill-all-local-variables) (use-local-map bf-mode-map) (set-syntax-table bf-mode-syntax-table) (setq local-abbrev-table bf-mode-abbrev-table) (setq major-mode 'bf-mode) (setq mode-name "Brainfuck") (run-hooks 'bf-mode-hook)) ;;; bf.el ends here ------------------------------ Date: Sun, 1 Jul 2001 02:55:01 +0200 (CEST) From: "Rafal M. Sulejman" Subject: [lang] Re: [BF] Emacs mode On Sun, 1 Jul 2001, markus.kliegl wrote: > I worked up a little emacs mode for brainfuck including an > interpreter and odd debugging functionality. Heh, someone to implement this in vi? emacsers just won a battle in this world domination war ;) (Use of BF against civil population was forbidden by UN in 1999) > Please look at my comments for more details, though. They > comprise more than a third of the file, after all :-) FSF will probably send you four six-packs of Virtual Beer :) Now I have *real* reason to learn eeee-macs. BTW. -- where do you live, Markus? Ich bin ein M'jlabbacher (not exactly ;) -- . . . . | Rafal M. Sulejman . # . . | # . # . | "It's a small world, but I wouldn't want . . # . | to have to paint it." --Steven Wright ------------------------------ Date: Sun, 01 Jul 2001 17:27:53 -0500 From: Chris Pressey Subject: pretentious rant about operating systems I'll get right to the point. WHERE'S THE METADATA? The answer: when it is there, it is spread across umpteen configuration files that don't know about each other, OR, it is lumped willy-nilly into an obscure, convoluted centralized database... but it's usually not there. The fact is that system objects (files, devices, etc) have relationships with each other. Relationships are nothing new in computer science, we've been using relational databases for years. So why hasn't the rather simple concept of "Object-A somehow-relates-to Object-B" filtered down to the operating system level yet? We have dependencies in 'make' scripts, and we have 'install' and 'uninstall' scripts, and we have 'file type detection' in the form of either file suffix or 'magic'... but these are all loose and scattered functions. In the meantime, you can rename a file and none of these measures are the wiser. (And when an OS does have enough seatbelts to warn you, the warning is seemingly never specific enough: "file may be required by some installed packages" - well I'd love to know which ones?!?) There are some other little things that are just shell/filesystem issues. I'd love for a shell to be able to change to a directory just by typing its name. It should be obvious that "enter" is the verb to be implied when you just name a directory. I'd also love for the FS to be able to treat archive files (.tar, .tgz, .zip, .bz2 etc) as directories, so you could "enter" them and copy files in and out with the usual commands (not to mention having programs like "find" able to search within archives automatically :) Chris -- This sentence was false half an hour ago, did you notice? ------------------------------ Date: Mon, 2 Jul 2001 01:54:30 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Re: pretentious rant about operating systems On Sun, 1 Jul 2001, Chris Pressey wrote: > > There are some other little things that are just shell/filesystem > issues. I'd love for a shell to be able to change to a directory just > by typing its name. It should be obvious that "enter" is the verb to be > implied when you just name a directory. I'm quite sure that would be simple enough to change... most shells these days hash executables in the PATH, so it should be easy enough to detect if there's an executable of the same name as the directory and then ask the user. Interesting, though... having to type 'cd' to change directories never bothered me. 'some_dir' just doesn't seem right to me for changing to directory some_dir. Why not have ENTER mean 'ls ' instead? Or something else if you get what I mean :-) > I'd also love for the FS to be > able to treat archive files (.tar, .tgz, .zip, .bz2 etc) as directories, > so you could "enter" them and copy files in and out with the usual > commands (not to mention having programs like "find" able to search > within archives automatically :) There was an interesting article on freshmeat a while ago... http://freshmeat.net/articles/view/237/ > > Chris > > -- > This sentence was false half an hour ago, did you notice? > ------------------------------ From: Gerson.Kurz@t-online.de (Gerson Kurz) Subject: a linux rant (was: pretentious rant about operating systems) Date: Mon, 2 Jul 2001 20:48:41 +0100 Chris, Metadata: you have problems! Here is what happened the weekend before last: I am dissatisfied with Windows again and decide its once again try working my way with Linux. I buy the brand new Suse Linux 7.2 Pro at 130,- DM. "Free", really. I go home and boot the DVD -> the boot process stops and shows a blank screen. For the 130,- DM I have the right to phone a support line; I stay for nearly one hour on the phone with support until they give up: my machine is not supported, write down all hardware & send a letter. [Its a brandnew Dual 1GHz P3 with 512m]. I try changing various BIOS settings until I find a combination that lets me install Linux. I am greeted with a resolution of 800x600. (I have a GeForce 2MX, plus a 21" monitor capable of 2048x1536, but I am willing to do with 1600x1200). I install XFree 4.1.0. When I switch to 1600x1200, the mouse disappears. But this is not a problem, since SAX2 has lost its capability of changing the resolution anyway: I can change it, but it has no effect on KDE. I reinstall Linux. SAX2 works, but my Monitor is not on the monitor list (its an USB monitor on a OS with USB support, btw) & I have to fiddle with settings. I have to edit vertical & horizontal freqs; If I enter the specs given by my monitors vendor, the screen gets blank; I thus enter the specs of a different 21" monitor I use @ work, finally I see something somewhat resembling 1600x1200, but with a total different screen dimension than W2K on the same machine. I'm probably ruining my monitor as I speak. I enable font antialiasing. The font for Konsole changes to something unreadable and unrenderable. I quickly decide to disable font antialising again. I start Mozilla. It fails on my homepage. It shows a clearly faulty error message for out-of-memory, on a 512mB machine. I start Konqueror. It has trouble with my homepage. I cannot change the fonts in Konqueror, even though the default fonts designed for 800x600 look really really really really tiny on a 1600x1200 resolution. I edit some python scripts in NEdit. NEdit uses unreadble oldskool X-Widgets (I never figured out if that checkbutton is now pressed or depressed). NEdits Ctrl+S doesn't work 50% of the time. NEdit mixes \t and spaces so the source gets unreadble for python. I use Komodo instead. Initially, I am pleased to see something somewhat resembling the good Pythonwin. I am used to scroll mainly with the mousewheel. I like that I can use the mousewheel in Komodo, until I find out that the mousewheel is also associated with "Paste", so when I scroll the sourcecode gets sprinkled with pasted bits. ***AAAARG***! And the UNDO doesn't seem to care. I want to import my IE bookmarks to konqueror. I wrote an import script some months ago, and after a few editings (see above) the bookmark files are created; then I notice that in the meanwhile they have changed the bookmarks to a goddamn XML-file. (Did I tell you that I hate XML, and with a reason ? XML is worth a rant in its own right, so maybe nexttime) I change the script to create an XML file. Konqueror doesn't like it because it contains Windows characters (German Umlauts et al). I ask Yast to install the gtk-bindings for Python. The gtk-samples don't start. I ask Yast to install the Qt-Bindings for Python. I didn't find any samples and/or docs. RealPlayer keeps hanging up. I am a vivid real-listener, for I am addicted to www.klubradio.de. Needless to say, there are no DFX plugins for Linux MP3 players. Knowing ext2, I had choosen ReiserFS during install. It is impossible to access the ReiserFS-parition from windows. Hans Reiser claims this is because the M$ IFS kit is too expensive and M$ is a bad monopoly anyway. Even though he most likely is right on the latter part, I once bought "Windows NT File System Internals" (http://www.oreilly.com/catalog/wininternals/index.html) for $49.95, and I can read raw NTFS ondisk structures; but then again, I am no Hans Reiser. I have a Dual P3 1Ghz, but somehow copying from large files from NTFS to my ~ brings down the whole multitasking SMP system speed to the point that the mouse cursor freezes until the file has been copied. I attempt to enable UDMA support, but am greeted with horrendous warnings about possible bugs in IDE chipsets. (Windows doesn't seem to mind, though). I try working as a normal user. (Yes, I logged on as root before). I cannot start mozilla. "shutdown" is an unknown command. When I "su", I cannot start X apps. I have to re-edit .bashrc to get some aliases i can remember. I logoff and logon as root again. Logged on as root, my DSL connection keeps hanging up - it only keeps alive if I am logged on as non-root-user. In psychology, you call that double-bind. I start reading the documentation for "Suse personal firewall". Being used to AtGuard on Windows, I quickly stop doing so. I have a *really* quick look at IPChains before I reboot to Windows. Somehow, all that reminds me *terribly* of OS/2. And you all know what I think of OS/2. OT: I really miss Amiga-Style ASSIGN:s ------------------------------ Date: Mon, 02 Jul 2001 12:29:11 -0700 From: Robert Subject: Re: a linux rant (was: pretentious rant about operating systems) None of you probly know me, I subscribed to the mailing list right when the number-base thread started, then never bothered to subscribe to the two new lists until now hehehe. For some reason I don't seem to run in to any of these problems with linux. But I've been using slackware since 1996..if you have the bandwidth it's easier in the long run just to download the iso and burn it to a cd. Can't run X programs when you su? Well export DISPLAY=":0" if using bash. USB support means it supports the usb chipset, and we won't go in to how messed up via's usb chipset is..*cough* FreeBSD has better usb support..not to mention overall is just better for the most part. With regards to xml, I'd rather have the standard document format as xml over m$ word format. ;) I guess some people are just better off with windows :) Windows once you install win2k is bearable to some degree. I'm using WinXP Beta-2 going to upgrade to rc1 tonight hehe. ----- Original Message ----- From: "Gerson Kurz" To: "ESOLANG" Sent: Monday, July 02, 2001 12:48 PM Subject: a linux rant (was: pretentious rant about operating systems) > > Chris, Metadata: you have problems! Here is what happened the weekend before > last: > > I am dissatisfied with Windows again and decide its once again try working > my way with Linux. I buy the brand new Suse Linux 7.2 Pro at 130,- DM. > "Free", really. I go home and boot the DVD -> the boot process stops and > shows a blank screen. For the 130,- DM I have the right to phone a support > line; I stay for nearly one hour on the phone with support until they give > up: my machine is not supported, write down all hardware & send a letter. > [Its a brandnew Dual 1GHz P3 with 512m]. > > I try changing various BIOS settings until I find a combination that lets me > install Linux. I am greeted with a resolution of 800x600. (I have a GeForce > 2MX, plus a 21" monitor capable of 2048x1536, but I am willing to do with > 1600x1200). I install XFree 4.1.0. When I switch to 1600x1200, the mouse > disappears. But this is not a problem, since SAX2 has lost its capability of > changing the resolution anyway: I can change it, but it has no effect on > KDE. I reinstall Linux. SAX2 works, but my Monitor is not on the monitor > list (its an USB monitor on a OS with USB support, btw) & I have to fiddle > with settings. I have to edit vertical & horizontal freqs; If I enter the > specs given by my monitors vendor, the screen gets blank; I thus enter the > specs of a different 21" monitor I use @ work, finally I see something > somewhat resembling 1600x1200, but with a total different screen dimension > than W2K on the same machine. I'm probably ruining my monitor as I speak. > > I enable font antialiasing. The font for Konsole changes to something > unreadable and unrenderable. I quickly decide to disable font antialising > again. > > I start Mozilla. It fails on my homepage. It shows a clearly faulty error > message for out-of-memory, on a 512mB machine. I start Konqueror. It has > trouble with my homepage. I cannot change the fonts in Konqueror, even > though the default fonts designed for 800x600 look really really really > really tiny on a 1600x1200 resolution. > > I edit some python scripts in NEdit. NEdit uses unreadble oldskool X-Widgets > (I never figured out if that checkbutton is now pressed or depressed). > NEdits Ctrl+S doesn't work 50% of the time. NEdit mixes \t and spaces so the > source gets unreadble for python. > > I use Komodo instead. Initially, I am pleased to see something somewhat > resembling the good Pythonwin. I am used to scroll mainly with the > mousewheel. I like that I can use the mousewheel in Komodo, until I find out > that the mousewheel is also associated with "Paste", so when I scroll the > sourcecode gets sprinkled with pasted bits. ***AAAARG***! And the UNDO > doesn't seem to care. > > I want to import my IE bookmarks to konqueror. I wrote an import script some > months ago, and after a few editings (see above) the bookmark files are > created; then I notice that in the meanwhile they have changed the bookmarks > to a goddamn XML-file. (Did I tell you that I hate XML, and with a reason ? > XML is worth a rant in its own right, so maybe nexttime) > > I change the script to create an XML file. Konqueror doesn't like it because > it contains Windows characters (German Umlauts et al). > > I ask Yast to install the gtk-bindings for Python. The gtk-samples don't > start. I ask Yast to install the Qt-Bindings for Python. I didn't find any > samples and/or docs. > > RealPlayer keeps hanging up. I am a vivid real-listener, for I am addicted > to www.klubradio.de. Needless to say, there are no DFX plugins for Linux MP3 > players. > > Knowing ext2, I had choosen ReiserFS during install. It is impossible to > access the ReiserFS-parition from windows. Hans Reiser claims this is > because the M$ IFS kit is too expensive and M$ is a bad monopoly anyway. > Even though he most likely is right on the latter part, I once bought > "Windows NT File System Internals" > (http://www.oreilly.com/catalog/wininternals/index.html) for $49.95, and I > can read raw NTFS ondisk structures; but then again, I am no Hans Reiser. > > I have a Dual P3 1Ghz, but somehow copying from large files from NTFS to my > ~ brings down the whole multitasking SMP system speed to the point that the > mouse cursor freezes until the file has been copied. I attempt to enable > UDMA support, but am greeted with horrendous warnings about possible bugs in > IDE chipsets. (Windows doesn't seem to mind, though). > > I try working as a normal user. (Yes, I logged on as root before). I cannot > start mozilla. "shutdown" is an unknown command. When I "su", I cannot start > X apps. I have to re-edit .bashrc to get some aliases i can remember. I > logoff and logon as root again. Logged on as root, my DSL connection keeps > hanging up - it only keeps alive if I am logged on as non-root-user. In > psychology, you call that double-bind. > > I start reading the documentation for "Suse personal firewall". Being used > to AtGuard on Windows, I quickly stop doing so. I have a *really* quick look > at IPChains before I reboot to Windows. > > Somehow, all that reminds me *terribly* of OS/2. And you all know what I > think of OS/2. > > OT: I really miss Amiga-Style ASSIGN:s > > > > > ------------------------------ Date: Mon, 02 Jul 2001 22:47:19 +0100 From: Keith Gaughan Subject: Bloody Listar! Ok, every time I've posted stuff to the list, it's gone back to the sender. Why? Listar doesn't seem to include something as simple as a Reply-To: in the email header when it forwards on mail. Is this just happening to me? Could it be Eudora that's causing the problems? Panu, if it really is Listar, could you please fix this? K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Mon, 02 Jul 2001 22:49:03 +0100 From: Keith Gaughan Subject: Fwd: Re: pretentious rant about operating systems He're one of those emails which *didn't* get sent to the correct address: >Date: Mon, 02 Jul 2001 21:37:43 +0100 >To: Chris Pressey >From: Keith Gaughan >Subject: Re: pretentious rant about operating systems > >At 17:27 01/07/2001 -0500, Chris Pressey wrote: > >>I'll get right to the point. WHERE'S THE METADATA? >> >>The answer: when it is there, it is spread across umpteen configuration >>files that don't know about each other, OR, it is lumped willy-nilly >>into an obscure, convoluted centralized database... but it's usually not >>there. > >The Windows registry was a Bad Idea. The filing system should be the >registry. Or at least that's how we did it over on RISC OS. The closest we >ever got to a registry was that there was a directory in the !Boot >application which stored software settings. > >>The fact is that system objects (files, devices, etc) have relationships >>with each other. Relationships are nothing new in computer science, >>we've been using relational databases for years. So why hasn't the >>rather simple concept of "Object-A somehow-relates-to Object-B" filtered >>down to the operating system level yet? > >Ooh! I've been thinking about this sort of thing for ages and it's >recently crystallised that all these separate threads in my head were >connected. I envisioned some way of creating objects (not in the dumbass >OOP way, these would be autonomous) where the user would give them >semantic value by relating them to one another. Goodbye stupid >hierarchical filing system, hello Data Network. Currently, I'm thinking >that simple unidirectional and bidirectional links would be made into >complex ones by associating a controlling object to that link. > >Ok, so maybe it's still a bit messy, but I'm getting there. Anybody I've >suggested this idea to has thought I'm a raving lunatic, but I think it >has virtue. > >>We have dependencies in 'make' scripts, > >That's because C had flaws which later languages such as Java (for all >it's flaws) cured. > >>and we have 'install' and 'uninstall' scripts, > >We shouldn't need these. It should simply be a matter of deleting the >application. > >>and we have 'file type detection' in the form of >>either file suffix or 'magic'... > >Nobody said *nixen were perfect, did they? > >>but these are all loose and scattered >>functions. In the meantime, you can rename a file and none of these >>measures are the wiser. > >The file name should be independent of the semantic relations of a file. >In fact, why do we need filenames anyway? All they are are summaries of >the file's content. > >>There are some other little things that are just shell/filesystem >>issues. I'd love for a shell to be able to change to a directory just >>by typing its name. It should be obvious that "enter" is the verb to be >>implied when you just name a directory. > >Hmmm... > >>I'd also love for the FS to be >>able to treat archive files (.tar, .tgz, .zip, .bz2 etc) as directories, > >We could do this on RISC OS. Such things were called Image Filing Systems. >It rocked. > >>This sentence was false half an hour ago, did you notice? > >No, I was too busy playing XKobo :-) K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Mon, 02 Jul 2001 22:49:37 +0100 From: Keith Gaughan Subject: Fwd: Re: a linux rant (was: pretentious rant about operating And the other one: >Date: Mon, 02 Jul 2001 21:16:52 +0100 >To: Gerson.Kurz@t-online.de (Gerson Kurz) >From: Keith Gaughan >Subject: Re: a linux rant (was: pretentious rant about operating systems) > >At 20:48 02/07/2001 +0100, Gerson Kurz wrote: > >>Chris, Metadata: you have problems! Here is what happened the weekend before >>last: >> >>I am dissatisfied with Windows again and decide its once again try working >>my way with Linux. I buy the brand new Suse Linux 7.2 Pro at 130,- DM. >>"Free", really. I go home and boot the DVD -> the boot process stops and >>shows a blank screen. For the 130,- DM I have the right to phone a support >>line; I stay for nearly one hour on the phone with support until they give >>up: my machine is not supported, write down all hardware & send a letter. >>[Its a brandnew Dual 1GHz P3 with 512m]. > > > >I've been meaning to do it for a while now, and at the weekend I took the >plunge: I formatted one of the partitions on my laptop and installed >Linux. (Well, I backed up all the data I needed from Windows, reinstalled >Windows 2000 on the first partition and then installed Linux on the other. >I'd had a bad weekend and I needed something to cheer me up. > >I'm burned a copy of Mandrake Linux 8 onto CD which I'd got with PC Plus >Magazine (I intend making a contribution to Mandrakesoft now). Set up my >partitions with about the best partitioning tool I've ever used and only >when it came to setting up X did I have problems. It forced me to use >800x600 in the beginning even though I said my screen was capable of >displaying 1024x768--I fixed that after the OS had been installed. > >I haven't had anywhere near the number of problems you had, Gerson. Apart >from the fact that Enlightenment won't display the User Menu (I prefer >Window Maker anyway so it doesn't matter that much), one of the themes in >Enlightenment is buggy (Blue Metal, AFAIR) and KDevelop didn't install >properly (a problem I've heard others have had), I've had no problems. I'm >addicted to XKobo at the moment. > >I'm quite happy with things. KDE's anti-aliasing mightn't be quite as good >as RISC OS's, but it's definitely better than the terrible attempt at it >in Windows. > >Aside: >I have to go back to getting JServ to recognise what the WEB-INF directory >is for. I've far more experience with JRun and Tomcat and they work just >fine, but JServ seems to be too dumb. It makes no sense to me at all. Of >course, it could be all the crap that Oracle contaminated my copy of >Apache with. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Mon, 2 Jul 2001 15:01:43 -0700 (MST) From: Jeff Johnston Subject: Re: Bloody Listar! nope its not just you... I'm using unix Pine.. and from what I can see, basically what is going on is the To: and Cc: are switched around the wrong way. To: should be misc@esoteric.sange.fi and Cc: should be whoever we are repling to. That way if we hit Reply, it automatically defaults to the list rather than a single person, but if you hit Reply to All you can still mail the person (just gotta strip out the list). Jeff On Mon, 2 Jul 2001, Keith Gaughan wrote: > Ok, every time I've posted stuff to the list, it's gone back to the sender. > Why? Listar doesn't seem to include something as simple as a Reply-To: in > the email header when it forwards on mail. Is this just happening to me? > Could it be Eudora that's causing the problems? Panu, if it really is > Listar, could you please fix this? > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! > > > ------------------------------ Date: Mon, 02 Jul 2001 23:36:45 +0100 From: Keith Gaughan Subject: Re: Bloody Listar! Being very careful to enter the list address... At 15:01 02/07/2001 -0700, Jeff Johnston wrote: >nope its not just you... > >I'm using unix Pine.. and from what I can see, basically what is going on >is the To: and Cc: are switched around the wrong way. To: should be >misc@esoteric.sange.fi and Cc: should be whoever we are repling to. That >way if we hit Reply, it automatically defaults to the list rather than a >single person, but if you hit Reply to All you can still mail the person >(just gotta strip out the list). That doesn't quite make sense to me. Here's the header of the email you sent to the list as Eudora gives it: Delivered-To: di2834984@mail-03 X-Sent-via: StarNet http://www.azstarnet.com/ Date: Mon, 2 Jul 2001 15:01:43 -0700 (MST) From: Jeff Johnston X-X-Sender: To: Subject: Re: Bloody Listar! Sender: misc-bounce@esoteric.sange.fi X-original-sender: jeffryj@azstarnet.com Which is all fine, but why would it take the respondee's address from To:? It makes more sense to me to take it from Reply-To: or From:. Here's a header (with that person's details and the list's address--it's invitation only--blanked out) from another list I'm on: Delivered-To: di2834984@mail-03 X-Authentication-Warning: newsvendor.com: ewsvend set sender to owner-open@***.ie using -f From: "***" <***@rte.ie> To: Subject: Open: Micropayments and the New/Old Economy Date: Thu, 28 Jun 2001 15:28:36 +0100 X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) Importance: Normal Sender: owner-open@***.ie Reply-To: open@***.ie This all makes sense to me and I've had no problems with any other lists. Here's the header of a post Steve sent to the Cats Eye list in it's dying days: Delivered-To: di2834984@mail-02 Mailing-List: contact list-help@catseye.mb.ca; run by ezmlm Reply-To: list@catseye.mb.ca Delivered-To: mailing list list@catseye.mb.ca From: Steve Mosher To: list@catseye.mb.ca Subject: [Tara MacLean] Poor Boy Date: Tue, 22 May 2001 10:41:32 -0300 X-Mailer: KMail [version 1.0.28] Sender: misc-bounce@esoteric.sange.fi X-original-sender: goat@isn.net All fine, makes perfect sense. My question is: why does Listar leave out the Reply-To: address in the header. When it receives a mail, it should take the address in the To: field and insert a Reply-To: field with this address in it. I can't see why it's working for everybody else and not me--is there something I'm missing? >On Mon, 2 Jul 2001, Keith Gaughan wrote: > > > Ok, every time I've posted stuff to the list, it's gone back to the sender. > > Why? Listar doesn't seem to include something as simple as a Reply-To: in > > the email header when it forwards on mail. Is this just happening to me? > > Could it be Eudora that's causing the problems? Panu, if it really is > > Listar, could you please fix this? K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Mon, 02 Jul 2001 23:44:55 +0100 From: Keith Gaughan Subject: Fwd: Re: Bloody Listar! Would you believe, I just did it again! Fucking L*st*r! >Date: Mon, 02 Jul 2001 23:43:01 +0100 >To: Keith Gaughan >From: Keith Gaughan >Subject: Re: Bloody Listar! > >At 23:36 02/07/2001 +0100, Keith Gaughan wrote: > >>At 15:01 02/07/2001 -0700, Jeff Johnston wrote: >> >>>nope its not just you... >> >>That doesn't quite make sense to me. > >To be a bit more precise, I meant to say that Listar's behaviour makes no >sense to me. When I read this after I sent it, it seemed as if I was >contradicting you. Sorry about that. Still, I can see a light irony in the >whole thing... :) K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Tue, 3 Jul 2001 01:21:34 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Re: pretentious rant about operating systems On Mon, 2 Jul 2001, Brian Connors wrote: > > --- "markus.kliegl" wrote: > > > > > > On Sun, 1 Jul 2001, Chris Pressey wrote: > > > > > > > > There are some other little things that are just > > shell/filesystem > > > issues. I'd love for a shell to be able to change > > to a directory just > > > by typing its name. It should be obvious that > > "enter" is the verb to be > > > implied when you just name a directory. > > > > I'm quite sure that would be simple enough to > > change... most shells these > > days hash executables in the PATH, so it should be > > easy enough to detect > > if there's an executable of the same name as the > > directory and then ask > > the user. > > Yeah... it's definitely a shell issue, nothing more. > You still have to execute a "cd" somewhere. Changing to a different directory is simply a matter of setting PWD to a new value (and cd sets OLDPWD as well). I think I don't quite understand you... the way I see it, changing to a different directory requires changing to a different directory :-) In bash, due to auto-completion features, one can type , press TAB and it will automatically expand it to / so maybe if a single word that ends in '/' is read on a line, bash could assume that we want to change directories... that would resolve ambiguaties about whether to execute a command or change to a different directory as well. Or maybe could be expanded into "cd ". If someone really feels like changing a shell, I've included a simple patch for bash (for the execute_command.c file) that most likely won't work (in particular I have no idea what the flags are supposed to be), but might be a step in the right direction (or in the wrong place altogether). > > > Interesting, though... having to type 'cd' to change > > directories never > > bothered me. 'some_dir' just doesn't seem right to > > me for changing to > > directory some_dir. Why not have ENTER > > mean 'ls ' > > instead? Or something else if you get what I mean > > :-) > > > > > I'd also love for the FS to be > > > able to treat archive files (.tar, .tgz, .zip, > > .bz2 etc) as directories, > > > so you could "enter" them and copy files in and > > out with the usual > > > commands (not to mention having programs like > > "find" able to search > > > within archives automatically :) > > > > There was an interesting article on freshmeat a > > while ago... > > http://freshmeat.net/articles/view/237/ > > I think it's been done here and there anyway -- I > believe gmc (the file manager from GNOME) does this > fairly transparently. Konqueror doesn't, which is > mildly annoying. > > > ===== > -- > > __________________________________________________ > Do You Yahoo!? > Get personalized email addresses from Yahoo! Mail > http://personal.mail.yahoo.com/ > -- Attached file included as plaintext by Listar -- -- File: exec-patch 3393c3393 < --- > 3404c3404,3416 < internal_error ("%s: is a directory", command); --- > { > Function *cd_builtin; > WORD_LIST *args; > > args = xmalloc (sizeof (WORD_LIST)); > args->next = xmalloc (sizeof (WORD_LIST)); > args->word->word = xmalloc (3); > args->next->word->word = xmalloc (strlen(command) + 1); > strcpy (args->word->word, "cd"); > strcpy (args->next->word->word, command); > cd_builtin = find_shell_builtin ("cd"); > execute_builtin (cd_builtin, args, 0, 0); > } ------------------------------ Date: Mon, 2 Jul 2001 17:08:41 -0700 (MST) From: Jeff Johnston Subject: Re: Bloody Listar! Well something IS different, because before I could just hit reply, but now I have to Reply, to all recipients, then modify it so it sends to the list and not just the person On Mon, 2 Jul 2001, Keith Gaughan wrote: > Being very careful to enter the list address... > > At 15:01 02/07/2001 -0700, Jeff Johnston wrote: > > >nope its not just you... > > > >I'm using unix Pine.. and from what I can see, basically what is going on > >is the To: and Cc: are switched around the wrong way. To: should be > >misc@esoteric.sange.fi and Cc: should be whoever we are repling to. That > >way if we hit Reply, it automatically defaults to the list rather than a > >single person, but if you hit Reply to All you can still mail the person > >(just gotta strip out the list). > > That doesn't quite make sense to me. Here's the header of the email you > sent to the list as Eudora gives it: > > Delivered-To: di2834984@mail-03 > X-Sent-via: StarNet http://www.azstarnet.com/ > Date: Mon, 2 Jul 2001 15:01:43 -0700 (MST) > From: Jeff Johnston > X-X-Sender: > To: > Subject: Re: Bloody Listar! > X-original-sender: jeffryj@azstarnet.com > > Which is all fine, but why would it take the respondee's address from To:? > It makes more sense to me to take it from Reply-To: or From:. Here's a > header (with that person's details and the list's address--it's invitation > only--blanked out) from another list I'm on: > > Delivered-To: di2834984@mail-03 > X-Authentication-Warning: newsvendor.com: ewsvend set sender to > owner-open@***.ie using -f > From: "***" <***@rte.ie> > To: > Subject: Open: Micropayments and the New/Old Economy > Date: Thu, 28 Jun 2001 15:28:36 +0100 > X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) > Importance: Normal > Reply-To: open@***.ie > > This all makes sense to me and I've had no problems with any other lists. > Here's the header of a post Steve sent to the Cats Eye list in it's dying days: > > Delivered-To: di2834984@mail-02 > Mailing-List: contact list-help@catseye.mb.ca; run by ezmlm > Reply-To: list@catseye.mb.ca > Delivered-To: mailing list list@catseye.mb.ca > From: Steve Mosher > To: list@catseye.mb.ca > Subject: [Tara MacLean] Poor Boy > Date: Tue, 22 May 2001 10:41:32 -0300 > X-Mailer: KMail [version 1.0.28] > X-original-sender: goat@isn.net > > All fine, makes perfect sense. My question is: why does Listar leave out > the Reply-To: address in the header. When it receives a mail, it should > take the address in the To: field and insert a Reply-To: field with this > address in it. I can't see why it's working for everybody else and not > me--is there something I'm missing? > > >On Mon, 2 Jul 2001, Keith Gaughan wrote: > > > > > Ok, every time I've posted stuff to the list, it's gone back to the sender. > > > Why? Listar doesn't seem to include something as simple as a Reply-To: in > > > the email header when it forwards on mail. Is this just happening to me? > > > Could it be Eudora that's causing the problems? Panu, if it really is > > > Listar, could you please fix this? > > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! > > > ------------------------------ Date: Tue, 03 Jul 2001 09:37:03 +0200 From: =?iso-8859-1?Q?Fr=E9d=E9ric?= van der Plancke Subject: Re: Bloody Listar! I have a similar but different problem... Netscape's "Reply-To" button yields "misc-bounce@esoteric.sange.fi" as reply address to my new message. If I don't delete the "-bounce" the message never reaches the misc list. Here are the headers I receive (some of them obviously added by our MDaemon internal router; others added by lovely Netscape). Given these headers the "misc-bounce" thing is no surprise. |Return-path: |Received: [snipped] |Message-Id: <5.0.2.1.0.20010702234412.00a8fdf0@mail.9netave.com> |X-Sender: di2834984@mail.9netave.com |X-Mailer: QUALCOMM Windows Eudora Version 5.0.2 |Date: Mon, 02 Jul 2001 23:44:55 +0100 |To: misc@esoteric.sange.fi |From: Keith Gaughan |Subject: Fwd: Re: Bloody Listar! |Mime-Version: 1.0 |Content-Type: text/plain; charset="us-ascii"; format=flowed |X-listar-version: Listar v0.129a |Sender: misc-bounce@esoteric.sange.fi |Errors-to: misc-bounce@esoteric.sange.fi |X-original-sender: keith@digital-crew.com |Precedence: bulk |X-list: misc |X-MDRcpt-To: fvdp@decis.be |X-MDRemoteIP: 195.170.158.190 |X-Return-Path: misc-bounce@esoteric.sange.fi |X-MDaemon-Deliver-To: FPlancke@decis.be |Reply-To: misc-bounce@esoteric.sange.fi |X-Mozilla-Status: 8001 |X-Mozilla-Status2: 00000000 |X-UIDL: MD50000005915:MSG:3431:133710979 Frédéric vdP Keith Gaughan wrote: > > Ok, every time I've posted stuff to the list, it's gone back to the sender. > Why? Listar doesn't seem to include something as simple as a Reply-To: in > the email header when it forwards on mail. Is this just happening to me? > Could it be Eudora that's causing the problems? Panu, if it really is > Listar, could you please fix this? > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! ------------------------------ Date: Tue, 3 Jul 2001 16:55:49 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Re: pretentious rant about operating systems [replying to myself...] On Tue, 3 Jul 2001, markus.kliegl wrote: > > > On Mon, 2 Jul 2001, Brian Connors wrote: > > If someone really feels like changing a shell, I've included a simple > patch for bash (for the execute_command.c file) that most likely won't > work (in particular I have no idea what the flags are supposed to be), but > might be a step in the right direction (or in the wrong place altogether). > Sorry about that code... looking at it again, it's incredibly ugly (I forgot to free the allocated memory, causing memory leaks and forgot to set args->next->next to NULL). By the way, does anyone know what 'xmalloc' is? It was used all over the code, so I suppose it's simply some sort of wrapper of malloc. Markus ------------------------------ Date: Tue, 3 Jul 2001 18:32:40 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Fwd: Announce: ICFP programming contest I hope you all participate with Brainfuck, Intercal and the lot :-) Unfortunately I'm gone at the time :-( Hope you have fun, Markus From: Damien Doligez Newsgroups: comp.programming,comp.lang.apl,comp.lang.c,comp.lang.c++,comp.lang.dylan,comp.lang.eiffel,comp.lang.forth,comp.lang.functional,comp.lang.icon,comp.lang.lisp,comp.lang.misc,comp.lang.modula2,comp.lang.modula3,comp.lang.oberon,comp.lang.objective-c,comp.lang.perl.misc,comp.lang.prolog,comp.lang.python,comp.lang.sather,comp.lang.scheme,comp.lang.smalltalk,comp.lang.tcl Subject: Announce: ICFP programming contest Date: 3 Jul 2001 15:37:49 GMT We are pleased to announce: The Fourth Annual ICFP PROGRAMMING CONTEST July 26-29, 2001 http://cristal.inria.fr/ICFP2001/prog-contest/ Convinced your favorite programming language provides unbeatable productivity? Convinced you and your friends are world-class programmers? If so, we're providing you the opportunity to prove it! We are pleased to announce the Fourth Annual ICFP Programming Contest to be held in conjunction with the 2001 International Conference on Functional Programming (ICFP 2001). All programmers are invited to enter the contest, either individually or in teams; we especially encourage students to enter. You may use any programming language (or combination of languages) to show your skill. On Thursday, July 26, 2001 at 15:00 UTC, we will publish a challenge task on the Web site and by e-mail to registered participants. Teams will have until Sunday, 29 July 2001, 15:00 UTC (72 hours) to implement a program to perform this task and submit it to the contest judges. We've designed the contest for direct, head-to-head comparison of language technology and programming skill. We have a range of prizes for the winners: cash awards, special student prizes, and, of course, unlimited bragging rights. For more information about the contest, prizes, and registration, point your browser to: http://cristal.inria.fr/ICFP2001/prog-contest/ For more information about ICFP 2001, see: http://cristal.inria.fr/ICFP2001/ -- Damien Doligez, Luc Maranget, Pierre Weis ------------------- Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/ To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr ------------------------------ From: "Chris Pressey" Subject: various Date: Tue, 3 Jul 2001 16:23:49 -0500 Re Listar... Panu has said a couple of times that misc can't add a Reply-To header for some reason (because traffic from lang is mirrored onto misc, or something.) So just hit 'Reply to All' instead of 'Reply to Sender' (which actually makes a certain amount of sense.) You can doctor the To: and Cc: fields so that you don't send off useless duplicates to the poster, or, if you don't, I'm sure the poster won't mind (this is the situation on the Erlang list, last I checked.) But... the strange thing is... me, being the poster of one of those e-mails, should have received the single-replies that some people sent.... but, I didn't. And I have no idea what "misc-bounce" is all about. So I don't have the whole story, here... Re mailing lists and e-mail in general... maybe we should be bitching about the general awkwardness and unreliability e-mail instead of operating systems. :) Re metadata... I tend to believe in "the filesystem IS the database" - making things like a seperate "registry" are just patching on contingency measures instead of fixing the problem. (The problem is even worse in the Unix world where there are n "third-party" registries; who cares if they're open source iffen they're not sponsored by the operating system? Metadata is too integral a component for everyone to be using a slightly different way of storing it...) Re filenames... "just" a summary of the contents of a file? Well, no, there's got to be more to it than that. It's hard to find a book when you don't know it's title... how does an app find it's configuration file, if files don't have names? Look through every nearby file until it finds one that fits the bill? The name serves as a "global relationship"... even if it were replaced by a relationship, it'd still have to be name-like... like... a named relationship. Or something :) It does make a certain amount of sense for a name to only be a 'link', so that a file could be referred to by more than one name (but hopefully, y'know, something a little more elegant than Unix' symbolic links...) Re Linux & FreeBSD. Unix will certainly not threaten the 'user' market of Mac and Windows until it's packaged in such a way that little 'hacky' tricks like Robert mentioned don't have to be used to use X windows. Really, I've encountered enough of those little *non*intuitive things already ("Whattya mean, I forgot to chown the .xsession file?" "Whattya mean, *CTRL* plus alt plus function key to swap screens?" etc) I also used to think browser software for Windows was buggy - now I've found web sites that not only silently crash my browser (NS4.76), I've actually found ones that crash X windows completely, with nary a chance to save my work. This, ladies and gentlemen, is why there is an option in Netscape to disable Java... because, heh, it doesn't work. Imagine that! :) I've said it before, I'll say it again: web pages are *documents*, and documents should *not* crash. The idea of embedding "scripting" languages and such in documents is just... shall we say conceptually flawed, perhaps even doomed but no one has noticed yet :) If there's something that pisses me off about FreeBSD it's availability of commercial products for it. When a company (such as, oh, I dunno, RealNetworks,) develops a product, why they want the greatest market share of course, they have to maximize the number of small green pieces of paper they acquire of course, so of course they're going to target Windows. Then maybe Mac, you know. And sure, why not Solaris... but what about people running "free" operating systems? Why, that'd be the "community" project... which, of course, has done it's best to make it available for Linux... and presumably FreeBSD can't be far behind... but what pisses me off is that in the meantime between releases, Real is kind enough to make old versions of their software unobtainable. Practical upshot of all this is that I can't listen to half the sound files on the net... unless I want to be bothered into hacking a bunch of Netscape-related code. Or running Linux Netscape under the Linux-emulation-layer (which I haven't had the best of luck with so far.) Also, I'm really not sure why downloading the "community" version of RealPlayer should elicit Real to be even nosier with their download form, but they are. I mean, if it's a community version, I ought to be able to get it from somewhere besides Real, right? Well - I haven't seen it anywhere else, so far... Ok, ok. Enough bitching for now :) Chris ------------------------------ Date: Tue, 03 Jul 2001 14:58:51 -0700 From: Russell Bornschlegel Subject: Re: various Chris Pressey wrote: > I've said it before, I'll say it again: web pages are *documents*, and > documents should *not* crash. The idea of embedding "scripting" languages > and such in documents is just... shall we say conceptually flawed, perhaps > even doomed but no one has noticed yet :) It may be conceptually flawed, but in the same way as the idea that "normal" people should be expected to use computers, it's an idea that's here to stay. You could have fought this in 1993, but not in 2001. Some web pages are documents. Some are active, dynamic environments. I will meet you halfway on this and say that there should be a clearer distinction between the two -- different protocols or something (dhttp: for dynamic), but it's just pissing into the tide [sic] to try and complain about this now. > ...but what pisses me off is that in > the meantime between releases, Real is kind enough to make old versions of > their software unobtainable. I would really love to see more companies releasing copyright and distributing source to their old software... ahhh, it's too hot to start this rant. -RB ------------------------------ Date: Wed, 04 Jul 2001 00:13:25 +0100 From: Keith Gaughan Subject: Re: various (fortified with Keith yacking on about RISC OS) "Taking Chris's advice on 'Reply to All', Keith decided to do just that..." At 16:23 03/07/2001 -0500, Chris Pressey wrote: >Re Listar... Panu has said a couple of times that misc can't add a Reply-To >header for some reason (because traffic from lang is mirrored onto misc, or >something.) Hmmm... (my favourite noise at the moment) I'm compelled to hack listar's source to see if I could fix this little anomaly. Probably not though, but it'd still be fun to try. Then again, I hardly have time to work on my Eloi language, never mind hack a listserver. >So just hit 'Reply to All' instead of 'Reply to Sender' (which actually >makes a certain amount of sense.) But it'd make even more sense if both the sender's address and the list's address didn't both appear :( >Re mailing lists and e-mail in general... maybe we should be bitching about >the general awkwardness and unreliability e-mail instead of operating >systems. :) I wonder--I've never found email unreliable, you must just be unlucky Chris. >Re metadata... I tend to believe in "the filesystem IS the database" - >making things like a seperate "registry" are just patching on contingency >measures instead of fixing the problem. I've always found the idea of a system registry outside of the filing system a bit silly, though all my godless Windows-loving friends can't see how things'd work without one. On RISC OS (an OS I get slagged off for still loving to use and program for), you could throw an application anywhere you wanted -- it was all up to you. This was because the structure of applications on RISC OS was well thought out. Example: I wrote a small utility many moons ago called 'Pollster', which allowed me to calculate certain magic numbers which help keep my software polling the OS efficiently. It's structure was a good example of how a RISC OS application was structured. Here's a tree: !Pollster <-- The application directory. Notice the exclamation mark. | |-- !Boot <-- A script file executed the first time the application is 'seen' by | the OS. Sets up various bits and pieces. |-- !Help <-- A file run if help is requested for this application. Usually text. | |-- !Run <-- A script file executed when the application is executed. It is the | application's bootstrapping program. |-- !RunImage <-- The primary application binary image, that it to say, the real thing. | |-- !Sprites <-- The application's icons. This file is usually loaded into memory by | the !Boot file, or the OS if the !Boot file wasn't present. |-- Resources <-- Directory holding the application's resources such as extra | icons/images, window templates, files containing version's of the | app's messages in various languages, modules (think Unix DSOs) etc... : . . A typical !Boot file would be something like: | | > !Boot file for FooBar v0.01 (01-Apr-1999) | Set FooBar$Dir Iconsprites .!Sprites If "" = "" Then Set Alias$RunType_666 .!Run %%0 If "" = "" Then Set File$Type_666 EvilWord And a typical !Run file would be something like: | | > !Boot file for FooBar v0.01 (01-Apr-1999) | Set FooBar$Dir Iconsprites .!Sprites Set Alias$RunType_666 .!Run %%0 Set File$Type_666 EvilWord WimpSlot -min 640k -max 640k Run .!RunImage %*0 It looks like gobbledegook, I know, but this is all quite useful stuff. The 'Set' command sets system variables. If you want to read the value of system variables, you enclose them in angle brackets. There's no real registry anywhere saying which application is run when file xyz is run, instead the closest is that applications register themselves by setting system variables (there's a set of system variable 'shapes' which serve different purposes) if the place isn't gone. That's what the third line of code in each of the files is about. In this example, the !Boot file checks to see if another program has registered itself to handle files with the numerical file type 666. If there is none, it sets itself up to handle them, and on the next line, gives that file type a human-readable name. The !Run file does much the same thing in this case, but forces itself to be registered as that filetype's handler. After that, it's just the program itself being run. There's a lot more to it. It's all pretty cool, I think. It's a pity RISC OS never shook off the image it got in its early days. >Re filenames... "just" a summary of the contents of a file? Well, no, >there's got to be more to it than that. It's hard to find a book when you >don't know it's title... That's because there's no such thing as semantic links in the real world beyond names and associations we keep in our heads. If these links and associations were part of the real word and we could manipulate them, then we'd have no need for names. That's what I mean. >how does an app find it's configuration file, if >files don't have names? Easily enough, as long as you've not constrained yourself to 'modern day' filing systems and place semantic associations between the files, which don't require a name, their existence is enough. >Look through every nearby file until it finds one >that fits the bill? Ack! That's so 1970's! My application object just takes a look at the objects associated with it and queries the object managing it's configuration details... No, no, wait... it's semantic relations to other objects *are* it's configuration. Nice slight of hand, eh! >The name serves as a "global relationship"... even if >it were replaced by a relationship, it'd still have to be name-like... >like... a named relationship. Or something :) Possibly, but not necessarily. You could give a relationship a name, or an object a name, just it's still little more than a mnemonic for your own convenience. You could easily get by without it. >I've said it before, I'll say it again: web pages are *documents*, and >documents should *not* crash. The idea of embedding "scripting" languages >and such in documents is just... shall we say conceptually flawed, perhaps >even doomed but no one has noticed yet :) I'm sort of with Russell on this one. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Tue, 03 Jul 2001 19:15:04 -0500 From: Chris Pressey Subject: Re: various Russell Bornschlegel wrote: > Chris Pressey wrote: > > I've said it before, I'll say it again: web pages are *documents*, and > > documents should *not* crash. The idea of embedding "scripting" languages > > and such in documents is just... shall we say conceptually flawed, perhaps > > even doomed but no one has noticed yet :) > It may be conceptually flawed, but in the same way as the idea that > "normal" people should be expected to use computers, it's an idea > that's here to stay. You could have fought this in 1993, but not > in 2001. I doubt that I could have fought it in '93, actually. HTML just kin'a took over in that path-of-least-resistance way; it was never intended to be a layout format, much less a carrier for code... Java is a good idea, in that it runs in a 'sandbox'; Javascript is much less of a good idea. When you think about it, Javascript is probably the "crashingest" language on the planet. Supporting n different scripting languages is even a WORSE idea... So the problem here is really complexity. The Powers that Be got together and said, "We shall make this Far More Complex Than it Really Needs to Be," as they so often do. And the result is that there's seemingly no browser on earth that's sufficiently "compliant" enough to not crash/fail-to-display-effectively on at least one of a hundred web sites selected at random... > Some web pages are documents. Some are active, dynamic environments. > I will meet you halfway on this and say that there should be a > clearer distinction between the two -- different protocols or > something (dhttp: for dynamic), but it's just pissing into the tide > [sic] to try and complain about this now. Well, I'm certainly not expecting anything to *change* because of it, I'm just venting... :) I dunno, y'know, web "page" is almost a total misnomer these days. I'm guessing it would be easier to go the other way around; surrender http as the dynamic protocol, and resort to PDF/DVI/PS/something/other for static documents. And I should clarify that dynamic environments ideally shouldn't crash either of course! But it's easier to forgive a programmer for messing that up, than it is to forgive them for messing up something that should be *simple*. Think of how many pages you've seen with Java[script] that is merely gratuitous, hardly counting as 'dynamic environment'. > > ...but what pisses me off is that in > > the meantime between releases, Real is kind enough to make old versions of > > their software unobtainable. > I would really love to see more companies releasing copyright and > distributing source to their old software... ahhh, it's too hot to > start this rant. The problem here is the lawyers, of course. The companies would actually *profit* from doing this kin'a stuff. (If not monetarily, then in customer confidence.) But, they can't see the forest because their lawyers keep shouting "TREES! TREES!" at them. :) It's silly to think that there's "still profit to be had" from obsolete software; if there was, then the software wouldn't be *obsolete*, now would it?!? :) To me that's a strong argument that "damages" based on the "piracy" of obsolete-but-still-copyrighted software would be very, very difficult to obtain in any court of law where the judge has his or her eyes open... But, that could just be my imagination :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Tue, 03 Jul 2001 19:44:53 -0500 From: Chris Pressey Subject: Re: various (fortified with Keith yacking on about RISC OS) Keith Gaughan wrote: > >So just hit 'Reply to All' instead of 'Reply to Sender' (which actually > >makes a certain amount of sense.) > But it'd make even more sense if both the sender's address and the list's > address didn't both appear :( Well... yes... but you can't have everything :) > >Re mailing lists and e-mail in general... maybe we should be bitching about > >the general awkwardness and unreliability e-mail instead of operating > >systems. :) > I wonder--I've never found email unreliable, you must just be unlucky Chris. Or maybe you've just never noticed? :) Granted it seems mostly reliable. but I worry a lot when people reply to messages on a mailing list that *I've never seen*. This has happened both with my ol' list and this list; e.g. I don't seem to have ever received Brian's "If someone really feels like changing a shell," message (although maybe that's because he only sent it to markus, I dunno.) Further to that - if someone *is* gonna hack Listar, could you *please* also make it so that it adjusts the timestamp on each message to match the *list's* local time, i.e. pretend it was sent at the time the list manager receieved it? I'm kind of sick of receiving replies before their original posts, simply because someone's local internal clock is messed up. > >Re metadata... I tend to believe in "the filesystem IS the database" - > >making things like a seperate "registry" are just patching on contingency > >measures instead of fixing the problem. > > I've always found the idea of a system registry outside of the filing > system a bit silly, though all my godless Windows-loving friends can't see > how things'd work without one. Well, these are probably the same people who can't see how you'd live without filename extensions... > On RISC OS (an OS I get slagged off for > still loving to use and program for), you could throw an application > anywhere you wanted -- it was all up to you. This was because the structure > of applications on RISC OS was well thought out. [...] > It's all pretty cool, I think. It's a pity RISC > OS never shook off the image it got in its early days. I dunno... for me, it's a problem with the name... RISC has nothing to do with OS. Unless it's a "reduced-instruction-set" OS, I suppose...? Hell, "RISC" has little or nothing to do with seperate load and store instructions, so... in for a penny, in for a pound :) > >Re filenames... "just" a summary of the contents of a file? Well, no, > >there's got to be more to it than that. It's hard to find a book when you > >don't know it's title... > That's because there's no such thing as semantic links in the real world > beyond names and associations we keep in our heads. If these links and > associations were part of the real word and we could manipulate them, then > we'd have no need for names. That's what I mean. Well, yeah... names act as a proxy for chemistry. Sorta. > >how does an app find it's configuration file, if > >files don't have names? > Easily enough, as long as you've not constrained yourself to 'modern day' > filing systems and place semantic associations between the files, which > don't require a name, their existence is enough. Hm, yes - but then, you can only find the files by going through the application. Which may not be a bad thing, it just runs counter-intuitive to probably most OS-user's sense of "what should be" (which has simply been fostered by "what currently is")... > >Look through every nearby file until it finds one > >that fits the bill? > > Ack! That's so 1970's! My application object just takes a look at the > objects associated with it and queries the object managing it's > configuration details... No, no, wait... it's semantic relations to other > objects *are* it's configuration. Nice slight of hand, eh! But if you take it too far, then all of your objects are empty and ALL the useful information exists only in the relationships between them! Which is fascinating, but maybe not totally efficient :) > >The name serves as a "global relationship"... even if > >it were replaced by a relationship, it'd still have to be name-like... > >like... a named relationship. Or something :) > Possibly, but not necessarily. You could give a relationship a name, or an > object a name, just it's still little more than a mnemonic for your own > convenience. You could easily get by without it. Sometimes those mnemonics for your own convenience are pretty damned convenient, though; like, if I make four different versions of the same image, I want to differentiate them, probably with names like "foo1" "foo2" "foo3" etc. I suppose one could use "next" and "prev" relationships, like a linked list, but that strikes me as at least a little awkward. At the very least, you still practically need an "annotation" field on a file if nothing else. I guess what I'm getting to is that, names or no, you still have a single 'hub' object to which everything has a (direct or indirect) link to. Unix has /, Mac has the desktop, Windows has... well it's hard to tell, given that the desktop resides in \WINDOWS\Desktop, it's circular :) btw, if I can't save a document in "My Computer" directly, howcum it shows up in file dialogs :) > >I've said it before, I'll say it again: web pages are *documents*, and > >documents should *not* crash. The idea of embedding "scripting" languages > >and such in documents is just... shall we say conceptually flawed, perhaps > >even doomed but no one has noticed yet :) > > I'm sort of with Russell on this one. I should say that I am, too. I can't change it and I wouldn't even want it to change, really, but I would like to see some way, some measure of ensuring that a page most likely won't crash. I mean, some way besides self-restraint, which not enough webweavers seem to have very much of :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Tue, 03 Jul 2001 17:57:31 -0700 From: Russell Bornschlegel Subject: Re: various Chris Pressey wrote: > To me that's a strong argument that "damages" based on the "piracy" of > obsolete-but-still-copyrighted software would be very, very difficult to > obtain in any court of law where the judge has his or her eyes open... Legal fees at some point will vastly eclipse damages, and work just as well as a deterrent. Personally, I believe piracy of obsolete-but- still-copyrighted software is ethically Wrong; it's not up to the pirate to decide when the software deserves to be free, it's up to the owner of the IP. I may wish they would release their IP sooner rather than later, but my wishes are not a valid claim on their IP. I mean, I wish Bill Gates would give me a million dollars -- he wouldn't miss it -- but I don't expect him to. -R ------------------------------ Date: Tue, 03 Jul 2001 18:06:16 -0700 From: Russell Bornschlegel Subject: Re: various (fortified with Keith yacking on about RISC OS) Chris Pressey wrote: > > > > Ack! That's so 1970's! My application object just takes a look at the > > objects associated with it and queries the object managing it's > > configuration details... No, no, wait... it's semantic relations to other > > objects *are* it's configuration. Nice slight of hand, eh! > > But if you take it too far, then all of your objects are empty and ALL > the useful information exists only in the relationships between them! > Which is fascinating, but maybe not totally efficient :) I'll once again launch the "filesystem node = raw binary block + key/value pair property data" meme, and see if anyone salutes. > > >I've said it before, I'll say it again: web pages are *documents*, and > > >documents should *not* crash. The idea of embedding "scripting" languages > > >and such in documents is just... shall we say conceptually flawed, perhaps > > >even doomed but no one has noticed yet :) > > > > I'm sort of with Russell on this one. > > I should say that I am, too. I can't change it and I wouldn't even want > it to change, really, but I would like to see some way, some measure of > ensuring that a page most likely won't crash. I mean, some way besides > self-restraint, which not enough webweavers seem to have very much of :) How about a restricted or private TLD called .static or something? Any use of any Javascript, other embedded scripting language, or other escape route in an .html file in a .static domain would be grounds for the managers of the TLD to kill that domain. :) -RB ------------------------------ Date: Tue, 03 Jul 2001 20:49:46 -0500 From: Chris Pressey Subject: intellectual property (was Re: various) Russell Bornschlegel wrote: > Chris Pressey wrote: > > To me that's a strong argument that "damages" based on the "piracy" of > > obsolete-but-still-copyrighted software would be very, very difficult to > > obtain in any court of law where the judge has his or her eyes open... > Legal fees at some point will vastly eclipse damages, and work just > as well as a deterrent. Personally, I believe piracy of obsolete-but- > still-copyrighted software is ethically Wrong; I'd rather leave ethics out of it; arguably, merely using a computer would be "ethically wrong" to a luddite, right? :) > it's not up to the > pirate to decide when the software deserves to be free, it's up to > the owner of the IP. Well, speaking legally, that's the case (subject to what, exactly, you or anyone means by "free"... :) Internet Explorer is "free" in the sense that you can (in theory) obtain the binaries for no charge; but it is far from free in the sense of unrestrictedness of development. Otoh, FSF seems the other way around; vehemently against anyone (except, perhaps, themselves) making a buck off their software, even if it is unrestricted in terms of development. I mean, pleeeez. "Why you should not use the LGPL." FMH grumble grumble... Anyway, that's why I admire the BSD license, and why I consider it "freer" than most commercial software, and even GNU stuff, too. And if anyone would like to toss in their 2c about the FSF, I will be *quite* happy to moan and gripe and carp further upon the subject of how utterly silly I think the GPL is... :) > I may wish they would release their IP sooner > rather than later, but my wishes are not a valid claim on their IP. > I mean, I wish Bill Gates would give me a million dollars -- he > wouldn't miss it -- but I don't expect him to. Hell no. On the other hand, the law is pretty clear about this: if I download, I dunno, C64 Ultimate Wizard, without paying for the license, the onus is on EA - and, AFAICT, no one else - to find me and to sue me for infringement. This obviously constitutes a huge waste of resources on EA's part should they choose to do so, resources they could be using to develop new, "non-obsolete" games, which is, arguably, their job - otherwise, they'd still be selling Ultimate Wizard, which, last I checked, they ain't. Or, take something even more interesting - var'aq. Like all "fanart," it's technically in violation of Paramount's license, because it uses elements from their copyrighted material (Klingons.) But if you could argue that var'aq is a parody of Star Trek, rather than a derivative work, it would be excepted from infringement. Do you understand that distinction? I'm not sure I do - especially considering the sophisticated sense of humour that can appear in some parodies. It really freaks me out to think that this is what courts of the land are busy doing, trying to decide if some story or programming language or other is a joke or not :) Or, take the license on NT, which states you can't apply a patch to it, which contradicts what the courts have AFAIK decided about software in general: if you own it (that is, if you own a licensed copy,) then you have the right to modify it, just like you have the right to install a new muffler in a car that you own; call it the "right to aftermarket upgrading" if you like. You can put whatever you like in a license, but federal and international law take precedence. If you thought this was bad you ought to hear me rant about patents sometime :) Chris -- Epimenides sez "Feel free, Cretan" ------------------------------ Date: Tue, 03 Jul 2001 20:57:48 -0500 From: Chris Pressey Subject: Re: various (fortified with Keith yacking on about RISC OS) Russell Bornschlegel wrote: > Chris Pressey wrote: > > > Ack! That's so 1970's! My application object just takes a look at the > > > objects associated with it and queries the object managing it's > > > configuration details... No, no, wait... it's semantic relations to other > > > objects *are* it's configuration. Nice slight of hand, eh! > > But if you take it too far, then all of your objects are empty and ALL > > the useful information exists only in the relationships between them! > > Which is fascinating, but maybe not totally efficient :) > I'll once again launch the "filesystem node = raw binary block + > key/value pair property data" meme, and see if anyone salutes. I think the question we're coming to here is, if you have a key/value property list, what do you need the raw binary block for (except, perhaps, efficiency?) > > > >I've said it before, I'll say it again: web pages are *documents*, and > > > >documents should *not* crash. [...] > How about a restricted or private TLD called .static or something? Any > use of any Javascript, other embedded scripting language, or other > escape > route in an .html file in a .static domain would be grounds for the > managers of the TLD to kill that domain. :) Eek! That's a bit... fundamentalist... but I kinda like it :) I dunno, wasn't it, like, years ago (like, 1996) that everybody was talking about how we were going to have new TLD's like .store, .info and so forth? Well, here we are five years later and... I dunno about you, but I've yet to encounter one of those on the internet at large. Chris -- "Enthusiasm! Enthusiasm! Rah, rah, rah!" ------------------------------ From: "Cal Henderson" Subject: Re: various (fortified with Keith yacking on about RISC OS) Date: Wed, 4 Jul 2001 07:16:28 +0100 : It looks like gobbledegook, I know, but this is all quite useful stuff. : The 'Set' command sets system variables. If you want to read the value of : system variables, you enclose them in angle brackets. There's no real : registry anywhere saying which application is run when file xyz is run, : instead the closest is that applications register themselves by setting : system variables (there's a set of system variable 'shapes' which serve : different purposes) if the place isn't gone. That's what the third line of : code in each of the files is about. In this example, the !Boot file checks : to see if another program has registered itself to handle files with the : numerical file type 666. If there is none, it sets itself up to handle : them, and on the next line, gives that file type a human-readable name. The : !Run file does much the same thing in this case, but forces itself to be : registered as that filetype's handler. After that, it's just the program : itself being run. : : There's a lot more to it. It's all pretty cool, I think. It's a pity RISC : OS never shook off the image it got in its early days. It's a good OS, but within the !boot lies one of it's problems. Until you've opened the folder containing your program, the file association hasn't been made. The basic file associations (text, sprites, draw, obey, command etc.) are made by the !Resources booting on startup (correct?) which suggests that the !Resources folder is a kind of registry. There's also the limitation of one (or two) applications per file in terms of association. Each file type (and file types are a good thing - MacOS is fundementally wrong here) only has two actions - dblclick and shift-dblclick, which means it can't be associated with more than two programs. To open it in another program, you need that program already open. Hmmm. --cal ------------------------------ Date: Wed, 4 Jul 2001 17:33:59 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Re: intellectual property (was Re: various) On Tue, 3 Jul 2001, Chris Pressey wrote: > > Well, speaking legally, that's the case (subject to what, exactly, you > or anyone means by "free"... :) Internet Explorer is "free" in the > sense that you can (in theory) obtain the binaries for no charge; but it > is far from free in the sense of unrestrictedness of development. Otoh, > FSF seems the other way around; vehemently against anyone (except, > perhaps, themselves) making a buck off their software, even if it is > unrestricted in terms of development. Anyone can sell their software, as long as you give whomever you sell it to the same rights that you have, i.e. they can get the souce code and can do anything they want with the software under the terms and conditions of the GPL. The idea is free software for free software in return. The GPL allows that... taking free software, making it proprietary and selling it is forbidden. > I mean, pleeeez. "Why you should > not use the LGPL." FMH grumble grumble... Anyway, that's why I admire > the BSD license, and why I consider it "freer" than most commercial > software, and even GNU stuff, too. The problem with the FSF stuff is that they take it so strict that you can't use their free software with your free software (even if it fits the FSF's definition of "free"), unless it is "GPL-compatible". This is, for example, a problem with Ocaml. Ocaml is distributed under the QPL and can't use the GNU readline library (LGPL). The GNU's "free" is more of an ensuring that your software remains free, with the trade-off of the software in total being a little less free that way :-) > > And if anyone would like to toss in their 2c about the FSF, I will be > *quite* happy to moan and gripe and carp further upon the subject of how > utterly silly I think the GPL is... :) I always wonder... how many people that find the GPL utterly silly have ever actually had a _problem_ with it? The QPL-LGPL example above is a sort of stupid thing, but how often does something like that happen? So please, go on with your "utterly silly" rant... but give me an example where you actually had a problem. I personally like putting stuff under the GPL for sake of simplicity... it is for what I want, i.e. free software that can be modified, etc. If somebody wants to do something more with the software than the GPL allows, they can send me an email asking me to dual-license the software and I will, if I don't see any problem with it (ignoring for a moment the fact that I don't write any useful software, anyway :-) > > Chris > Markus ------------------------------ Date: Wed, 04 Jul 2001 16:59:36 +0100 From: Keith Gaughan Subject: Re: various (fortified with Keith yacking on about RISC OS) At 07:16 04/07/2001 +0100, Cal Henderson wrote: >: It looks like gobbledegook, I know, but this is all quite useful stuff. >: The 'Set' command sets system variables. If you want to read the value of >: system variables, you enclose them in angle brackets. There's no real >: registry anywhere saying which application is run when file xyz is run, >: instead the closest is that applications register themselves by setting >: system variables (there's a set of system variable 'shapes' which serve >: different purposes) if the place isn't gone. That's what the third line of >: code in each of the files is about. In this example, the !Boot file checks >: to see if another program has registered itself to handle files with the >: numerical file type 666. If there is none, it sets itself up to handle >: them, and on the next line, gives that file type a human-readable name. The >: !Run file does much the same thing in this case, but forces itself to be >: registered as that filetype's handler. After that, it's just the program >: itself being run. >: >: There's a lot more to it. It's all pretty cool, I think. It's a pity RISC >: OS never shook off the image it got in its early days. > >It's a good OS, but within the !boot lies one of it's problems. Until you've >opened the folder containing your program, the file association hasn't been >made. The basic file associations (text, sprites, draw, obey, command >etc.) are >made by the !Resources booting on startup (correct?) which suggests that the >!Resources folder is a kind of registry. Are you sure you're not thinking of the !Boot application? The !Boot application is more of a repository than a registry, a standard place to put things. The basic associations are part of the OS (such as absolute files invoking the Run command, Obey files, etc.) set up when it boots first. Then it boots the applications in Resources:$.Apps. What happens to sprites, textfiles, drawfiles and the like is all up to that application. When !Edit is booted, it sets itself up to load text files, the same with !Paint with sprites, and !Draw with drawfiles. These aren't set up initially. There is a way around the whole application-must-be-seen-before-!boot-is-run problem. I had a wee program !AutoBoot which allowed me to edit a list of applications with would be Filer_Boot-ed by another small program sitting in the Tasks directory in the !Boot application. I'd just drag the application onto !AutoBoot's icon and it would add it to the list if it wasn't already there. It also had a small editor which would allow me to edit the list and could try to find applications if it got moved. If you dragged a normal directory onto it, it would register all the applications in it to be autobooted on startup. It was pretty nifty. >There's also the limitation of one (or two) applications per file in terms of >association. Each file type (and file types are a good thing - MacOS is >fundementally wrong here) only has two actions - dblclick and shift-dblclick, >which means it can't be associated with more than two programs. To open it in >another program, you need that program already open. I never understood (happens to me a lot, doesn't it!) quite why we had to set those system variables directly, rather than maybe using commands. Something like: RegisterFileHandler "" "" so rather than Set Alias$RunType_666 .!Run %%0 I'd do RegisterFileHandler "Evil Word" 666 "EvilWord" .!Run %%0 This would do the same as the line above, but the OS could also keep a record of all the previous programs which registered themselves as handlers for this kind of file *and* we'd be able to get rid of the If "" = "" Then Set Alias$RunType_666 .!Run %%0 nonsense that we have to include in !Boot files. One nice thing about Windows is the 'Open With...' submenu. If the above scheme was implemented, RISC OS could do this too and wouldn't need a registry. Shift-double-click *always* sends the file to whatever application is currently handling textfiles so I wonder if that counts. Despite poor attempts, something which RISC OS had which other desktops would benefit from is the iconbar. The iconbar made the desktop truly document-centric, something which no other desktop/GUI has ever been able to achieve. Hats off to Sophie! K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ From: Gerson.Kurz@t-online.de (Gerson Kurz) Subject: Abandonware for Programmers (was: Re: various) Date: Wed, 4 Jul 2001 19:13:10 +0100 Am I the only one for whom http://community.borland.com/museum brings up nice (?) memories ? (In case you don't want to register: They provide - Turbo Pascal 5.5 - Turbo C++ 1.01 - Turbo C 2.01 4free) ------------------------------ Date: Wed, 04 Jul 2001 13:31:28 -0500 From: Chris Pressey Subject: Re: intellectual property (was Re: various) "markus.kliegl" wrote: > On Tue, 3 Jul 2001, Chris Pressey wrote: > > FSF seems the other way around; vehemently against anyone (except, > > perhaps, themselves) making a buck off their software, even if it is > > unrestricted in terms of development. > Anyone can sell their software, as long as you give whomever you sell it > to the same rights that you have, i.e. they can get the souce code and can > do anything they want with the software under the terms and conditions of > the GPL. Which is as it should be, I was mainly rolling my eyes at FSF's general attitude, here. > The idea is free software for free software in return. The GPL allows > that... taking free software, making it proprietary and selling it is > forbidden. ...without express consent of the copyright holder, I should add; if you can get that, you can do anything you like with it. As with any license. > > I mean, pleeeez. "Why you should > > not use the LGPL." FMH grumble grumble... Anyway, that's why I admire > > the BSD license, and why I consider it "freer" than most commercial > > software, and even GNU stuff, too. > The problem with the FSF stuff is that they take it so strict that you > can't use their free software with your free software (even if it fits > the FSF's definition of "free"), unless it is "GPL-compatible". This is, > for example, a problem with Ocaml. Ocaml is distributed under the QPL and > can't use the GNU readline library (LGPL). See, that's just idiotic. Especially considering it's *L*GPL, that's supposed to relieve problems like this. In fact I suspect it might not be a problem, but the legal text is so thick that people are merely frightened that it might be a problem. The fact is a single piece of software is nigh useless alone, and no one can scientifically define the difference between "mere aggregation" and "integration." So why doesn't OCaml's author write the FSF and ask for one of the "special exceptions" that they "sometimes grant." Surely theirs is not against the "spirit of free software." Imagine, for example, if OCaml just went ahead and included gnureadline anyway. Would FSF sue them? I'd guess no. They obviously have similar, if not the same, intent. See, maybe part of the problem is that people just aren't communicating enough... > The GNU's "free" is more of an ensuring that your software remains free, > with the trade-off of the software in total being a little less free that > way :-) Try telling *them* that: "If you develop a new program, and you want it to be of the *greatest* possible use to the public, the *best* way to achieve this is to make it free software which everyone can redistribute and change under *these* terms." (emphasis mine) > > And if anyone would like to toss in their 2c about the FSF, I will be > > *quite* happy to moan and gripe and carp further upon the subject of how > > utterly silly I think the GPL is... :) > I always wonder... how many people that find the GPL utterly silly have > ever actually had a _problem_ with it? The QPL-LGPL example above is a > sort of stupid thing, but how often does something like that happen? Well, FreeBSD has it's problems as well, unable to distribute gmake et al as part of it's installation. Thank heavens for the ports collection :) > So please, go on with your "utterly silly" rant... but give me an example > where you actually had a problem. I'll jump to the silliest part: "Also add information on how to contact you by electronic and paper mail." Mmm hmm, right, paper mail. I see the author's home address on a *lot* of free software, I do I do :) Other stuff: "7. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License." Nope... law of the land takes precedence over license, regardless of what the license states, as I mentioned earlier. So when I see something like this I have to ask myself "who are they trying to fool?" "9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns." That's not a term nor a condition. It should be in the preamble... if it needs to be stated at all! > I personally like putting stuff under the GPL for sake of simplicity... > it is for what I want, i.e. free software that can be modified, etc. If > somebody wants to do something more with the software than the GPL allows, > they can send me an email asking me to dual-license the software and I > will, if I don't see any problem with it (ignoring for a moment the fact > that I don't write any useful software, anyway :-) Well, dual licenses are their own cup of stew. I have no idea how that's supposed to work, legally. Apparently nothing is stopping me from putting two (quite possibly mutually contradictory) licenses on a program, hell, why not put it under ALL the licenses listed at http://www.opensource.org/licenses/ , that would sure confuse the hell out of the lawyers wouldn't it :) I don't like the GPL because it's *not* simple (compared to something like BSD, which *is* simple and *doesn't* seem to beg for 'dual licensing'.) So, I like putting stuff under the *BSD* license for the sake of simplicity. It does "the same thing" as the GPL, but: - it's way shorter - it avoids the "you should have received a copy" problem entirely - it's written in legal language instead of the vague idealistic prattle found in the GPL and the Artistic license I mean, face it, if you hired a lawyer to come up with a license for your product, and what they gave you looked like the GPL or the Artistic license... would *you* pay the guy? :) Chris -- THE VIEWS EXPRESSED IN THE PRECEDING E-MAIL ARE SOLELY THE VIEWS OF CHRIS PRESSEY (WHO IS NOT A LAWYER) AND MAY OR MAY NOT REFLECT REALITY WHERE YOU COME FROM. YOUR MILEAGE MAY VARY. VOID WHERE PROHIBITED. POST NO BILLS. WATCH THIS SPACE. ------------------------------ Date: Wed, 04 Jul 2001 12:48:08 -0700 From: Russell Bornschlegel Subject: Re: intellectual property (was Re: various) Chris Pressey wrote: > > Russell Bornschlegel wrote: > > Chris Pressey wrote: > > > To me that's a strong argument that "damages" based on the "piracy" of > > > obsolete-but-still-copyrighted software would be very, very difficult to > > > obtain in any court of law where the judge has his or her eyes open... > > Legal fees at some point will vastly eclipse damages, and work just > > as well as a deterrent. Personally, I believe piracy of obsolete-but- > > still-copyrighted software is ethically Wrong; > > I'd rather leave ethics out of it; arguably, merely using a computer > would be "ethically wrong" to a luddite, right? :) Sure, so it's "ethically wrong" to force a Luddite to use a computer. Leaving ethics aside, I started going in circles here: Obsoleteware piracy violates copyright law. I believe in obeying "just" laws and in disobeying certian "unjust" laws. I consider portions of copyright law just, and other portions unjust (like the DMCA, which I think is dangerously restrictive and gives too much power to large holders of IP). However, I just realized how to get obsolete software in a reasonable way without violating copyright law. (It may violate EULAs, but I don't believe them to be legally binding for the most part.) Buy old software from people who have upgraded to new software. Take a quick trip to eBay. Borland makes their abandonware available for free, but I see on eBay a sealed-box copy of Turbo C++ 3.0 going for $2.25 -- that's a bargain for printed manuals. MS-DOS 6 runs around $10. If I had to get a kid set up with a system where he could learn to program, this might be a viable alternative to a free Unixoid OS -- particularly if I wanted to keep him off the internet ("Here's the RFCs. Let me know when you've written a TCP/IP stack."). If you need "modern" but not latest software, you're gonna pay a little more: MS Office is stil pretty pricey -- Office 95 goes for about $40, Office 97 upgrades around $50, Office 97 full versions $100. Lotus SmartSuite Millennium edition is under $20. Photoshop 4, $30. > And if anyone would like to toss in their 2c about the FSF, I will be > *quite* happy to moan and gripe and carp further upon the subject of how > utterly silly I think the GPL is... :) Well, it depends on what you mean by silly. Something shifted in my head a few months ago that made the GPL make sense to me. Every time GPL'd code is taken and reused, something of value has changed hands; the condition, the value given in return, is the expectation that code of value will be returned to the GPL community, eventually trickling down (in a statistical, if not literally traceable way) to the person who released the code in the first place. If you don't want to participate in this, don't -- but it doesn't make sense to me to bitch about it. It's a neighborhood, a community. It's going next door to borrow a cup of sugar from the neighbors. If you borrow from everyone else and don't let anyone else borrow from you, eventually the neighbors get pissed off and send nasty letters. The less restrictive open source licenses are merely neighbors who don't worry so much about bad apples in the neighborhood. (Chris, I just got your response to markus after I wrote all this, so I see that much of your problem with GPL is the phrasing, rather than the intent. > On the other hand, the law is pretty clear about this: if I download, I > dunno, C64 Ultimate Wizard, without paying for the license, the onus is > on EA - and, AFAICT, no one else - to find me and to sue me for > infringement. Hmm, can't find C64 Ultimate Wizard on eBay. :) > Or, take something even more interesting - var'aq. Like all "fanart," > it's technically in violation of Paramount's license, because it uses > elements from their copyrighted material (Klingons.) As usual, Paramount has to balance community goodwill against the money they could get by launching a competing product. They could try to suppress var'aq, but they would annoy some people and they couldn't show _any_ financial damage -- there's no market for var'aq. So yeah, they would only consider doing it as a deterrent to others. > But if you could > argue that var'aq is a parody of Star Trek, rather than a derivative > work, it would be excepted from infringement. Do you understand that > distinction? I'm not sure I do - especially considering the > sophisticated sense of humour that can appear in some parodies. AFAIK, it's not an objective distinction, and there are a lot of interacting factors. You can do a blatant parody of Star Trek, down to costume design and character names, as a single skit on Saturday Night Live (IIRC, the fact that both SNL and ST-TOS were on the same TV network helped make this work, but I think another network could get away with a one-off like that). Doing a feature film (Galaxy Quest) you have to disguise things a bit more. I'm not sure how I'd try to defend var'aq. It uses parts of the Klingon language; I don't know what exactly the relationship between Paramount and the KLI is. It's more a parody of programming languages (as with any esolang) than a parody of Star Trek. If you did a string search-n-replace on the var'aq page, you'd get exactly the same logical content, but it's significance would change drastically -- the giggle value of var'aq is tied intimately to Paramount's IP. That sounds like parody to me... > It > really freaks me out to think that this is what courts of the land are > busy doing, trying to decide if some story or programming language or > other is a joke or not :) No worse than some people's careers and reputations hanging on whether some court of the land decides is something is art or obscenity.[1] -RB [1] http://www.cbldf.org/casefiles/index.shtml -- Mike Diana's case is IMO the most significant of these. ------------------------------ Date: Wed, 4 Jul 2001 23:58:46 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: Re: intellectual property (was Re: various) On Wed, 4 Jul 2001, Chris Pressey wrote: > "markus.kliegl" wrote: > > On Tue, 3 Jul 2001, Chris Pressey wrote: > > > FSF seems the other way around; vehemently against anyone (except, > > > perhaps, themselves) making a buck off their software, even if it is > > > unrestricted in terms of development. > > Anyone can sell their software, as long as you give whomever you sell it > > to the same rights that you have, i.e. they can get the souce code and can > > do anything they want with the software under the terms and conditions of > > the GPL. > > Which is as it should be, I was mainly rolling my eyes at FSF's general > attitude, here. I fail to understand you here. > > > The idea is free software for free software in return. The GPL allows > > that... taking free software, making it proprietary and selling it is > > forbidden. > > ...without express consent of the copyright holder, I should add; if you > can get that, you can do anything you like with it. As with any > license. GPL'd software, even if used in a commercial and/or proprietary product, is still freely available to everyone. If a company wants to use free software, rather than reinventing the wheel, in a commercial and proprietary manner, why shouldn't they pay the copyright holder? Of course, the copyright holder of GNU software is the FSF, which would never dare think of selling a proprietary company software. > > > > I mean, pleeeez. "Why you should > > > not use the LGPL." FMH grumble grumble... Anyway, that's why I admire > > > the BSD license, and why I consider it "freer" than most commercial > > > software, and even GNU stuff, too. > > The problem with the FSF stuff is that they take it so strict that you > > can't use their free software with your free software (even if it fits > > the FSF's definition of "free"), unless it is "GPL-compatible". This is, > > for example, a problem with Ocaml. Ocaml is distributed under the QPL and > > can't use the GNU readline library (LGPL). > > See, that's just idiotic. Especially considering it's *L*GPL, that's > supposed to relieve problems like this. I don't know why the LGPL exists. It goes against the philosophical ideal of the FSF and still isn't all too helpful, but I'm not one to decide that. > > In fact I suspect it might not be a problem, but the legal text is so > thick that people are merely frightened that it might be a problem. It is a problem. > > The fact is a single piece of software is nigh useless alone, and no one > can scientifically define the difference between "mere aggregation" and > "integration." > > So why doesn't OCaml's author write the FSF and ask for one of the > "special exceptions" that they "sometimes grant." Surely theirs is not > against the "spirit of free software." The problem is one of the QPL being incompatible with the GPL and thus (I suppose) some sort of weird legal problems that could in some way crop up are trying to be avoided at all costs. It's a matter of not being dragged into the world of unsafe free software. > > Imagine, for example, if OCaml just went ahead and included gnureadline > anyway. Would FSF sue them? I'd guess no. They obviously have > similar, if not the same, intent. See, maybe part of the problem is > that people just aren't communicating enough... No, but the FSF would probably ask them to stop including it. I by all means don't know very much about law, but the way I know laws and have experienced the pedanticness of GNU software (recording *every* change, etc.), probably there's something about invalidness of the license if the FSF is aware and doesn't do anything. OTOH, I may be completely wrong. In the Objective Caml distribution, all files marked "Copyright INRIA" in the following directories and their sub-directories are distributed under the terms of the GNU Library General Public License: asmrun, byterun, config, maccaml, otherlibs, stdlib All files marked "Copyright INRIA" in the other directories and their sub-directories are distributed under the terms of the Q Public License version 1.0. So INRIA is actually quite willing to play along. > > > The GNU's "free" is more of an ensuring that your software remains free, > > with the trade-off of the software in total being a little less free that > > way :-) > > Try telling *them* that: > > "If you develop a new program, and you want it to be of the *greatest* > possible use to the public, the *best* way to achieve this is to make it > free software which everyone can redistribute and change under *these* > terms." (emphasis mine) No, not "of the greatest use to the public", but "of the greatest use to those that play along and *only* those that play along", which leads back to the philosophical ideal, which I approve of. > > > > And if anyone would like to toss in their 2c about the FSF, I will be > > > *quite* happy to moan and gripe and carp further upon the subject of how > > > utterly silly I think the GPL is... :) > > I always wonder... how many people that find the GPL utterly silly have > > ever actually had a _problem_ with it? The QPL-LGPL example above is a > > sort of stupid thing, but how often does something like that happen? > > Well, FreeBSD has it's problems as well, unable to distribute gmake et > al as part of it's installation. Thank heavens for the ports collection > :) Yes, that's really ridiculous. The GNU project is about having free software and fighting proprietary software in all manner. Unfortunately we do not live in that ideal world, so in order to still not serve proprietary software the GPL is very strict about anything that isn't GPL. So it requires the "play along"... if everyone would play along, all these problems wouldn't exist. > > > So please, go on with your "utterly silly" rant... but give me an example > > where you actually had a problem. I was asking for a problem you had with using GPL'd software, not what you find utterly silly about the text of the license. So I assume you only find it utterly silly to use it yourself. > > I'll jump to the silliest part: > > "Also add information on how to contact you by electronic and paper > mail." > > Mmm hmm, right, paper mail. I see the author's home address on a *lot* > of free software, I do I do :) That's in the 'Appendix: How to Apply These Terms to Your New Programs'. So it's more of a guideline. Putting your email address somewhere makes sense to me, but you don't seem to object to that, either. Paper mail... yes, seems a bit weird, but the GPLv2 dates back to June 1991, where not everybody had electronic mail. > > Other stuff: > > "7. If, as a consequence of a court judgment or allegation of patent > infringement or for any other reason (not limited to patent issues), > conditions are imposed on you (whether by court order, agreement or > otherwise) that contradict the conditions of this License, they do not > excuse you from the conditions of this License." > > Nope... law of the land takes precedence over license, regardless of > what the license states, as I mentioned earlier. So when I see > something like this I have to ask myself "who are they trying to fool?" I'm no lawyer, either, but if you're sure about that, go ahead and send the FSF an email asking them for clarification (or ask one of the many GPL-specialists around). > > "9. The Free Software Foundation may publish revised and/or new versions > of the General Public License from time to time. Such new versions will > be similar in spirit to the present version, but may differ in detail to > address new problems or concerns." > > That's not a term nor a condition. It should be in the preamble... if > it needs to be stated at all! Perhaps it is the wrong place, but taking the "...either version 2 of the License, or (at your option) any later version." found in most GPL'd files, it is something important to note. > > > I personally like putting stuff under the GPL for sake of simplicity... > > it is for what I want, i.e. free software that can be modified, etc. If > > somebody wants to do something more with the software than the GPL allows, > > they can send me an email asking me to dual-license the software and I > > will, if I don't see any problem with it (ignoring for a moment the fact > > that I don't write any useful software, anyway :-) > > Well, dual licenses are their own cup of stew. I have no idea how > that's supposed to work, legally. Apparently nothing is stopping me > from putting two (quite possibly mutually contradictory) licenses on a > program, hell, why not put it under ALL the licenses listed at > http://www.opensource.org/licenses/ , that would sure confuse the hell > out of the lawyers wouldn't it :) It works by giving people the option of using the software under either the conditions of license 1 or license 2, not a mix of both or anything like that. Sure, put a piece of software under all those licenses, I'm sure it wouldn't confuse lawyers, exactly because it's like the same piece of software distributed in 10 different versions, each with a different license. > > I don't like the GPL because it's *not* simple (compared to something > like BSD, which *is* simple and *doesn't* seem to beg for 'dual > licensing'.) By simplicity I mean that I don't have to worry about anything. The GPL has everything worked out for me, i.e. I want to distribute free community software and I don't want to deal with legal details. It guarantees me safety. > > So, I like putting stuff under the *BSD* license for the sake of > simplicity. It does "the same thing" as the GPL, but: > - it's way shorter > - it avoids the "you should have received a copy" problem entirely > - it's written in legal language instead of the vague idealistic prattle > found in the GPL and the Artistic license It doesn't do the same thing. Someone can take the software and relicense it... * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. Thus the software can be used in proprietary products... this is completely contradictory of the GNU ideal. > > I mean, face it, if you hired a lawyer to come up with a license for > your product, and what they gave you looked like the GPL or the Artistic > license... would *you* pay the guy? :) From a legal standpoint the GPL has been checked by several lawyers. The whole point of the FSF being founded (it's an official organization) was to protect software and sue any companies trying to steal any of it. In order for this to work, they needed an absolutely safe license and that's what the GPL is. Which license to use depends on your philosophy. Saying that one license is "utterly silly" (why, actually? You don't like the wording?) is utterly silly :-) > > Chris > > -- > THE VIEWS EXPRESSED IN THE PRECEDING E-MAIL ARE SOLELY THE VIEWS OF > CHRIS PRESSEY (WHO IS NOT A LAWYER) AND MAY OR MAY NOT REFLECT REALITY > WHERE YOU COME FROM. YOUR MILEAGE MAY VARY. VOID WHERE PROHIBITED. > POST NO BILLS. WATCH THIS SPACE. > > Markus ------------------------------ Date: Thu, 5 Jul 2001 00:09:31 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: [lang] Re: [BF] Emacs mode (Sorry I didn't respond earlier) On Sun, 1 Jul 2001, Rafal M. Sulejman wrote: > On Sun, 1 Jul 2001, markus.kliegl wrote: > > > I worked up a little emacs mode for brainfuck including an > > interpreter and odd debugging functionality. > > Heh, someone to implement this in vi? emacsers just won a battle > in this world domination war ;) (Use of BF against civil > population was forbidden by UN in 1999) I've only really used vi (more specifically vim) for a few months, so I'm not sure how much is possible (to my understanding, vim now supports scripting in perl, python and ruby). But here are some things to note: vim doesn't have an array data type, so it has to be simulated by a buffer (I read that somewhere, I have no idea how it works). [ and ] can be, depending on a[p], mapped to either '%' or 'l', which is basically what I used in bf.el (bf-matching-bracket). Of course this is terribly inefficient, but it makes BF seem like the perfect editor language sometimes :-) Anyone to implement a bf interpreter in TECO? > > > Please look at my comments for more details, though. They > > comprise more than a third of the file, after all :-) > > FSF will probably send you four six-packs of Virtual Beer :) > Now I have *real* reason to learn eeee-macs. I wouldn't dare show any emacs lisp hacker the code... > > BTW. -- where do you live, Markus? Ich bin ein M'jlabbacher (not > exactly ;) I live in Stahnsdorf, about 7 km from Berlin and 3 from Potsdam. > -- > . . . . | Rafal M. Sulejman > . # . . | > # . # . | "It's a small world, but I wouldn't want > . . # . | to have to paint it." --Steven Wright > ------------------------------ Date: Thu, 5 Jul 2001 00:24:38 +0200 (CEST) From: "Rafal M. Sulejman" Subject: [lang] Re: [BF] Emacs mode On Thu, 5 Jul 2001, markus.kliegl wrote: > > (Sorry I didn't respond earlier) > > On Sun, 1 Jul 2001, Rafal M. Sulejman wrote: > > > On Sun, 1 Jul 2001, markus.kliegl wrote: > > > > > I worked up a little emacs mode for brainfuck including an > > > interpreter and odd debugging functionality. > > > > Heh, someone to implement this in vi? emacsers just won a > > battle in this world domination war ;) (Use of BF against > > civil population was forbidden by UN in 1999) > > I've only really used vi (more specifically vim) for a few > months, so I'm not sure how much is possible (to my > understanding, vim now supports scripting in perl, python and > ruby). > But here are some things to note: > vim doesn't have an array data type, so it has to be simulated > by a buffer (I read that somewhere, I have no idea how it > works). [ and ] can be, depending on a[p], mapped to either '%' > or 'l', which is basically what I used in bf.el > (bf-matching-bracket). Of course this is terribly inefficient, > but it makes BF seem like the perfect editor language sometimes > :-) I already "created" something -- set of macros. Instead of array you can use actual buffer of vi :) Just split. In one window/buffer you have BF code in second -- something like: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Because you can map: :map + ^V^A :map - ^V^X you can easily increase or decrease "cell" next to cursor. Movements can be "implemented" using :map > W :map < B To add input/output you need to open next window/buffer and copy back and forth values in it (so far it worked with numeric,space-separated values. > > Anyone to implement a bf interpreter in TECO? > bf? TECO is powerful enough to emulate how WE think :) (someone said it can be done with 18(?) macros and 4000 Q-registers :) > > > Please look at my comments for more details, though. They > > > comprise more than a third of the file, after all :-) > > > > FSF will probably send you four six-packs of Virtual Beer :) > > Now I have *real* reason to learn eeee-macs. > > I wouldn't dare show any emacs lisp hacker the code... > THEY ALREADY KNOW. > > BTW. -- where do you live, Markus? Ich bin ein M'jlabbacher (not > > exactly ;) > > I live in Stahnsdorf, about 7 km from Berlin and 3 from Potsdam. Ossi with attitude ? ;) [I'm still more "eastern" than you]. -- . . . . | Rafal M. Sulejman /"\ . . . . | rms@spam-haters.poczta.onet.pl> \ / . # . . | ASCII Ribbon Campaign for HTML-free mail X # # . . | / \ ------------------------------ Date: Thu, 5 Jul 2001 01:49:13 +0200 (CEST) From: markus.kliegl@t-online.de (markus.kliegl) Subject: [lang] Re: [BF] Emacs mode On Thu, 5 Jul 2001, Rafal M. Sulejman wrote: > On Thu, 5 Jul 2001, markus.kliegl wrote: > > > > > (Sorry I didn't respond earlier) > > > > On Sun, 1 Jul 2001, Rafal M. Sulejman wrote: > > > > > On Sun, 1 Jul 2001, markus.kliegl wrote: > > > > > > > I worked up a little emacs mode for brainfuck including an > > > > interpreter and odd debugging functionality. > > > > > > Heh, someone to implement this in vi? emacsers just won a > > > battle in this world domination war ;) (Use of BF against > > > civil population was forbidden by UN in 1999) > > > > I've only really used vi (more specifically vim) for a few > > months, so I'm not sure how much is possible (to my > > understanding, vim now supports scripting in perl, python and > > ruby). > > But here are some things to note: > > vim doesn't have an array data type, so it has to be simulated > > by a buffer (I read that somewhere, I have no idea how it > > works). [ and ] can be, depending on a[p], mapped to either '%' > > or 'l', which is basically what I used in bf.el > > (bf-matching-bracket). Of course this is terribly inefficient, > > but it makes BF seem like the perfect editor language sometimes > > :-) > > I already "created" something -- set of macros. > > Instead of array you can use actual buffer of vi :) > > Just split. In one window/buffer you have BF code in second -- > something like: > > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > > Because you can map: > > :map + ^V^A > :map - ^V^X > > you can easily increase or decrease "cell" next to cursor. > > Movements can be "implemented" using > > :map > W > :map < B > > To add input/output you need to open next window/buffer and copy > back and forth values in it (so far it worked with > numeric,space-separated values. Wow! I really like that, great for debugging. Time to learn some more vi(m). > > > > > Anyone to implement a bf interpreter in TECO? > > > > bf? TECO is powerful enough to emulate how WE think :) > (someone said it can be done with 18(?) macros and 4000 > Q-registers :) I wasn't questioning if it _could_ be implemented in TECO :-) If hope I get around to reading more of the manual sometime. > > > > > Please look at my comments for more details, though. They > > > > comprise more than a third of the file, after all :-) > > > > > > FSF will probably send you four six-packs of Virtual Beer :) > > > Now I have *real* reason to learn eeee-macs. > > > > I wouldn't dare show any emacs lisp hacker the code... > > > THEY ALREADY KNOW. Your initials are RMS... hmm :-) > > > > > BTW. -- where do you live, Markus? Ich bin ein M'jlabbacher (not > > > exactly ;) > > > > I live in Stahnsdorf, about 7 km from Berlin and 3 from Potsdam. > > Ossi with attitude ? ;) [I'm still more "eastern" than you]. I was born in western Berlin :-) > > -- > . . . . | Rafal M. Sulejman /"\ > . . . . | rms@spam-haters.poczta.onet.pl> \ / > . # . . | ASCII Ribbon Campaign for HTML-free mail X > # # . . | / \ > Markus ------------------------------ Date: Thu, 5 Jul 2001 02:13:49 +0200 (CEST) From: "Rafal M. Sulejman" Subject: [lang] Re: [BF] Emacs mode On Thu, 5 Jul 2001, markus.kliegl wrote: > > > On Thu, 5 Jul 2001, Rafal M. Sulejman wrote: > Wow! I really like that, great for debugging. > Time to learn some more vi(m). > > > > > > > > > Anyone to implement a bf interpreter in TECO? > > > > > > > bf? TECO is powerful enough to emulate how WE think :) > > (someone said it can be done with 18(?) macros and 4000 > > Q-registers :) > > I wasn't questioning if it _could_ be implemented in TECO :-) > If hope I get around to reading more of the manual sometime. Smokey! ;) > > > > > Please look at my comments for more details, though. They > > > > > comprise more than a third of the file, after all :-) > > > > > > > > FSF will probably send you four six-packs of Virtual Beer :) > > > > Now I have *real* reason to learn eeee-macs. > > > > > > I wouldn't dare show any emacs lisp hacker the code... > > > > > THEY ALREADY KNOW. > > Your initials are RMS... hmm :-) Not RMS, but rms... unices are case sensitive. Just trying to make you paranoid ;) > > > > BTW. -- where do you live, Markus? Ich bin ein > > > > M'jlabbacher (not exactly ;) > > > > > > I live in Stahnsdorf, about 7 km from Berlin and 3 from > > > Potsdam. > > > > Ossi with attitude ? ;) [I'm still more "eastern" than you]. > > I was born in western Berlin :-) Oooops! %) -- . . . . | Rafal M. Sulejman /"\ . # . . | rms@spam-haters.poczta.onet.pl> \ / # . . . | ASCII Ribbon Campaign for HTML-free mail X # . . . | / \ ------------------------------ Date: Wed, 04 Jul 2001 20:43:41 -0500 From: Chris Pressey Subject: Re: intellectual property (was Re: various) Russell Bornschlegel wrote: > Sure, so it's "ethically wrong" to force a Luddite to use a computer. It's probably ethically wrong to force anyone to use a computer against their will. OK, um, never mind... :) > Leaving ethics aside, I started going in circles here: > > Obsoleteware piracy violates copyright law. I believe in obeying > "just" laws and in disobeying certian "unjust" laws. Well - yeah. There has to be some flexibility in law; people who treat it as sacrosanct just bug me. ("It must be wrong, 'cos it's against the law...") Laws are trying in some imperfect way to reflect ethics rather than the other way around... > I consider > portions of copyright law just, and other portions unjust (like > the DMCA, which I think is dangerously restrictive and gives too > much power to large holders of IP). I agree that recent "extensions" to copyright law have been contrived to protect companies like Disney from having to release 70-some-yr-old material into the public domain... to me this goes against the spirit of copyright, regardless of their nonsensical arguments of "oh, if we don't keep protecting intellectual property rights, artists will stop creating art"... bull! you just wanna keep making money off Mickey Mouse. :) > However, I just realized how to get obsolete software in a reasonable > way without violating copyright law. (It may violate EULAs, but I > don't believe them to be legally binding for the most part.) Not when "actual" laws precede them (hmm, I seem to be detecting a theme here.) > Buy old software from people who have upgraded to new software. Makes sense... > The less restrictive open source licenses are merely neighbors who > don't worry so much about bad apples in the neighborhood. Which strikes me as canny. Being paranoid about people "stealing the code" is falling prey to the same thought patterns that keep code closed in the first place! Many, many companies are VERY reluctant to use software they perceive as "too free" because it goes against their economic sensibilities. This is the case even when the free software is undeniably superior to what they're currently using. Because they've paid for what they're currently using, the idea of getting the same (or better!) software gratis makes them very nervous. Understandable? Maybe... I dunno. Anyway there does seem to be a huge rift between two equally zealous camps, and that division hurts everyone in a small way. > (Chris, I just got your response to markus after I wrote all this, > so I see that much of your problem with GPL is the phrasing, rather > than the intent. The intent is great. It's the implementation that bugs me (and the people who can't seem to tell the implementation apart from the intent...) Plus you have to realize I'm just bitching for the sake of bitching. It's fun, I don't expect anything to change, but maybe these thoughts can help spark some ideas or something. > > On the other hand, the law is pretty clear about this: if I download, I > > dunno, C64 Ultimate Wizard, without paying for the license, the onus is > > on EA - and, AFAICT, no one else - to find me and to sue me for > > infringement. > Hmm, can't find C64 Ultimate Wizard on eBay. :) http://www.lemon64.com/ Of course, be aware that you are (AFAIK) VIOLATING EA's COPYRIGHT if you choose to download it from there :) I dunno. In most ways, I don't believe there is such a thing as obsolescence... but I do believe that other people believe that there is. If they want to abandon old software without relaxing copyright on it, that's fine, but they have to realize that it becomes not much more than a formality if they don't have the resources to enforce it. And a formality isn't going to dissuade anyone who's determined to obtain it. > > Or, take something even more interesting - var'aq. Like all "fanart," > > it's technically in violation of Paramount's license, because it uses > > elements from their copyrighted material (Klingons.) > As usual, Paramount has to balance community goodwill against the > money they could get by launching a competing product. They could > try to suppress var'aq, but they would annoy some people and they > couldn't show _any_ financial damage -- there's no market for > var'aq. So yeah, they would only consider doing it as a deterrent > to others. Well, the "good neighbours" thing applies here, too, or at least it should. If Paramount explicitly tells people "fanart is OK," then they would theoretically be duly rewarded, and this would overall profit for them and their community. but, if they act paranoid and suppressive, then theoretically they should be identified as "bad blood." Of course - this sort of assumes that people will think for themselves, which, often as not isn't the case :) > > It > > really freaks me out to think that this is what courts of the land are > > busy doing, trying to decide if some story or programming language or > > other is a joke or not :) > No worse than some people's careers and reputations hanging on > whether some court of the land decides is something is art or > obscenity.[1] "The judge has just ruled that the legal proceedings themselves are both obscene and a joke, and, complaining of a headache, has called a recess..." :) Chris -- Add the next term to this series: 1. hunt and gather 2. farm 3. industry-commerce 4. ___________________ --R.A.W. ------------------------------ Date: Wed, 04 Jul 2001 20:59:46 -0700 From: Russell Bornschlegel Subject: re: intellectual property Crap. Replied to sender again. Chris Pressey wrote: > > Russell Bornschlegel wrote: > > Obsoleteware piracy violates copyright law. I believe in obeying > > "just" laws and in disobeying certian "unjust" laws. > > Well - yeah. There has to be some flexibility in law; people who treat > it as sacrosanct just bug me. ("It must be wrong, 'cos it's against the > law...") Laws are trying in some imperfect way to reflect ethics rather > than the other way around... Sure. Obeying the law as a rule-of-thumb or default position is probably good. Obey if you can, protest if you want, willfully disregard if you must. Since I have trouble even understanding why the DMCA hasn't been shot down in the courts yet, I'll be disregarding that one... > I agree that recent "extensions" to copyright law have been contrived to > protect companies like Disney from having to release 70-some-yr-old > material into the public domain... to me this goes against the spirit of > copyright, regardless of their nonsensical arguments of "oh, if we don't > keep protecting intellectual property rights, artists will stop creating > art"... bull! you just wanna keep making money off Mickey Mouse. :) Note that all the digital copy protection schemes being devised have _no_ facility for expiring the copyright in 70 or 90 or 120 or 1357 years. > > The less restrictive open source licenses are merely neighbors who > > don't worry so much about bad apples in the neighborhood. > > Which strikes me as canny. Being paranoid about people "stealing the > code" is falling prey to the same thought patterns that keep code closed > in the first place! Then you can look at it as the FSF people wanting to have the control of source code somewhere other than the status quo, because more control for the status quo is less control for the FSF. Like any revolution, they have decided on some level that even if a utopia doesn't arise, they can be better off than what the status quo provides. I'm all for bloodless revolutions that aren't actually being forced upon anyone, so again, more power to 'em. Just... not _too much_ more power to them. ;) I'll take a quick leap out of IP law & philosophy into sociology here. Consider all of human society as an emergent feature of the interaction between xenophilia/neophilia and xenophobia/neophobia. We get bored and seek out the new; we become enriched by the new, want to hold and possess it, to defend it (or our ownership of it) against the outsiders. We leave our hometowns to meet exotic MOTASes[1], settle down with them, make friends with the neighbors in the new neighborhood, and freak out when the children rebel against the status quo. > Many, many companies are VERY reluctant to use software they perceive as > "too free" because it goes against their economic sensibilities. This > is the case even when the free software is undeniably superior to what > they're currently using. I wonder how many companies are using free software and don't realize it. Companies aren't monoliths any more than governments are. I can grab perl and use it and there's no need to notify my manager, let alone purchasing and accounting. Think anyone at Microsoft uses perl? > Anyway there does seem to be a huge rift between two equally zealous > camps, and that division hurts everyone in a small way. Yup. It's just so hard to rally the masses around the banner of pragmatism and moderation and compromise, though. At least, things are perceived that way. > > Hmm, can't find C64 Ultimate Wizard on eBay. :) > > http://www.lemon64.com/ > > Of course, be aware that you are (AFAIK) VIOLATING EA's COPYRIGHT if > you choose to download it from there :) I was trying to see if I could steer _you_ towards a more-legal method of obtaining the software, so your soul wouldn't be burdened with the heinous crime of abandonware piracy, my son. I've never even owned a C64. :) > > I dunno. In most ways, I don't believe there is such a thing as > obsolescence... but I do believe that other people believe that there > is. If they want to abandon old software without relaxing copyright on > it, that's fine, but they have to realize that it becomes not much more > than a formality if they don't have the resources to enforce it. And a > formality isn't going to dissuade anyone who's determined to obtain it. Yeah, I'll give you that. I would love to see more companies stating a common-sense IP policy. Example: Ani DiFranco records on her own family-run Righteous Babe label. Her CDs all come with the following statement: "Unauthorized reproduction, while sometimes necessary, is never as good as the real thing." This is just beautiful. It works just as well as "...is prohibited by law," but has the added feature of demonstrating some sort of connection with reality, of some respect for the person reading the warning. I was going to compose an Estarcion Media License for if/when I make music good enough to sell (hah), defining in excruciating detail what sort of reproduction I considered acceptable. Given all the possible analog and digital representations of music, it would have been absurdly long and dull. I may just use the Ani License. > Well, the "good neighbours" thing applies here, too, or at least it > should. If Paramount explicitly tells people "fanart is OK," then they > would theoretically be duly rewarded, and this would overall profit for > them and their community. Paramount's strategy seems to be simply to put out more and crappier Star Trek stuff than the fan community has the resources to produce, thereby collecting money from the fan community so they don't have any left to buy fanart. :) -R [1] Members Of The Appropriate Sex ------------------------------ Date: Thu, 5 Jul 2001 09:37:48 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: re: intellectual property [ Using "Reply to all" in Pine, I got To: Russel Bornschlegel CC: misc@esoteric.sange.fi I've just always been manually deleting the To: and moving the CC: there. ] On Wed, 4 Jul 2001, Russell Bornschlegel wrote: > > > Many, many companies are VERY reluctant to use software they perceive as > > "too free" because it goes against their economic sensibilities. This > > is the case even when the free software is undeniably superior to what > > they're currently using. > > I wonder how many companies are using free software and don't > realize it. Companies aren't monoliths any more than governments > are. I can grab perl and use it and there's no need to notify > my manager, let alone purchasing and accounting. Think anyone > at Microsoft uses perl? Companies are reluctant to use free software, because it usually isn't supported by an organization. "No warranty" is scary if there's a bug that won't let them do whatever they want to do after they've written 90 % of whatever they want to do. Or support... if something doesn't work, they don't know how to do something, there's no commercial support line available. Then there's of course the problem of the free software sometimes requiring its output to be free as well, which I believe is the case with GCC. (Bison and Flex I think allow people to do whatever they want with generated code.) Other than that, who knows what companies use what free software? The reason Perl and Linux are used commercially so much is that they are a) marketed quite well, i.e. are a "buzzword" in relation to eBusiness b) there are lots of books available c) they can get commercial support (from vendors, etc.) GNU wants Linux to be called GNU/Linux... how about 'Linux' in relation to eBuiseness uses and 'GNU/Linux' in relation to the open-source programmer/idealist/hobbyist community that surrounds GNU/Linux? > > -R > Markus -- Time to get rid of some quotes :-) 'An operating system is just a name you give the features you left out of your editor.' -- Per Abrahamsen in comp.emacs ------------------------------ Date: Thu, 5 Jul 2001 18:30:51 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: [lang] [BF] Common Repository There doesn't seem to be a BF repository. I know this has been suggested several times, but I was thinking in terms of having a Wiki. It would make it a lot easier for anyone to add their interpreter, compiler, text, bf program, link, whatever and anyone could post ideas. A simple FTP repository wouldn't be bad, either. Or maybe a Common Brainfuck Archive Network. There's plenty of possibilities... does someone have a server they can use for this? (Panu, can you set up something on esoteric.sange.fi?) Markus ------------------------------ Date: Thu, 05 Jul 2001 11:24:43 -0500 From: Chris Pressey Subject: Re: intellectual property (was Re: various) I'm not trying to start an oslic war here, so I'll try to skip the minor details and get to the heart of the matter... "markus.kliegl" wrote: > On Wed, 4 Jul 2001, Chris Pressey wrote: > > "markus.kliegl" wrote: [...] > > See, that's just idiotic. Especially considering it's *L*GPL, that's > > supposed to relieve problems like this. > I don't know why the LGPL exists. My understanding is that it exists because of the 'virulent' nature of the GPL. If the code in, say, gcc's crt1.o was GPL'ed, then every C program compiled with gcc would (apparently) suddenly become GPL'ed too. > It goes against the philosophical ideal > of the FSF and still isn't all too helpful, but I'm not one to decide > that. It's pretty silly that they simultaneously provide the LGPL and discourage it's use. If they're serious about their philosophy, then why not just refuse to make an "L" version of their license, and tell everyone who uses gcc to just "put up and shut up?" > > "If you develop a new program, and you want it to be of the *greatest* > > possible use to the public, the *best* way to achieve this is to make it > > free software which everyone can redistribute and change under *these* > No, not "of the greatest use to the public", but "of the greatest use to > those that play along and *only* those that play along", That's what I don't like. Don't confuse the two. That's no better than Microso~1. The point I'm trying to make here is that BSD is "freer" than GPL largely because it doesn't suffer under any such philosophical pretense that everyone feels the same way the FSF does. > which leads back > to the philosophical ideal, which I approve of. And I think it's silly in it's idealism. Open source is about licensing, not philosophy. Try searching www.fsf.org for "manifesto" (eight pages.) Try searching www.freebsd.org for "manifesto" (no documents.) > > > So please, go on with your "utterly silly" rant... but give me an example > > > where you actually had a problem. > I was asking for a problem you had with using GPL'd software, not what > you find utterly silly about the text of the license. I wasn't trying to say that I have a problem using GPL'd software, I hope I didn't give you that impression. I was trying to say that the license, moreso the philosophy, strikes me as a silly way to go about it. > > "Also add information on how to contact you by electronic and paper > > mail." > > Mmm hmm, right, paper mail. I see the author's home address on a *lot* > > of free software, I do I do :) > That's in the 'Appendix: How to Apply These Terms to Your New Programs'. > So it's more of a guideline. If you have a license, why do you need guidelines too? To me, it just runs the danger of people confusing the two. When you put legally-binding rules together with non-legally-binding suggestions, you're sending an inconsistent message. > > Well, dual licenses are their own cup of stew. I have no idea how > > that's supposed to work, legally. Apparently nothing is stopping me > > from putting two (quite possibly mutually contradictory) licenses on a > > program, hell, why not put it under ALL the licenses listed at > > http://www.opensource.org/licenses/ , that would sure confuse the hell > > out of the lawyers wouldn't it :) > It works by giving people the option of using the software under either > the conditions of license 1 or license 2, not a mix of both or anything > like that. Sure, put a piece of software under all those licenses, I'm > sure it wouldn't confuse lawyers, exactly because it's like the same piece > of software distributed in 10 different versions, each with a different > license. Oh... so if I choose to look at Perl as being GPL'ed, I can assume artistic control if I like? While that strikes me as probably true... I have the impression Mr. Wall doesn't look at it that way... > > So, I like putting stuff under the *BSD* license for the sake of > > simplicity. It does "the same thing" as the GPL, but: > > - it's way shorter > > - it avoids the "you should have received a copy" problem entirely > > - it's written in legal language instead of the vague idealistic prattle > > found in the GPL and the Artistic license > It doesn't do the same thing. Someone can take the software and relicense > it... [...] > Thus the software can be used in proprietary products... this is > completely contradictory of the GNU ideal. Yeah, but what's *wrong* with it? If someone makes a proprietary version of gcc, how does that hurt the "community"? > > I mean, face it, if you hired a lawyer to come up with a license for > > your product, and what they gave you looked like the GPL or the Artistic > > license... would *you* pay the guy? :) > >From a legal standpoint the GPL has been checked by several lawyers. The > whole point of the FSF being founded (it's an official organization) was > to protect software and sue any companies trying to steal any of it. Not unlike Microso~1. > Which license to use depends on your philosophy. Saying that one license > is "utterly silly" (why, actually? You don't like the wording?) is utterly > silly :-) I never claimed it wasn't :) To me it looks like the FSF is, well, at war with proprietary software (and by extension for-profit software) out of idealistic notions and the assumption that "if it's proprietary, it must necessarily be evil"... Whereas I don't see how that's the case. Chris -- This sentence will self-destruct in five... four... three... Tuesday... ------------------------------ Date: Thu, 05 Jul 2001 09:36:38 -0700 From: Russell Bornschlegel Subject: Re: intellectual property Markus Kliegl wrote: > > [ Using "Reply to all" in Pine, I got > To: Russel Bornschlegel > CC: misc@esoteric.sange.fi > > I've just always been manually deleting the To: and moving the CC: > there. ] I get the same result when I remember to reply-to-all; my trouble is that I'm trained to not hit reply-to-all by another mailing list that I'm involved in which is intended as a one-to-many communication. There, reply-to-all is the Wrong Thing, and other people's careless use of reply-to-all has started flamewars and embarrassed people. My problem, I know, but there you have it. > Then there's of course the problem of the free software > sometimes requiring its output to be free as well, which I believe is the > case with GCC. (Bison and Flex I think allow people to do whatever they > want with generated code.) I don't believe that's the case with gcc. I've used gcc for game development in the past, with support from Cygnus arranged through Sega of America. Believe me, there have been times when I've wished I could point at something like a licensing issue to avoid having to use it... :) Can anyone here cite a license that explicitly says that program output is covered by the same license as the program itself? > GNU wants Linux to be called GNU/Linux... how about 'Linux' in relation to > eBuiseness uses and 'GNU/Linux' in relation to the open-source > programmer/idealist/hobbyist community that surrounds GNU/Linux? Okay, this is one of _my_ hot button issues. Does the GPL address naming? Is there anything in the license which prevents one from messing with some internal details of the Linux kernel and re- releasing the product as "Foonix", "Narcissux", or "I Am Jack's Operating System"? If not, then GNU can release their own damn distribution, and call it GNU/Linux or LiGNUx or GNUX -- or GNU/Hurd for that matter. But if it isn't in the license, don't even try to tell me what I can call _my_ product. -R ------------------------------ Date: Thu, 5 Jul 2001 10:39:48 -0600 (MDT) From: Ben Olmstead Subject: Re: GEB (was Re: number-o-mania...) On Sat, 23 Jun 2001, Russell Bornschlegel wrote: Sorry for how late this is; I've been on an actual vacation for the past week and a half. > Chris Pressey wrote: > > > I didn't really like it enough to read the whole thing... kind of geeky > > for my taste. It confirmed what I didn't really like about his style. > > Too much awe, not enough whimsey. > > I don't believe I've actually read the whole book, but I have read > bits and pieces of it over and over and over. Certainly it's not > flawless, but it did introduce me to a lot of then-new-to-me ideas > when I started trying to read it years-n-years ago. Ironically, I just picked it up and read it during my vacation (well, I read about the first third, skimmed the second third, and skipped the last third). My impression of the book was that the first half was trying to explore/explain the ramifications of self-reference, while the second half was simply trying to apply the first half to the question of 'Is true AI possible?'. I also was talking to a friend of mine who is basically of geek of anatomy and geology. (As opposed to a geek of computer science, like me.) She had a very difficult time getting through even the first few pages, just because she had never really thought about self-reference; I suspect these kinds of thoughts are a mathematics/comp sci thing, and not that many people overall have actually thought about it. Ben Olmstead bem@mad.scientist.com Inertia: Simple. Addictive. Free. http://www.mines.edu/students/b/bolmstea/games/inertia/ ------------------------------ Date: Thu, 05 Jul 2001 11:38:42 -0500 From: Chris Pressey Subject: Re: intellectual property Markus Kliegl wrote: > On Wed, 4 Jul 2001, Russell Bornschlegel wrote: > > > Many, many companies are VERY reluctant to use software they perceive as > > > "too free" because it goes against their economic sensibilities. This > > > is the case even when the free software is undeniably superior to what > > > they're currently using. > > I wonder how many companies are using free software and don't > > realize it. Companies aren't monoliths any more than governments > > are. I can grab perl and use it and there's no need to notify > > my manager, let alone purchasing and accounting. Think anyone > > at Microsoft uses perl? Well, certainly yes in the sense that MS is (last i heard) trying to assimilate "every known programming language" into it's VM/C# thingy. But I assume you mean beyond that, and the answer is probably still yes, 'cos it's a really useful rapid-development-gluing-scripting language. But I was referring less to software companies and moreso to companies that have a significant technological investment: insurance, telephony, and such. > Companies are reluctant to use free software, because it usually isn't > supported by an organization. "No warranty" is scary if there's a bug that > won't let them do whatever they want to do after they've written 90 % of > whatever they want to do. I have yet to see a commercial software product that comes with a warranty. > Or support... if something doesn't work, they > don't know how to do something, there's no commercial support line > available. Even when there is... Gerson can attest to how useful it is, I'm sure :) Seriously, a support line is largely a security blanket. One of those corporate things that makes other corporate things feel comfortable... just because most corporations don't recognize (say) the FreeBSD-questions mailing list as a support line, doesn't mean it doesn't provide as good or better support than some toll-free number. > Then there's of course the problem of the free software > sometimes requiring its output to be free as well, which I believe is the > case with GCC. Not when the c-libraries are LGPL'ed. Which, last I checked, they are. IIRC that was the reason for making the LGPL in the first place. > Other than that, who knows what companies use what free software? I can only speak for companies I've seen locally that don't. Or at least, their attitude of reluctance. > GNU wants Linux to be called GNU/Linux... how about 'Linux' in relation to > eBuiseness uses and 'GNU/Linux' in relation to the open-source > programmer/idealist/hobbyist community that surrounds GNU/Linux? How about whatever Linus Torvalds damn well pleases to call it? He's the guy who owns it, after all. Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Thu, 05 Jul 2001 18:25:14 +0100 From: Keith Gaughan Subject: Re: intellectual property (was Re: various) At 11:24 05/07/2001 -0500, Chris Pressey wrote: >The point I'm trying to make here is that BSD is "freer" than GPL >largely because it doesn't suffer under any such philosophical pretense >that everyone feels the same way the FSF does. The BSD licence is freer, no doubt. It's also a licence for angels, the same way as communism is a political system for angels. Both things have a flaw -- the don't fit in with human nature and require people to act like angels for them to work. The GPL and LGPL, for all their flaws, are more pragmatic. They don't presume that people are angels. This is why, even though dictatorships are the most efficent system of government, we live in democracies -- you can't be assured that your dictator will be benevolent (an angel) or malevolent (a devil). The GPL's pragmatism comes from how it gives up some of the freedom to make sure it stays free. We could live under efficient dictatorships, but we choose not to because in an inefficent democracy you're somewhat assured that you won't be able to be pushed around by a despot. The BSD licence is a licence of a perfect world. This is not a perfect world. That's the advantage of the GPL. Cut away all the philosophical pretense and that's what you're left with. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 05 Jul 2001 18:44:20 +0100 From: Keith Gaughan Subject: re: intellectual property At 09:37 05/07/2001 +0200, Markus Kliegl wrote: >Companies are reluctant to use free software, because it usually isn't >supported by an organization. "No warranty" is scary if there's a bug that >won't let them do whatever they want to do after they've written 90 % of >whatever they want to do. It's worth noting that virtually all software is provided without warranty. It's not just something which occurs in Free Software. >Then there's of course the problem of the free software >sometimes requiring its output to be free as well, which I believe is the >case with GCC. (Bison and Flex I think allow people to do whatever they >want with generated code.) That's why there's licences like the LGPL (so that the likes of Bison, which needs to have specific libraries attached to use it's output) exist. As a sidenote, the output of a program is never covered by the licence of the generating software, that's why you could write as much proprietary software as you wanted, using gcc as your compiler. If this was not the case, every MSWord file on the planet would be covered by a MS EULA -- scary! Software is a process, not it's input or output. >Other than that, who knows what companies use what free software? > >The reason Perl and Linux are used commercially so much is that they are >a) marketed quite well, i.e. are a "buzzword" in relation to eBusiness >b) there are lots of books available >c) they can get commercial support (from vendors, etc.) Don't forget Apache! >GNU wants Linux to be called GNU/Linux... how about 'Linux' in relation to >eBuiseness uses and 'GNU/Linux' in relation to the open-source >programmer/idealist/hobbyist community that surrounds GNU/Linux? Hmmm... ugly. The kernel is called Linux, the GNU part is just bundled software. There's much more than the GPL'd software in there, remember. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 05 Jul 2001 19:18:00 +0100 From: Keith Gaughan Subject: Re: intellectual property (was Re: various) I think this should have gone to the list... At 11:02 05/07/2001 -0700, Brian Connors wrote: >--- Keith Gaughan wrote: > > At 11:24 05/07/2001 -0500, Chris Pressey wrote: > > > > >The point I'm trying to make here is that BSD is > > "freer" than GPL > > >largely because it doesn't suffer under any such > > philosophical pretense > > >that everyone feels the same way the FSF does. > > > > The BSD licence is freer, no doubt. It's also a > > licence for angels, the > > same way as communism is a political system for > > angels. Both things have a > > flaw -- the don't fit in with human nature and > > require people to act like > > angels for them to work. > >I don't see the BSD license as having been concieved >with any particular ideals at all, actually. The BSD licence predates the GPL, the FSF and the OSI. It's essentially a college licence. The need for people to be angels only arose when the copyright clause was removed. By itself, code released under the the BSD licence (note that I like the BSD licence and wish I could always use it) isn't protected from individuals and groups who wish to take the code as their own. This is it's flaw. The GPL seeks to correct this flaw by including terms which oblige those who change code covered by the GPL to release the source code of any changed versions to the public if these changed versions are distributed. This ensures that nobody can take code written under the GPL and make it proprietary. That's what makes it more pragmatic. >The way >it's written, it simply asks that credit be given >where credit is due, and chooses to leave well enough >alone otherwise. Anyone arguing BSD licensing from a >moral standpoint higher than that is missing the >point. I'm not arguing from a moral POV, I'm just saying that if one releases software under the BSD licence and expects it to remain Free, then one is presuming that people act like angels. > > The GPL and LGPL, for all their flaws, are more > > pragmatic. They don't > > presume that people are angels. This is why, even > > though dictatorships are > > the most efficent system of government, we live in > > democracies -- you can't > > be assured that your dictator will be benevolent (an > > angel) or malevolent > > (a devil). > >This is entirely true -- the GPL is an entirely >different animal from the BSD license for precisely >that reason. > > > The GPL's pragmatism comes from how it gives up some > > of the freedom to make > > sure it stays free. We could live under efficient > > dictatorships, but we > > choose not to because in an inefficent democracy > > you're somewhat assured > > that you won't be able to be pushed around by a > > despot. > > > > The BSD licence is a licence of a perfect world. > > This is not a perfect > > world. That's the advantage of the GPL. Cut away all > > the philosophical > > pretense and that's what you're left with. > >I personally favor MozPL/NPL because MozPL/NPL gives >me privileges as the creator of the work that maybe I >don't want extended to others using it (in other >words, you're free to use my toys in any way you want >as long as you don't lock things up, but I'm not bound >by the same rules nyah nyah nyah). (I consider it a >CYA policy in case I ever have to pull some code out >of, say, var'aq for a consulting project.) I am not, >however, inherently averse to the GPL -- I'm more than >happy to slap GPL on anything I feel is too important >to be locked up by anybody (not that I've ever written >anything quite that earthshattering). K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 05 Jul 2001 19:25:16 +0100 From: Keith Gaughan Subject: IRC Channel It just popped into my head -- wouldn't it be great if we had an IRC channel... K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ From: Steve Mosher Subject: Re: IRC Channel Date: Thu, 5 Jul 2001 15:38:38 -0300 On Thu, 05 Jul 2001, Keith Gaughan pressed some keys and this came out: > It just popped into my head -- wouldn't it be great if we had an IRC channel... > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! That would be neat. It would be even neater if it were on slashnet, 'cause that's the IRC network I use (because #kuro5hin is there.) ;) -- Steve Mosher, Mad Scientist Be brave, all must face their soul, but keep the mirror close to the youth. Sometimes, the other becomes insane, you must keep the clarity close to the friend of strength. ------------------------------ From: Gerson.Kurz@t-online.de (Gerson Kurz) Subject: Google 4 language HAXORS Date: Thu, 5 Jul 2001 20:51:07 +0100 Try out the language box on the google preferences page (http://www.google.com/preferences). Includes the languages - Hacker - Bork ! Bork ! Bork ! - Elmer Fudd - Tagalog (I think Intercal knows this one, too) ------------------------------ From: "Chris Pressey" Subject: Re: intellectual property (was Re: various) Date: Thu, 5 Jul 2001 13:53:38 -0500 Keith Gaughan wrote: >The BSD licence predates the GPL, the FSF and the OSI. It's essentially a >college licence. The need for people to be angels only arose when the >copyright clause was removed. I'm not sure what you mean by that. All BSD-licensed software (in fact all O.S.S.) is copyrighted. The only thing rescinded about the BSD license was the advertising clause. >By itself, code released under the the BSD >licence (note that I like the BSD licence and wish I could always use it) >isn't protected from individuals and groups who wish to take the code as >their own. This is it's flaw. How is this a flaw? >I'm not arguing from a moral POV, I'm just saying that if one releases >software under the BSD licence and expects it to remain Free, If someone copies a BSD-licensed product, relicenses it, and sells it as proprietary software, that doesn't affect the original product. The original product remains free. >then one is >presuming that people act like angels. Just curious... if you're not talking about morals, then why the Christian terminology? Chris ------------------------------ Date: Thu, 05 Jul 2001 20:33:40 +0100 From: Keith Gaughan Subject: Re: intellectual property (was Re: various) At 01:53 05/07/2001 -0500, Chris Pressey wrote: >Keith Gaughan wrote: > >The BSD licence predates the GPL, the FSF and the OSI. It's essentially a > >college licence. The need for people to be angels only arose when the > >copyright clause was removed. > >I'm not sure what you mean by that. All BSD-licensed software (in fact all >O.S.S.) is copyrighted. The only thing rescinded about the BSD license was >the advertising clause. Um, that's sort of what I meant. I said that because the original licence stated that that Berkeley had copyright on the software. > >By itself, code released under the the BSD > >licence (note that I like the BSD licence and wish I could always use it) > >isn't protected from individuals and groups who wish to take the code as > >their own. This is it's flaw. > >How is this a flaw? It's a flaw if you don't want the code to be made proprietary. > >I'm not arguing from a moral POV, I'm just saying that if one releases > >software under the BSD licence and expects it to remain Free, > >If someone copies a BSD-licensed product, relicenses it, and sells it as >proprietary software, that doesn't affect the original product. The >original product remains free. But the original product doesn't benefit from this and it also introduces a fork, which isn't healthy. The Free Software can't avail from any improvements in the proprietary software, but the proprietary software can sponge off improvements to the Free one. > >then one is > >presuming that people act like angels. > >Just curious... if you're not talking about morals, then why the Christian >terminology? I'm not talking about morality at all. I'm using the word 'angel' to refer to a behaviour pattern that people just don't have--being nice all the time & not taking advantage of others. We've all been stung by this enough to know that 'angels' like this don't exist. The supposed Judeochristian/Islamic/Blah-de-blah terminology was used because it has other meanings besides denoting a lesser kind of divine being. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 05 Jul 2001 20:49:45 +0100 From: Keith Gaughan Subject: Re: intellectual property (was Re: various) If you want a licence which is even more free than the BSD licence, take a look at the one Peter Hartley uses for his software: http://www.chaos.org.uk/~pdh/software/ K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 5 Jul 2001 23:24:26 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: Re: intellectual property On Thu, 5 Jul 2001, Russell Bornschlegel wrote: > Markus Kliegl wrote: > > > Then there's of course the problem of the free software > > sometimes requiring its output to be free as well, which I believe is the > > case with GCC. (Bison and Flex I think allow people to do whatever they > > want with generated code.) > > I don't believe that's the case with gcc. I've used gcc for game > development in the past, with support from Cygnus arranged through > Sega of America. Believe me, there have been times when I've > wished I could point at something like a licensing issue to avoid > having to use it... :) I think I confused something here, but I'm sure I remember having read something about bison explicitly having to allow unrestricted use of code generated by it. > > Can anyone here cite a license that explicitly says that program > output is covered by the same license as the program itself? From the GPL: 'The act of running the Program is not restricted, and the output from the Program is covered only if its contents constitute a work based on the Program (independent of having been made by running the Program). Whether that is true depends on what the Program does.' Which seems very ambiguous to me. > > > GNU wants Linux to be called GNU/Linux... how about 'Linux' in relation to > > eBuiseness uses and 'GNU/Linux' in relation to the open-source > > programmer/idealist/hobbyist community that surrounds GNU/Linux? > > Okay, this is one of _my_ hot button issues. Does the GPL address > naming? Is there anything in the license which prevents one from > messing with some internal details of the Linux kernel and re- > releasing the product as "Foonix", "Narcissux", or "I Am Jack's > Operating System"? If not, then GNU can release their own damn > distribution, and call it GNU/Linux or LiGNUx or GNUX -- or > GNU/Hurd for that matter. But if it isn't in the license, don't > even try to tell me what I can call _my_ product. I was trying to point out the differences between the two communities that surround Linux (the programmers/open-source idealists and the "be pragmatic and use this Linux software for our business" people). > > -R > Markus ------------------------------ Date: Thu, 5 Jul 2001 23:49:48 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: Re: intellectual property (was Re: various) On Thu, 5 Jul 2001, Chris Pressey wrote: > "markus.kliegl" wrote: > > On Wed, 4 Jul 2001, Chris Pressey wrote: > > > "markus.kliegl" wrote: [...] > > > See, that's just idiotic. Especially considering it's *L*GPL, that's > > > supposed to relieve problems like this. > > I don't know why the LGPL exists. > > My understanding is that it exists because of the 'virulent' nature of > the GPL. If the code in, say, gcc's crt1.o was GPL'ed, then every C > program compiled with gcc would (apparently) suddenly become GPL'ed too. I just saw this: http://www.gnu.org/copyleft/gpl-faq.html#TOCCanIUseGPLToolsForNF It explains the bison thing, too. > > > It goes against the philosophical ideal > > of the FSF and still isn't all too helpful, but I'm not one to decide > > that. > > It's pretty silly that they simultaneously provide the LGPL and > discourage it's use. If they're serious about their philosophy, then > why not just refuse to make an "L" version of their license, and tell > everyone who uses gcc to just "put up and shut up?" Agreed. If they're going to make compromises that hurt the ideal like that, why not get rid of all of those restrictions causing problems? > > > > "If you develop a new program, and you want it to be of the *greatest* > > > possible use to the public, the *best* way to achieve this is to make it > > > free software which everyone can redistribute and change under *these* > > No, not "of the greatest use to the public", but "of the greatest use to > > those that play along and *only* those that play along", > > That's what I don't like. Don't confuse the two. That's no better than > Microso~1. I think that would be Micros~1 :-) I can't see where you draw up the connection to Microsoft. > > The point I'm trying to make here is that BSD is "freer" than GPL > largely because it doesn't suffer under any such philosophical pretense > that everyone feels the same way the FSF does. The way I understand the BSD license, it's a sort of public-domain with the addition of the software being copyrighted and the requirement of mentioning that it's copyrighted. Sort of like: "Do what you want, but give me credit". > > > which leads back > > to the philosophical ideal, which I approve of. > > And I think it's silly in it's idealism. Open source is about > licensing, not philosophy. I look at it the other way around: Open-source is about philosophy concerning what software is like. The licenses are what we use to ensure our philosophy in relation to _our_ work, i.e. _our_ work, which we decide should follow this philosophy. Pragmatic vs. Pure BSD vs. GPL ML vs. Haskell Prolog vs. Mercury Common Lisp vs. Scheme (well, sort of) > > > > So please, go on with your "utterly silly" rant... but give me an example > > > > where you actually had a problem. > > I was asking for a problem you had with using GPL'd software, not what > > you find utterly silly about the text of the license. > > I wasn't trying to say that I have a problem using GPL'd software, I > hope I didn't give you that impression. I was trying to say that the > license, moreso the philosophy, strikes me as a silly way to go about > it. > > > > "Also add information on how to contact you by electronic and paper > > > mail." > > > Mmm hmm, right, paper mail. I see the author's home address on a *lot* > > > of free software, I do I do :) > > That's in the 'Appendix: How to Apply These Terms to Your New Programs'. > > So it's more of a guideline. > > If you have a license, why do you need guidelines too? To me, it just > runs the danger of people confusing the two. When you put > legally-binding rules together with non-legally-binding suggestions, > you're sending an inconsistent message. It's guidelines on what to do if you want to be absolutely sure your software is safe. > > > So, I like putting stuff under the *BSD* license for the sake of > > > simplicity. It does "the same thing" as the GPL, but: > > > - it's way shorter > > > - it avoids the "you should have received a copy" problem entirely > > > - it's written in legal language instead of the vague idealistic prattle > > > found in the GPL and the Artistic license > > It doesn't do the same thing. Someone can take the software and relicense > > it... [...] > > Thus the software can be used in proprietary products... this is > > completely contradictory of the GNU ideal. > > Yeah, but what's *wrong* with it? If someone makes a proprietary > version of gcc, how does that hurt the "community"? Making a verbatim proprietary copy of something wouldn't make much sense. The software will get modified and if it's proprietary the original software doesn't get anything in return for allowing you to use and modify, improve, etc. it. > > > > I mean, face it, if you hired a lawyer to come up with a license for > > > your product, and what they gave you looked like the GPL or the Artistic > > > license... would *you* pay the guy? :) > > >From a legal standpoint the GPL has been checked by several lawyers. The > > whole point of the FSF being founded (it's an official organization) was > > to protect software and sue any companies trying to steal any of it. > > Not unlike Microso~1. Let me rephrase: "to protect the software from losing the philosophy under which it was intended to be". > > Chris > > -- > This sentence will self-destruct in five... four... three... Tuesday... > > Markus ------------------------------ Date: Thu, 05 Jul 2001 23:54:23 +0100 From: Keith Gaughan Subject: Re: intellectual property (was Re: various) At 11:49 05/07/2001 +0200, Markus Kliegl wrote: >Pragmatic vs. Pure > >BSD vs. GPL My god, no! That's backwards! The BSD licence is the pure licence. The GPL has clauses in it to make sure the software stays Free. The BSD licence is the licence of idealists, but this kind of idealism belies a certain naivety, but only if you want your code to stay Free. On the other hand, the GPL is a *puritan* licence. It demands that the code covered by it stays free. It's pragmatic because it requires a certain loss of freedom (the freedom to make the code proprietary) in order to keep it free. Granted, the GPL is too complicated. It could be vastly simplified. My perfect licence would be one based on the BSD licence which required that changes be released back to the public in the form of source, just as the GPL does. I wouldn't use it for everything, but I'd use this for software which I'd usually put under the GPL. You know, the Nokia licence is quite a nice one. It's worth a look if you don't like the others. K. -- Keith Gaughan Software Developer, Digital Crew Ltd. Take off every '.sig'!! ------------------------------ Date: Thu, 5 Jul 2001 21:15:18 -0700 (MST) From: Jeff Johnston Subject: [lang] Re: [BF] Common Repository Yes I would really love this! I have lots of BF code just hanging out, some of it doesnt have proper credit tho :( I just surfed around one day and saved everything I could find It would also be great include interpreters and compilers too. Jeff On Thu, 5 Jul 2001, Markus Kliegl wrote: > > There doesn't seem to be a BF repository. I know this has been suggested > several times, but I was thinking in terms of having a Wiki. > > It would make it a lot easier for anyone to add their interpreter, > compiler, text, bf program, link, whatever and anyone could post ideas. > > A simple FTP repository wouldn't be bad, either. > > Or maybe a Common Brainfuck Archive Network. > > There's plenty of possibilities... does someone have a server they can use > for this? (Panu, can you set up something on esoteric.sange.fi?) > > Markus > > > > > ------------------------------ Date: Thu, 5 Jul 2001 21:31:39 -0700 (MST) From: Jeff Johnston Subject: Re: IRC Channel Yes! I recommend EFNet.. its a large group of servers, which means a better possiblity of getting on quickly. I'll start hanging out on EFNet #esolang in the evenings Some EFNet servers (more at www.efnet.org): irc.prison.net irc.ins.net.uk irc.solidstreaming.net irc.east.gblx.net irc.mcs.net irc.mindspring.com irc.plur.net irc.emory.edu irc.light.se Jeff (calamari) On Thu, 5 Jul 2001, Keith Gaughan wrote: > It just popped into my head -- wouldn't it be great if we had an IRC channel... > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! > > > ------------------------------ Date: Thu, 05 Jul 2001 21:57:27 -0700 From: Robert Subject: Re: IRC Channel I'd not recommend EFnet...it's very unstable, too many netsplits per day to count. The summer makes it even worse. I can brag to my friends that I've been connected to Emory for over 24 hours (that hasn't happened on any irc server on efnet for months now with most of us) IRCnet or undernet would be a better choice, since they are far more stable. IRCnet most likely won't be a choice since it doesn't really have American servers (I connect to a Japanese one) So I'd recommend Undernet...we have a network but it has like 20 users so I figure that wouldn't be a place you'd want to go to hahaha. *kicks efnet* packeters take all the fun out of it hehe. Robert (Most of those efnet servers won't just let anyone on.) ----- Original Message ----- From: "Jeff Johnston" To: "Keith Gaughan" Cc: Sent: Thursday, July 05, 2001 9:31 PM Subject: Re: IRC Channel > > Yes! > > I recommend EFNet.. its a large group of servers, which means a better > possiblity of getting on quickly. > > I'll start hanging out on EFNet #esolang in the evenings > > Some EFNet servers (more at www.efnet.org): > irc.prison.net > irc.ins.net.uk > irc.solidstreaming.net > irc.east.gblx.net > irc.mcs.net > irc.mindspring.com > irc.plur.net > irc.emory.edu > irc.light.se > > Jeff (calamari) > > > On Thu, 5 Jul 2001, Keith Gaughan wrote: > > > It just popped into my head -- wouldn't it be great if we had an IRC channel... > > > > K. > > > > -- > > Keith Gaughan > > Software Developer, Digital Crew Ltd. > > Take off every '.sig'!! > > > > > > > > > ------------------------------ Date: Thu, 5 Jul 2001 22:04:29 -0700 (MST) From: Jeff Johnston Subject: Re: IRC Channel Just been on undernet for a few months.. its just as bad as efnet, and the servers never let you on :) But yeah either one.. lets choose one and go with it. Jeff, now on undernet #esolang too, but no one's there! (irc.undernet.org) On Thu, 5 Jul 2001, Robert wrote: > I'd not recommend EFnet...it's very unstable, too many netsplits per day to > count. > The summer makes it even worse. I can brag to my friends that I've been > connected to Emory for over 24 hours (that hasn't happened on any irc server > on efnet for months now with most of us) > IRCnet or undernet would be a better choice, since they are far more stable. > IRCnet most likely won't be a choice since it doesn't really have American > servers (I connect to a Japanese one) > So I'd recommend Undernet...we have a network but it has like 20 users so I > figure that wouldn't be a place you'd want to go to hahaha. > *kicks efnet* packeters take all the fun out of it hehe. > Robert > > (Most of those efnet servers won't just let anyone on.) > ----- Original Message ----- > From: "Jeff Johnston" > To: "Keith Gaughan" > Cc: > Sent: Thursday, July 05, 2001 9:31 PM > Subject: Re: IRC Channel > > > > > > Yes! > > > > I recommend EFNet.. its a large group of servers, which means a better > > possiblity of getting on quickly. > > > > I'll start hanging out on EFNet #esolang in the evenings > > > > Some EFNet servers (more at www.efnet.org): > > irc.prison.net > > irc.ins.net.uk > > irc.solidstreaming.net > > irc.east.gblx.net > > irc.mcs.net > > irc.mindspring.com > > irc.plur.net > > irc.emory.edu > > irc.light.se > > > > Jeff (calamari) > > > > > > On Thu, 5 Jul 2001, Keith Gaughan wrote: > > > > > It just popped into my head -- wouldn't it be great if we had an IRC > channel... > > > > > > K. > > > > > > -- > > > Keith Gaughan > > > Software Developer, Digital Crew Ltd. > > > Take off every '.sig'!! > > > > > > > > > > > > > > > > > > ------------------------------ Date: Thu, 05 Jul 2001 22:23:41 -0700 From: Robert Subject: RE: IRC Channel That's why I like IRCnet hehe Only one problem with that network, you cannot /list on any server really. It has more clients then efnet though... Most irc networks have gone to hell, wonder if M$ and Aol could be up to it to promote their messenger clients Robert ------------------------------ Date: Fri, 06 Jul 2001 09:37:46 +0200 From: =?iso-8859-1?Q?Fr=E9d=E9ric?= van der Plancke Subject: Re: Google 4 language HAXORS Gerson Kurz wrote: > > Try out the language box on the google preferences page > (http://www.google.com/preferences). Includes the languages > > - Hacker > - Bork ! Bork ! Bork ! > - Elmer Fudd > - Tagalog (I think Intercal knows this one, too) and "Pig Latin" ... Note that Tagalog is a real human language spoken in the Philippines. (search for "tagalog language"). Frederic vdP ------------------------------ Date: Fri, 6 Jul 2001 11:32:00 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [list-meta] Bloody Listar! On Mon, 2 Jul 2001, Keith Gaughan wrote: > Ok, every time I've posted stuff to the list, it's gone back to the sender. > Why? Listar doesn't seem to include something as simple as a Reply-To: in > the email header when it forwards on mail. Is this just happening to me? > Could it be Eudora that's causing the problems? Panu, if it really is > Listar, could you please fix this? I have the feeling that I've explained this a hundred of times, but here it goes again: 1) The lists are chained, ie every posting to lang or sci gets forwarded to misc. Listar is totally unaware of this chaining: it's made by subscribing the misc list as recipient to the other lists. 2) Because I (we?) don't want lang & sci threads to accidentally move to misc (so that lang-only subscribers won't be able to participate), lang & sci insert a reply-to header, whereas misc does not (so that it won't overwrite the lang/sci -inserted reply-to header). 3) The side-effect of this configuration is that when some message gets posted to misc directly, it won't get a reply-to header at all. 4a) The solution to this is using the "reply to all" feature which every mailer I know of supports. The feature sends the reply to the original's recipients as well as its sender. 4b) Another solution would be to make still another list, say chat@esoteric.sange.fi, and require people to send their misc postings via that list. But it would probably just add to the overall confusion. 4c) The ultimate solution would be to check if Listar supports list chaining internally, or some other feature, like reply-to appending instead of overwriting. However, Listar is very badly documented, though good piece of software, so I'm not sure if I'll ever find how to make those settings even if they exist. Besides, I'm very busy in my personal life just now. I hope this explains the matter a bit; if it does not, feel free to ask more. Panu ------------------------------ Date: Fri, 6 Jul 2001 11:38:48 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [SPAM?] Re: Bloody Listar! On Mon, 2 Jul 2001, Jeff Johnston wrote: > I'm using unix Pine.. and from what I can see, basically what is going on > is the To: and Cc: are switched around the wrong way. To: should be > misc@esoteric.sange.fi and Cc: should be whoever we are repling to. That True. If I can find a way to swap the headers in Listar (it's the mail clients that put misc@esoteric into Cc:), I'll do that. But definitely not today. Panu ------------------------------ Date: Fri, 6 Jul 2001 11:47:58 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [list-meta] various On Tue, 3 Jul 2001, Chris Pressey wrote: > But... the strange thing is... me, being the poster of one of those e-mails, > should have received the single-replies that some people sent.... but, I Um... what do you mean? > didn't. And I have no idea what "misc-bounce" is all about. So I don't > have the whole story, here... It's Listar's way to try to ensure that the error reports MTA's send on expired / bad addresses won't get posted to the list. Panu ------------------------------ Date: Fri, 6 Jul 2001 11:52:25 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [www] Re: various On Tue, 3 Jul 2001, Russell Bornschlegel wrote: > Chris Pressey wrote: > > I've said it before, I'll say it again: web pages are *documents*, and > > documents should *not* crash. The idea of embedding "scripting" languages > > and such in documents is just... shall we say conceptually flawed, perhaps > > even doomed but no one has noticed yet :) > > It may be conceptually flawed, but in the same way as the idea that > "normal" people should be expected to use computers, it's an idea > that's here to stay. You could have fought this in 1993, but not > in 2001. I'd say "normal" people should be expected to learn to program, at least in a deterministic language. Am I conceptually flawed? > Some web pages are documents. Some are active, dynamic environments. > I will meet you halfway on this and say that there should be a > clearer distinction between the two -- different protocols or > something (dhttp: for dynamic), but it's just pissing into the tide > [sic] to try and complain about this now. How very true. http is one of the worst protocols to be used as a base for interactive environments. I'm looking forward for something chat-type (IRCD kind of protocols, but meant for all kinds of interaction, not only talk) to appear as the base of dynamic services. Panu ------------------------------ Date: Fri, 06 Jul 2001 11:48:31 -0500 From: Chris Pressey Subject: Re: [list-meta] various Panu A Kalliokoski wrote: > On Tue, 3 Jul 2001, Chris Pressey wrote: > > But... the strange thing is... me, being the poster of one of those e-mails, > > should have received the single-replies that some people sent.... but, I > Um... what do you mean? Well, if I post a message, and someone else replies to it without Replying to All like they probably should, I get their message but the list doesn't. But, there seemed to be some replies to one of my posts, that I didn't get, and that didn't appear on the list either, AFAICT. But, I could have been mistaken. Chris -- "Security holes can't show up in features that don't exist." -- DJB ------------------------------ Date: Fri, 06 Jul 2001 12:06:08 -0500 From: Chris Pressey Subject: Re: intellectual property (was Re: various) Keith Gaughan wrote: > At 01:53 05/07/2001 -0500, Chris Pressey wrote: > >Keith Gaughan wrote: > > >The BSD licence predates the GPL, the FSF and the OSI. It's essentially a > > >college licence. The need for people to be angels only arose when the > > >copyright clause was removed. > >I'm not sure what you mean by that. All BSD-licensed software (in fact all > >O.S.S.) is copyrighted. The only thing rescinded about the BSD license was > >the advertising clause. > Um, that's sort of what I meant. I said that because the original licence > stated that that Berkeley had copyright on the software. Well - that's only because Berkeley was the party using the license. Just like the original GPL stated that the FSF had copyright on their software, right? > > >By itself, code released under the the BSD > > >licence (note that I like the BSD licence and wish I could always use it) > > >isn't protected from individuals and groups who wish to take the code as > > >their own. This is it's flaw. > >How is this a flaw? > It's a flaw if you don't want the code to be made proprietary. Yeah, but that's obviously not a requirement of open-sourcing, or BSD, MIT et al wouldn't qualify. > > >I'm not arguing from a moral POV, I'm just saying that if one releases > > >software under the BSD licence and expects it to remain Free, > >If someone copies a BSD-licensed product, relicenses it, and sells it as > >proprietary software, that doesn't affect the original product. The > >original product remains free. > But the original product doesn't benefit from this Hypothetical scenario: Micros~1 takes FreeBSD, makes it proprietary, calls it DearMSD, and sells it. The result: people who are exposed to DearMSD (which, naturally, would be advertised heavily by Redmond) learn of it's history, and they learn they can get essentially the same product for free. The outcome: the original product benefits from what is essentially free advertising. > and it also introduces a fork, which isn't healthy. GPL doesn't prohibit forks either. I'd also be more apt to believe this bit of "conventional wisdom" ("forks aren't healthy") if someone could point out a project which was actually hurt by forking. For that matter, I'd be more apt to believe that BSD is an 'unsafe' license if someone could refer me to a project which was hurt by someone using BSD'ed software to make proprietary software. We can talk theory all we like, but where's the evidence? > The Free Software can't avail from any > improvements in the proprietary software, but the proprietary software can > sponge off improvements to the Free one. Hmm... under that logic, every user of GPL'ed software has an *obligation* to give something back? If that's the case, then a friend of mine, who uses RedHat at work but doesn't know how to program, had better start learning Perl or something, or she's "sponging" off Linus and the OS community at large. > > >then one is > > >presuming that people act like angels. > >Just curious... if you're not talking about morals, then why the Christian > >terminology? > I'm not talking about morality at all. I'm using the word 'angel' to refer > to a behaviour pattern that people just don't have--being nice all the time > & not taking advantage of others. Oh. In Canada, we have the word 'selfless' to describe that sort of behaviour. > We've all been stung by this enough to > know that 'angels' like this don't exist. Heh... speak for yourself :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Fri, 06 Jul 2001 12:12:47 -0500 From: Chris Pressey Subject: Re: intellectual property (was Re: various) Keith Gaughan wrote: > > If you want a licence which is even more free than the BSD licence, take a > look at the one Peter Hartley uses for his software: > http://www.chaos.org.uk/~pdh/software/ He could have saved himself a lot of typing by resorting to this simple eight-word phrase: "These works are part of the public domain." Or, if you want a license which is even more Erisian than Hartley's: http://www.catseye.mb.ca/gwadfc/epl.cgi Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Fri, 6 Jul 2001 10:19:57 -0700 (MST) From: Jeff Johnston Subject: Re: intellectual property (was Re: various) > MS-DOS 6 runs > around $10. If I had to get a kid set up with a system where he > could learn to program, this might be a viable alternative to > a free Unixoid OS -- particularly if I wanted to keep him off the > internet ("Here's the RFCs. Let me know when you've written a > TCP/IP stack."). There are several free TCP/IP stacks for MS-DOS.. and PPPD seems to work pretty good. I have my dad set up on it (and MS-DOS 6.22) because he refuses to learn Windows and hates the mouse :) Of course its really handy to have a shell account for that. Jeff ------------------------------ Date: Fri, 06 Jul 2001 10:25:15 -0700 From: Russell Bornschlegel Subject: Re: [www] Re: various Panu A Kalliokoski wrote: > > On Tue, 3 Jul 2001, Russell Bornschlegel wrote: > > It may be conceptually flawed, but in the same way as the idea that > > "normal" people should be expected to use computers, it's an idea > > that's here to stay. You could have fought this in 1993, but not > > in 2001. > > I'd say "normal" people should be expected to learn to program, at least > in a deterministic language. Am I conceptually flawed? Depends on what you mean by "expect". I see lots of people who are endlessly frustrated with computers, who spend hours and hours for very little apparent benefit. If people are comfortable with typewriters and filing cabinets, I say let 'em use typewriters and filing cabinets. (Hey, when did the word "file" in reference to a collection of bits attached to a computer come into use?) As for programming, I've been screening people who claim to be programmers who can't seem to learn to program, despite what expectations you or I might have. :) -RB ------------------------------ Date: Fri, 06 Jul 2001 10:31:24 -0700 From: Russell Bornschlegel Subject: Re: [list-meta] various Chris Pressey wrote: > "Security holes can't show up in features that don't exist." -- DJB This puts me in mind of the brilliant piece of programming advice I got from the head of the SCSI system software company that I fled to get into the game industry. Really, it's a corollary of St-Exupery's statement on design perfection: "If there's no code there, it's really easy to debug." -- Clint Ballard -RB ------------------------------ Date: Fri, 06 Jul 2001 10:35:51 -0700 From: Russell Bornschlegel Subject: Re: intellectual property (was Re: various) Aaargh. Remembered reply-all two outta three times. Jeff Johnston wrote: > > > MS-DOS 6 runs > > around $10. If I had to get a kid set up with a system where he > > could learn to program, this might be a viable alternative to > > a free Unixoid OS -- particularly if I wanted to keep him off the > > internet ("Here's the RFCs. Let me know when you've written a > > TCP/IP stack."). > > There are several free TCP/IP stacks for MS-DOS.. and PPPD seems to work > pretty good. I have my dad set up on it (and MS-DOS 6.22) because he > refuses to learn Windows and hates the mouse :) > > Of course its really handy to have a shell account for that. Yes, but the goal was to keep the kid off the internet, not to get him on. Besides, where do you _get_ a TCP/IP stack if you don't _have_ a TCP/IP stack? :) -RB ------------------------------ Date: Fri, 06 Jul 2001 12:34:47 -0500 From: Chris Pressey Subject: Re: [www] Re: various Russell Bornschlegel wrote: > Panu A Kalliokoski wrote: > > On Tue, 3 Jul 2001, Russell Bornschlegel wrote: > > > It may be conceptually flawed, but in the same way as the idea that > > > "normal" people should be expected to use computers, it's an idea > > > that's here to stay. You could have fought this in 1993, but not > > > in 2001. > > I'd say "normal" people should be expected to learn to program, at least > > in a deterministic language. Am I conceptually flawed? > Depends on what you mean by "expect". I see lots of people who are > endlessly frustrated with computers, who spend hours and hours for > very little apparent benefit. If people are comfortable with > typewriters and filing cabinets, I say let 'em use typewriters and > filing cabinets. Indeed. Obsolescence doesn't really exist. Typewriters still *work*... It's not "good for the economy," though. It's really amazing how pie-in-the-sky a lot of this technology is. It's not crucial, just "nifty." > (Hey, when did the word "file" in reference to a collection of bits > attached to a computer come into use?) May 3, 1967, at 4:14 pm. :) > As for programming, I've been screening people who claim to be > programmers who can't seem to learn to program, despite what > expectations you or I might have. :) I'd like to refine Panu's proposal, then, and suggest that "normal" people should be expected to program - IF they want to call themselves "programmers". There's a world of difference between programming and software engineering, too. It doesn't take a genius to figure out what this is supposed to do: 10 PRINT "BIG BROTHER 2 SUCKS ROCKS" 20 GOTO 10 On the other hand, just because you've got a grip on that idea doesn't mean you have a clue about how to handle a binary tree or what makes a good design for a maintainable piece of software. Oh yeah, just because you know what twenty obscure acronyms stand for doesn't mean you know what's going on either... Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Fri, 06 Jul 2001 10:59:17 -0700 From: Russell Bornschlegel Subject: Re: [www] Re: various Chris Pressey wrote: > > Russell Bornschlegel wrote: > > (Hey, when did the word "file" in reference to a collection of bits > > attached to a computer come into use?) > > May 3, 1967, at 4:14 pm. :) *ahem* Cite? -R ------------------------------ Date: Fri, 06 Jul 2001 14:32:10 -0500 From: Chris Pressey Subject: Re: [www] Re: various Russell Bornschlegel wrote: > Chris Pressey wrote: > > Russell Bornschlegel wrote: > > > (Hey, when did the word "file" in reference to a collection of bits > > > attached to a computer come into use?) > > May 3, 1967, at 4:14 pm. :) > *ahem* > Cite? _Ambrose Bierce's Great Big Wonderful Guide to the Precise Date and Time of when Different Words were Introduced into Technical Vocabulary_, 1911 edition. Seriously, as soon as someone had the brilliant idea of storing data on magtape, they must have needed a word to describe any such discrete unit. The office analogy must have presented itself easily, as it has with so many other computer terms ("desktop", etc.) _The Penguin Dictionary of Computers_ (1970) defines a file as "An organized collection of records. The relationship of records on a file may be that of common purpose, format, or data source, and the records may or may not be sequenced." I'd be more curious as to how to word "file" came to mean "to store documents" - prior to that, it would have meant "wearing down or smoothing of a material with a rough flat metal thingy." :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Fri, 6 Jul 2001 22:42:09 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: intellectual property (was Re: various) On Wed, 4 Jul 2001, Chris Pressey wrote: > law...") Laws are trying in some imperfect way to reflect ethics rather > than the other way around... Better still, laws are trying to reflect "common agreement" on ethics in a bit "sophisticated" way. (Like, if you were a good person, what would you do?) Anyway, any society has its mostly-unspoken agreements on ethical values, be they written in law or not. So actually law is kind of a reflection of a reflection. > I agree that recent "extensions" to copyright law have been contrived to > protect companies like Disney from having to release 70-some-yr-old > material into the public domain... to me this goes against the spirit of > copyright, regardless of their nonsensical arguments of "oh, if we don't I'm one of those who claim that all patents and copyrights should be deprived of their legal status. They add a level of possible cheating in the world of innovation which is already wicked enough. They add totally counterproductive factors into innovative competition, such as competing standards, capital gathering capital, and making money by "selling" and "buying" rights. > Which strikes me as canny. Being paranoid about people "stealing the > code" is falling prey to the same thought patterns that keep code closed > in the first place! True. Anyway, if your code is not brilliant or coming to an actual need, the only way you can get money for it is to advertise it. If it is brilliant or fulfills some need, you get your profits irrespective of whether you keep it closed. Panu ------------------------------ Date: Fri, 6 Jul 2001 22:53:07 +0300 (EET DST) From: Panu A Kalliokoski Subject: re: intellectual property On Thu, 5 Jul 2001, Markus Kliegl wrote: > Companies are reluctant to use free software, because it usually isn't > supported by an organization. "No warranty" is scary if there's a bug that I've been in a company, and usually this was the reason. However, most open source products are better supported than their commercial counterparts. The "real" reason seems to be that if some system does not work and you can prove your customer the problem is not in your company's code, it's still your problem if the problem is in open source code but not if it's some other company's closed code. This goes along with the fact that commercial programming does not walk the line of "if we make a bug, it's a mistake" but instead "if somebody notices a bug, it's a mistake." See why I abhor commercial products? (Take Oracle, for example. Once you get a working Oracle server you _don't_ want to upgrade it.) > available. Then there's of course the problem of the free software > sometimes requiring its output to be free as well, which I believe is the > case with GCC. (Bison and Flex I think allow people to do whatever they > want with generated code.) GCC does too. GPL takes this into account. GCC is probably the most wide-spread compiler in the commercial world, so it wouldn't make much sense if it had such restrictions. > The reason Perl and Linux are used commercially so much is that they are > a) marketed quite well, i.e. are a "buzzword" in relation to eBusiness > b) there are lots of books available > c) they can get commercial support (from vendors, etc.) Snowball effect. If somebody uses it, we can use it. And most other free OS's are plain unknown to the commercial world. Besides, Linux is a wonderful development platform for e.g. using experimental hardware (been there, done that). Panu ------------------------------ Date: Fri, 6 Jul 2001 22:55:37 +0300 (EET DST) From: Panu A Kalliokoski Subject: [lang] Re: [BF] Common Repository On Thu, 5 Jul 2001, Markus Kliegl wrote: > There's plenty of possibilities... does someone have a server they can use > for this? (Panu, can you set up something on esoteric.sange.fi?) I could. But I have no time to write upload scripts and I won't give user accounts to the machine just to enable you to use ftp. So if somebody cares to make an upload script, I can make a http-based repository; or you can simply send your bf stuff to me, and I'll put it there. Panu ------------------------------ Date: Fri, 06 Jul 2001 13:00:08 -0700 From: Russell Bornschlegel Subject: Re: [www] Re: various Chris Pressey wrote: > I'd be more curious as to how to word "file" came to mean "to store > documents" - prior to that, it would have meant "wearing down or > smoothing of a material with a rough flat metal thingy." :) http://www.everything2.com/index.pl?node_id=239339 According to Webster, it's two separate words; soldiers and documents are in files derived from the French "file", row <- Latin "filum", thread. The metal thing is Anglo-Saxon "feol" <- various german and northern european forms. "It is upon a file with the duke's other letters." -- Shakespeare -R ------------------------------ Date: Fri, 06 Jul 2001 13:05:15 -0700 From: Russell Bornschlegel Subject: Re: intellectual property (was Re: various) Panu A Kalliokoski wrote: > I'm one of those who claim that all patents and copyrights should be > deprived of their legal status. They add a level of possible cheating in > the world of innovation which is already wicked enough. They add totally > counterproductive factors into innovative competition, such as competing > standards, capital gathering capital, and making money by "selling" and > "buying" rights. And you're okay with the side-effect that any author's work can be pirated by unauthorized publishers? Or do you see all the publishers collapsing for lack of IP, so everyone has to self-publish? -RB ------------------------------ Date: Fri, 6 Jul 2001 23:22:06 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [licences] BSD vs GPL war On Thu, 5 Jul 2001, Keith Gaughan wrote: > The BSD licence is freer, no doubt. It's also a licence for angels, the > same way as communism is a political system for angels. Both things have a > flaw -- the don't fit in with human nature and require people to act like > angels for them to work. Urg. Somehow I seem to disagree with most everything here. Okay, BSDL is freer, but I don't think it's a licence for angels, because you can't break it anyway (angels are those that _do_ have restrictions but won't break them). And communism is not the system for angels, because it's a measure in cooperation, and angels don't need cooperation. Anarchy is probably the system of angels. Okay, that was way too symbolistic to be taken as an argument, but my point about BSDL is this: there is no such thing one should deny but BSDL does not. There's absolutely _no harm_ somebody can do you by stealing your source, if it's open. And if it's closed, the only harm you can get is commercial damage. > The GPL and LGPL, for all their flaws, are more pragmatic. They don't > presume that people are angels. Erm? What they do seem to presume is that it creates more common good to be jammed in endless licence incompatibilities like QPL/GPL than to give your code to people who want to close it after modifications. Does not make much sense to me. > This is why, even though dictatorships are > the most efficent system of government, we live in democracies -- you can't Most efficient? I always thought the only reason why we don't live in dictatorships is that it's not an efficient enough system to use in a big state. Bad governers are killed or kill their state and themselves with it, whereas democracy makes the will of the masses a common law, _really_ digging up the devilish nature of human beings. > The GPL's pragmatism comes from how it gives up some of the freedom to make > sure it stays free. No piece of open source software can be deprived of its freedom (it's already released, so what's the big deal?). What GPL denies is closing your changes to GPL'd software. That does not make sense to me, because they're your modifications and undoubtedly somebody is willing to remake those changes if they're useful. (Of course, I don't see why anybody would close their changes anyway.) > The BSD licence is a licence of a perfect world. This is not a perfect > world. That's the advantage of the GPL. Cut away all the philosophical > pretense and that's what you're left with. Interestingly, I see it exactly the other way around. GPL as a philosophical licence, BSDL as a practical one. Panu ------------------------------ Date: Fri, 6 Jul 2001 23:26:54 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: intellectual property (was Re: various) On Thu, 5 Jul 2001, Keith Gaughan wrote: > >How is this a flaw? > > It's a flaw if you don't want the code to be made proprietary. I don't understand this at all. If my web page has a BSDL'd piece of software, how can you make that proprietary? I mean that specific code there. How do you make it proprietary? > But the original product doesn't benefit from this and it also introduces a > fork, which isn't healthy. The Free Software can't avail from any I'd say forks are very healthy. But they don't naturally really happen that often. Panu ------------------------------ Date: Fri, 06 Jul 2001 15:48:38 -0500 From: Chris Pressey Subject: Open Sores Illgol##: Upgrade now!!! Just lettin' y'all know that the Next Generation Product of the venerable ILLGOL line is now available: Open Sores Illgol## deprecates ILLGOL, Illgola-2, and Illberon, and as such those OBSOLETE products are now DISCONTINUED. Upgrade now!!! http://www.catseye.mb.ca/esoteric/illgol/%23%23/ To try to keep everyone happy, Illgol## (pronounced either "Illgol Pound Pound" or "Illgol double-sharp"... I'm not sure which) has been licensed under eight, count'em eight, open-source licenses: I. Artistic License II. BSD License III. GPL IV. LGPL V. MIT License VI. Mozilla Public License VII. Q Public License VIII. zlib/libpng License Pick one. Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Fri, 6 Jul 2001 23:55:16 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [ On Fri, 6 Jul 2001, Russell Bornschlegel wrote: > > I'd say "normal" people should be expected to learn to program, at least > > in a deterministic language. Am I conceptually flawed? > > Depends on what you mean by "expect". I see lots of people who are > endlessly frustrated with computers, who spend hours and hours for > very little apparent benefit. If people are comfortable with > typewriters and filing cabinets, I say let 'em use typewriters and > filing cabinets. Well, they shouldn't do their typewriting by learning to program, I can say that much... more like, being able to program is a basic ability of overall use, much like basic arithmetic, reading and writing, and good manners. > As for programming, I've been screening people who claim to be > programmers who can't seem to learn to program, despite what > expectations you or I might have. :) You should begin the upbringing very early. :) Panu ------------------------------ Date: Sat, 7 Jul 2001 00:04:22 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: intellectual property (was Re: various) On Fri, 6 Jul 2001, Russell Bornschlegel wrote: > Panu A Kalliokoski wrote: > > I'm one of those who claim that all patents and copyrights should be > > deprived of their legal status. They add a level of possible cheating in > > the world of innovation which is already wicked enough. They add totally > > counterproductive factors into innovative competition, such as competing > > standards, capital gathering capital, and making money by "selling" and > > "buying" rights. > > And you're okay with the side-effect that any author's work can be > pirated by unauthorized publishers? Or do you see all the publishers > collapsing for lack of IP, so everyone has to self-publish? My god, this would be a dream world. Because if everyone "had to" (could!) self-publish, the buying audience would be forced to find those self-publishers that meet their interests, instead of relying on commercial publishers. Besides, if you don't want to play with law-granted, unnatural rights of IP, there is no such thing as "piratism" and "unauthorised" publishers. Which is a huge advantage: every publisher is an authorised publisher. Think about that! If what you mean is "are you okay with the fact that you can't control how your work is used", I answer in the positive. Panu ------------------------------ Date: Fri, 6 Jul 2001 23:18:54 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: Re: intellectual property (was Re: various) On Thu, 5 Jul 2001, Keith Gaughan wrote: > At 11:49 05/07/2001 +0200, Markus Kliegl wrote: > >Pragmatic vs. Pure > > > >BSD vs. GPL > > My god, no! That's backwards! The BSD licence is the pure licence. The GPL > has clauses in it to make sure the software stays Free. The BSD licence is > the licence of idealists, but this kind of idealism belies a certain > naivety, but only if you want your code to stay Free. On the other hand, > the GPL is a *puritan* licence. It demands that the code covered by it > stays free. It's pragmatic because it requires a certain loss of freedom > (the freedom to make the code proprietary) in order to keep it free. I see it the other way around: the GPL enforces that the software be used in its philosophical pure manner, while the BSD license is more pragmatic and allows using the software for impure, i.e. proprietary, commercial, uses. > > Granted, the GPL is too complicated. It could be vastly simplified. My > perfect licence would be one based on the BSD licence which required that > changes be released back to the public in the form of source, just as the > GPL does. I wouldn't use it for everything, but I'd use this for software > which I'd usually put under the GPL. I'm sure something like that exists with all the licenses around these days. All this license, copyright, patent, etc. stuff is really getting on my nerves. I think I'll just put everything in the public domain, or under the Erisian Public License, for that matter. People can generally be trusted to give credit where credit is to be given. The philosophy of the FSF is what I believe in and it would be great if that would be the way things are. The GPL just causes too many problems with other people with the same intent, i.e. all those FreeBSD problems. I find the BSD license even worse, though. So, back to the way "things might be and ought to be [and once were]"... public domain. > > You know, the Nokia licence is quite a nice one. It's worth a look if you > don't like the others. It's even longer than the GPL! The Nethack General Public License seems to basically be the same as the GPL, but about two and a half times shorter (the OSI seems to have gotten the title of that document mixed up with another license). http://www.opensource.org/licenses/nethack.html > > K. > > -- > Keith Gaughan > Software Developer, Digital Crew Ltd. > Take off every '.sig'!! Markus ------------------------------ Date: Fri, 6 Jul 2001 23:22:06 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: Re: intellectual property (was Re: various) On Fri, 6 Jul 2001, Chris Pressey wrote: > Keith Gaughan wrote: > > > > If you want a licence which is even more free than the BSD licence, take a > > look at the one Peter Hartley uses for his software: > > http://www.chaos.org.uk/~pdh/software/ > > He could have saved himself a lot of typing by resorting to this simple > eight-word phrase: > > "These works are part of the public domain." > > Or, if you want a license which is even more Erisian than Hartley's: > http://www.catseye.mb.ca/gwadfc/epl.cgi I think I'll start using the Erisian Public License... or I'll write up an own static license that states about the same :-) > > Chris > > -- > This sentence will be false in 20 seconds starting..... NOW! > Markus ------------------------------ Date: Sat, 7 Jul 2001 00:22:25 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [licences] some examples On Fri, 6 Jul 2001, Markus Kliegl wrote: > I think I'll start using the Erisian Public License... or I'll write up an > own static license that states about the same :-) Self-referential: This licence follows the conditions stated in this licence. Otherwise, see this licence for the details. Negative self-referential: This software can be used as is, without any warranty, as you will, except in the ways stated in this licence. Collective self-referential: This licence follows the conditions stated in all licences that are compatible with themselves. Null licence: "Basic set" licence: This licence explicitly follows all conditions not stated in the null licence. And so on... Panu ------------------------------ Date: Fri, 06 Jul 2001 15:02:04 -0700 From: Russell Bornschlegel Subject: Re: intellectual property and silly licenses Panu A Kalliokoski wrote: > > On Fri, 6 Jul 2001, Russell Bornschlegel wrote: > > > Panu A Kalliokoski wrote: > > > I'm one of those who claim that all patents and copyrights should be > > > deprived of their legal status... [snip] > > > > And you're okay with the side-effect that any author's work can be > > pirated by unauthorized publishers? Or do you see all the publishers > > collapsing for lack of IP, so everyone has to self-publish? > > My god, this would be a dream world. Because if everyone "had to" > (could!) self-publish, the buying audience would be forced to find those > self-publishers that meet their interests, instead of relying on > commercial publishers. Yeah, that suddenly does seem to level the playing field. I'm not trying to claim it wouldn't have advantages; I'm just concerned with the costs. > Besides, if you don't want to play with > law-granted, unnatural rights of IP, there is no such thing as "piratism" > and "unauthorised" publishers. Which is a huge advantage: every publisher > is an authorised publisher. Think about that! > > If what you mean is "are you okay with the fact that you can't control how > your work is used", I answer in the positive. Are you okay with someone else taking your work -- source code or popular novel -- and making money off it? Someone else taking your novel, putting their name on it, and selling it as their own? Someone else taking your serious philosophical novel, adding more sex scenes to it, leaving your name on it, and selling it as the "unexpurgated version"? We've been there before, and that's why copyright law now exists. Panu also wrote: > Self-referential: > > This licence follows the conditions stated in this licence. Otherwise, see > this licence for the details. > [snip] Another Paradoxical License: This license covers all software which is not covered by any software license. Dirty Harry License: So you have to ask yourself -- does the holder of this copyright have the resources to enforce his copyright, or doesn't he? Can he afford a lawyer, or not? Do you feel lucky, pirate? Do you? Go ahead, make my day. Fear, Uncertainty, and Doubt License: You know The Company can enforce this copyright. You know The Company can afford a whole flotilla of lawyers. You know The Company could probably buy the judge if it came to a court of law. You know they've got black helicopters and wiretaps and parabolic mikes and Men In Black. You should just give yourself up right now. Chain Letter License: When you receive this software, you should immediately mail it to five other people. Markus Kliegl didn't pass the software along, and he stubbed his toe. Chris Pressey did pass the software along, and the next day he received a check for 50,000 Zorkmids from Bill Gates along with a note of apology. Nondeterministic Chain Letter License: When you receive this software, you must immediately select one of the following Illgol##-compatible licenses: I. Artistic License II. BSD License III. GPL IV. LGPL V. MIT License VI. Mozilla Public License VII. Q Public License VIII. zlib/libpng License ...and remove that license from this license definition. The software can then be redistributed according to any of the remaining licenses. Schroedinger's License: The first time you use this software, flip a coin. If it comes up heads, this software is covered by the BSD License. If it comes up tails, this software is covered by the GPL. If it lands on edge, you must delete the software from your computer immediately and kill your cat. Rascality License: Sure, go ahead and copy this software, but don't let me catch you at it, you rascal. The Counting License: (As any other open source license of your choice, but with the additional clause:) This is generation 0 of the license. Please increment the generation number immediately when you receive this software, before redistributing it. Delegation License: Please compose your own open-source license in this space. Licenses should be pragmatic yet idealistic, should not infringe on any freedoms including the freedoms of bad people, should be written in a form which is clearly understandable yet legally binding and unambiguous. The Local Maximum License: This software may be freely redistributed only after you have modified it to either add at least one feature or to fix at least one bug. -RB ------------------------------ Date: Sat, 7 Jul 2001 00:51:18 +0200 (CEST) From: markus.kliegl@t-online.de (Markus Kliegl) Subject: [Licenses] Contest Inspired by all these wonderful licenses around, I am pleased to announce the 'Software License Contest' (and I sure hope someone's going to submit the GPL as an entry :-)... so, might as well waste this weekend. Those licenses that Panu and Russell already sent are counted as entries. Goal: To create the most unusable license (free or not) and to enforce usage of it. Submitting: Submissions should be sent to misc@esoteric.sange.fi with CC to Bill Gates, Richard Stallman, Linus Torvalds, ET and Mickey Mouse. Submission deadline is this Sunday (23:50 whatever your local time zone is), giving you almost no time Judging: The winners will be announced this Monday by me. Of course, anyone can judge and announce the winners according to their idea of how this contest should be run... the idea is having a billion license contests, but everyone only has to submit once. Invited judge is Jenny, though actually I never invited her to judge. Entries will be judged on the following criterias: * Unusability * Usability * Affecting of users of the software * Recursiveness * Paradoxicality (?) * anything else that comes to mind Prizes: The authors of the three winning entries will receive a free copy of a special piece of software licensed under all entries and signed by Eris! These prizes will of course be sent to misc@esoteric.sange.fi Markus ------------------------------ From: "Plant Kingdom" Date: Fri, 6 Jul 2001 23:13:17 -0400 Subject: Re: Google 4 language HAXORS On Fri, 06 Jul 2001 09:37:46 +0200, Fr=E9d=E9ric van der Plancke wrote: > Gerson Kurz wrote: > > - Hacker > > - Bork ! Bork ! Bork ! > > - Elmer Fudd > > - Tagalog (I think Intercal knows this one, too) > > and "Pig Latin" ... > > Note that Tagalog is a real human language spoken > in the Philippines. (search for "tagalog language"). And the correct name for the dialect spoken by the Swedish Chef is "Mock Swedish." As all right-thinking, well-educated citizens of the world should know. ------------------------------ From: Gerson.Kurz@t-online.de (Gerson Kurz) Subject: The worlds first python-written virus Date: Sat, 7 Jul 2001 06:39:42 +0100 "In the Slovenian pavilion at the Venice Biennale a group of artists and hackers, 0100101110101101.ORG and EpidemiC, is exhibiting a new computer virus called 'biennale.py'." http://www.heise.de/tp/english/inhalt/sa/3642/1.html The Virus: http://www.0100101110101101.org/home/biennale_py/biennale.py.html ------------------------------ From: Gerson.Kurz@t-online.de (Gerson Kurz) Subject: kalk Date: Sun, 8 Jul 2001 15:01:39 +0100 kalk is a Qt GUI based Linux / Win32 calculator with a different approach. It is somewhat similar to a language interpreter; while at the same time continuously evaluating expressions, so you can model expressions to fit a result (try it & you'll see). Features of interest to this group: - You can define functions on the fly, e.g. f(x,y)=sin(x*x)+cos(y*y) f(1,2) - You can query truthtables, e.g. "TT(~p|~q)" will result in p | q = ? F | F = T F | T = T T | F = T T | T = F - Open Source. kalk is available for both Win32 & Linux @ http://www.p-nand-q.com/kalk.htm PS. Of course, I was screaming aloud during the Linux port. Here is why: Problem: You want to use your IDE CD Burner as a CD Burner. Solution: Logon as root. Start yast. Go to System Administration. Goto Kernel & Boot configuration. Goto Lilo configuration. Append a line hdX=ide-scsi where X is the number of your CD Burner device. You can find the name by searching thru /var/log/boot.msg. Edit /etc/init.d/boot.local. Append a line "/sbin/modprobe ide-scsi". Do a "ln -sf /dev/scd0 /dev/cdrecorder" from a xterm window. Edit /etc/fstab, so that it contains a line "/dev/cdrecorder /media/cdrecorder auto ro,noauto,user,exec 0 0" Reboot. Free translation from http://sdb.suse.de/de/sdb/html/tbraza_ide_brenner.html So much for Linux on the Desktop. ------------------------------ Date: Sun, 8 Jul 2001 21:50:08 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: intellectual property and silly licenses On Fri, 6 Jul 2001, Russell Bornschlegel wrote: > > If what you mean is "are you okay with the fact that you can't control how > > your work is used", I answer in the positive. > > Are you okay with someone else taking your work -- source code or > popular novel -- and making money off it? Someone else taking your > novel, putting their name on it, and selling it as their own? > Someone else taking your serious philosophical novel, adding more > sex scenes to it, leaving your name on it, and selling it as the > "unexpurgated version"? We've been there before, and that's why > copyright law now exists. You ask for my opinions? There are different degrees in this thing. For the first question, I'm okay with that. After all, anybody can write something, and making money is the real trouble. As for the second one, I consider that bad manners (because it's lying that it's their work) -- I think I'd be okay with that, but I don't expect others to be. As for the third one, it's a different scenario altogether. It's not only concerned by copyright law, but also defamation law. Defamation is different, because it can harm the original author, which the first two don't do. And again, it's lying - lying about what the author said. All in all, I see a big difference between "intellectual property" (which I don't think should necessarily exist) and "correct reference". The first one is counterproductive (IMO), the second one very useful. > Another Paradoxical License: > This license covers all software which is not covered by any software > license. ... and, from now on, mote this be called "Russell's paradox". Do you happen to know the way they rephrased the paradoxical axiom? It's really quite interesting. > Delegation License: > Please compose your own open-source license in this space. Licenses > should be pragmatic yet idealistic, should not infringe on any > freedoms including the freedoms of bad people, should be written > in a form which is clearly understandable yet legally binding and > unambiguous. Hey, good idea! But how come did you distribute the first copy? You're breaking the licence... > The Local Maximum License: > This software may be freely redistributed only after you have > modified it to either add at least one feature or to fix at least > one bug. ... or to add one? Panu ------------------------------ Date: Sun, 8 Jul 2001 22:02:33 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: [Licenses] Contest I'd like to add one entry, the One True Licence: You may do with this piece of software whatever you damn please, as long as you follow good manners, don't irritate anybody by your actions in an unfair situation, ask the author for advice in unclear situations (such as copying the software to a friend who, however, does not want it), take others' feelings into account, think forward the consequences of your deeds, and most important of all, don't tease or harm anybody. The meaning of this software is to provide people with joy, enlightenment, solved problems, fulfilled needs, and excited interest. All use of the software should aim towards these goals. THE SOFTWARE IS PROVIDED AS IS, WITHOUT ANY EVIL BACKTHOUGHTS, IMPLIED OR EXPRESS, AND ABSOLUTELY NO HARM SHOULD EVER EMERGE FROM USING THE SOFTWARE. IF SUCH A CASE APPEARS, LET EVERYBODY USE THEIR OWN BEST JUDGEMENT TO DECIDE WHO IS AT FAULT AND HOW THE SITUATION WOULD BEST BE CORRECTED. Go into the world, fear nothing, and be nice to your surroundings (nice by their standards, not yours). Panu Am fuar -> symb <- am fesh atehwa@iki.fi ------------------------------ Date: Sun, 08 Jul 2001 19:58:05 -0500 From: Chris Pressey Subject: Re: [Licenses] Contest Markus Kliegl wrote: > Goal: > To create the most unusable license (free or not) and to enforce usage > of it. Software Lie Sense [courtesy of Wonk] ------------------ By installing this software the end user hereby agrees that it will, under no circumstances, be used as birth control, a paperweight, religion, a doorstop, a pencil sharpener, forceps, a means of killing Kenny, or as chewing gum remover. The SEEKrit Erisian Girlscout Public License -------------------------------------------- There is no SEEKrit Erisian Girlscout Public License, and it's a SEEKrit besides, so shhhhhhhhhhhhhhhhhhhhhhhhhhh The Feline Public License ------------------------- Meow meow meow meow, meow meow meow meow, meow meow meow: Meow meow meow meow (meow meow meow); Meow meow meow, meow meow; Meow meow purrrrrrrrrrr. MEOW MEOW MEOW MEOW NO WARRANTY MEOW MEOW MEOW MEOW MEOW MEOW MEOW MEOW. The Crowleyan Public License ---------------------------- Do as Thou Wilt Shall be the Whole of this License. The Unconditional Assimilation Public License --------------------------------------------- ALL YOUR BASE ARE BELONG TO US Chris -- "This administration is doing everything we can to end the stalemate in an efficient way. We're making the right decisions to bring the solution to an end." -- guess who, Washington, D.C., April 10, 2001 ------------------------------ Date: Mon, 9 Jul 2001 11:29:28 +0300 (EET DST) From: Panu A Kalliokoski Subject: [lang] Re: [lex/yacc] Stack-based prefix/"Forward Polish" syntax? On Wed, 27 Jun 2001, Russell Bornschlegel wrote: > Forth's "Reverse Polish"/postfix operator syntax just had to go > for readability reasons[2]. How do you do a stack-based language > without postfix operators? Simple, the parser reads each _line_ > backwards, so the programmer writes: The other way to do this (and IMO better, as it gives more flexibility) is to use an operator stack instead of an operand stack. For operators that have got some of their operands but not all, you can use closures. add add 2 3 4 stack at each stage ------------------- add < add > add < add add > 2 < add (add 2) > 3 < add (add 2 3) > ==> < (add 5) > 4 < (add 5 4) > ==> < 9 > (probably to be further reduced) You can use mixed FPN and RPN to make the language more powerful: add 2 map ! mylist stack at each stage ------------------- add < add > 2 < (add 2) > map < (map (add 2)) > (map behaves in RPN fashion) ! < (map (add 2)) ! > mylist < (map (add 2)) (conts. of mylist) > ==> < (conts. of mylist, each item incr'd by 2) > (probably to be further reduced) > What's the best way to implement a prefix-notation language using > lex and yacc? It seems like lex would need to find the tokens on a > line, throw out comments, then reverse them before presenting them > to yacc. Maybe a lex-to-lex preprocessor? I'm very new to lex and If you use RPN (or if that's what the actual interpreter sees), you don't need yacc at all. But if using yacc is OK, I propose you throw away the reversing (because it's so uncertain a technique) and use proper yacc rules, because that's just as easy: prog : expr expr : add_expr | sub_expr | neg_expr | const | var add_expr : ADD expr expr ... and so on. or even: expr : unary_expr | binary_expr | trinary_expr | term unary_expr : UN_OP expr binary_expr : BIN_OP expr expr trinary_expr : TRIN_OP expr expr expr ... and so on. > yacc, so I don't know if they have any explicit features which > would help with this kind of thing, or if there's a standard idiom > for it... lex and yacc are exactly designed for this kind of thing, but not the way you want to do it. Instead, they do the stack and state tracking for you, letting you specify whatever unambiguous grammar you please. FPN is always unambiguous, so the grammar is very easy to make. b5's grammar makes an excellent example of basic yacc usage. (However, it does not execute the language straight in the parsing phase, but instead makes a representation tree from it to be later evaluated.) It's available at http://www.sange.fi/~atehwa-u/b5/b5-grammar.y > [2] Do speakers of German and other at-the-end-of-the-sentence- > the-verb-comes languages have more aptitude for Forth than do > we 'merkins? I don't think so. Besides, German is IMO very illogical as it comes to word order; that the _first_ verb of main clause is not at the end and that there are more prepositions than postpositions just doesn't make a very coherent view for me. See word order of Japanese. Panu ------------------------------ Date: Mon, 9 Jul 2001 11:21:40 +0200 (CEST) From: "Rafal M. Sulejman" Subject: [lang] Re: [lex/yacc] Stack-based prefix/"Forward Polish" syntax? On Mon, 9 Jul 2001, Panu A Kalliokoski wrote: > On Wed, 27 Jun 2001, Russell Bornschlegel wrote: > > > [2] Do speakers of German and other at-the-end-of-the-sentence- > > the-verb-comes languages have more aptitude for Forth than do > > we 'merkins? > > I don't think so. Besides, German is IMO very illogical as it > comes to word order; that the _first_ verb of main clause is not > at the end and that there are more prepositions than > postpositions just doesn't make a very coherent view for me. See > word order of Japanese. Heh. Try Polish if you want something REALLY illogical ;) AFAIK it's one of the hardest-to-parse natural languages... But we use "normal" algebraic notation ;) Few (many) years ago someone invented ENGLISH language for database munching ("Yes! You CAN program in ENGLISH using our tools..."). POLISH can be it's equivalent for spreadsheet macros ;)... (if you really hate s-sheet users)... -- # . . . | Rafal M. Sulejman /"\ # . . . | rms@spam-haters.poczta.onet.pl> \ / # . . . | ASCII Ribbon Campaign for HTML-free mail X . . . # | / \ ------------------------------ From: Steve Mosher Subject: Re: [Licenses] Contest Date: Mon, 9 Jul 2001 08:36:03 -0300 On Sun, 08 Jul 2001, Chris Pressey pressed some keys and this came out: > The Unconditional Assimilation Public License > --------------------------------------------- > > ALL YOUR BASE ARE BELONG TO US That's "ALL YOUR BASE-CODE ARE BELONG TO US" (See GPL) ;) -- Steve Mosher, Mad Scientist Never worry when you're strong, but keep the absurd absurdity close to the youth. Your clarity is insane; a youth disturbs the strong heart. ------------------------------ Date: Tue, 10 Jul 2001 14:57:19 -0500 From: Chris Pressey Subject: Re: [www] Re: various Russell Bornschlegel wrote: > Chris Pressey wrote: > > I'd be more curious as to how to word "file" came to mean "to store > > documents" [...] > According to Webster, it's two separate words; soldiers and documents > are in files derived from the French "file", row <- Latin "filum", > thread. Ah - phylum, filament, etc. Gotcha. > The metal thing is Anglo-Saxon "feol" <- various german and > northern european forms. OK... my point here, assuming I had one, which I didn't, was that, whereas in computer programming languages we go to great lengths to avoid namespace collisions, in spoken languages we actually encourage them, presumably to reduce the number of "terminal tokens" (phoneme-strings that you have to be able to pronounce and distinguish.) I mean, there's no good reason Interlish couldn't still use the word "feol", it's not like it's being used for something else now. But, while that pronunciation was maybe natural in the Anglo-Saxon tongue, it isn't really present in the Interlish phoneme-base, where it could easily be confused with a word like "feel". So instead Interlish makes up compound words and phrases. For specificity, we can distinguish a file from a file by calling one by it's full name, calling it a handfile. So I guess the question is, to what extent do filesystems and variable namespaces benefit from such approaches of "applied ambiguity"? On most operating systems there is the relatively crude measure of a "search path." The thing I've found about the search path is that it's one of the harder things for "non-techies" to understand. But the obstacle seems to be that the basic structure of a typical filesystem is hierarchical. (Although symbolic links can turn it into a graph, for better or worse.) The search path is needed to tack another "dimension" onto the hierarchy. Perhaps if the basic structure had a "higher dimensionality", it would be easier to handle in general. For example, I might decide to put a hypothetical game, written in a some hypothetical scripting language in a directory called foolang/games. Or, I might wnat to put it in a directory called games/foolang. Which one I choose is more or less arbitrary. This, to me, is a good argument for using *keywords* rather than "directory" names in a filesystem; I'd put it in games&foolang, to which foolang&games would also match. Adding wildcards to this approach lets you specify "virtual directories"; for example, asking to view the files for "games&*" would list files in games&foolang, perl&games, whatever&games&stuff, etc. This obviates the need for a search path; you just look in "executables&*" and lo and behold, there are all your executables. It's also largely backwards-compatible. Listing the files in usr&local&share&whatever would list the same files as /usr/local/share/whatever/ would in unix (assuming both filesystems had copies of the same files, of course.) The point is, existing filesystem-accessing code wouldn't break. Chris -- ERROR: Corrupt feolsystem. Spinning silk, please wait... ------------------------------ Date: Tue, 10 Jul 2001 14:13:54 -0700 From: Russell Bornschlegel Subject: Re: [www] Re: various (Argh. Missed reply-all again) Chris Pressey wrote: > OK... my point here, assuming I had one, which I didn't, was that, > whereas in computer programming languages we go to great lengths to > avoid namespace collisions, in spoken languages we actually encourage > them, presumably to reduce the number of "terminal tokens" > (phoneme-strings that you have to be able to pronounce and distinguish.) I'm not sure what the natural-linguistic _benefits_ are to allowing namespace collisions, but historically the _dangers_ of namespace collisions in spoken language haven't been very serious because of the usual speed with which humans do things. For example: X: "Do we turn left here?" Y: "Right." Prior to the 20th century, there was plenty of time to note this ambiguity and correct it before any damage was done. On the freeway at 120kph, this is more dangerous than it once was, and people who are paranoid about ambiguity (like programmers) may go out of their way to say "Correct" or "No, turn right" instead of just saying "Right". Now, with computer performance being rated in billions of mistakes per second, ambiguity can be a dangerous thing indeed, so the usual computer-system strategy for dealing with ambiguity is to stop and complain. But you knew all that. > So I guess the question is, to what extent do filesystems and variable > namespaces benefit from such approaches of "applied ambiguity"? > [...] > > But the obstacle seems to be that the basic structure of a typical > filesystem is hierarchical. [...] > > For example, I might decide to put a hypothetical game, written in a > some hypothetical scripting language in a directory called > foolang/games. Or, I might wnat to put it in a directory called > games/foolang. Which one I choose is more or less arbitrary. This, to > me, is a good argument for using *keywords* rather than "directory" > names in a filesystem; I'd put it in games&foolang, to which > foolang&games would also match. There's some implicit support in modern OSes for what you're talking about here. Consider the increasing reliance on find-file capabilities on modern machines with big hard disks where it's easy to lose stuff. Consider web search engines and the integration of filesystem browsers with web browsers. Consider symbolic links and the desktop -> my computer -> filesystem -> desktop-file circularity you mentioned a while back, turning the filesystem tree into a cyclic graph. An underlying tree representation is kept mainly to provide start and end points for enumerations, and for compatibility. > Adding wildcards to this approach lets you specify "virtual > directories"; for example, asking to view the files for "games&*" would > list files in games&foolang, perl&games, whatever&games&stuff, etc. > This obviates the need for a search path; you just look in > "executables&*" and lo and behold, there are all your executables. Yeah; dealing with the new ambiguities thus created, and keeping the system indexed efficiently, become the new problems to solve in this case. -RB ------------------------------ Date: Tue, 10 Jul 2001 14:38:41 -0700 From: Russell Bornschlegel Subject: Re: [www] Re: various Chris Pressey wrote: > But the obstacle seems to be that the basic structure of a typical > filesystem is hierarchical. (Although symbolic links can turn it into a > graph, for better or worse.) The search path is needed to tack another > "dimension" onto the hierarchy. Perhaps if the basic structure had a > "higher dimensionality", it would be easier to handle in general. Another thing about Everything2 that might be relevant to a filesystem is that the act of going from any node in the system to any other node, no matter how that transition is made, forges what's called a "soft link" in the system, and the N most-travelled softlinks are available to click on. Applying this to the filesystem, consider a directory with a bunch of audio files in it; sometimes you're copying files to other directories for archive or export, other times you're bringing them into an editor to mess with them, other times you're just playing them back. Lots of OSes are keeping track of recently used applications and recently accessed documents, but they're keeping track of them on the system- wide or per-user scale instead of on a per-directory basis. Essentially, I'm talking about the system evolving different "workspaces" (containing pointers/shortcuts/links to files, apps, and directories) as you use the system, so that in the "usability dimension", things that are frequently used together move closer together. This is exactly analogous to the (legendary/mythical?) college campus where they didn't put down any paths between buildings the first year, letting the students walk across the lawns as they wished, then later paved over the emergent paths where the grass had been stomped to death. -RB ------------------------------ Date: Tue, 10 Jul 2001 18:38:56 -0500 From: Chris Pressey Subject: ESO (was Re: [www] Re: various) So I guess I wouldn't be griping about OSen and licenses unless I had it in mind to do some project involving them. So this raises the question, am I still interested in building ESO, the esoteric operating system? Well, what does that mean anyway? Some people want an obfuscated OS, some people want a small OS, some people want something elegant, some people just want something strange. All of these things are honourable esoteric goals, of course, but I think we'd have be more specific if we didn't want the thing to turn out to be anything but a monstrosity (the bad kind). I'd personally like to explore lots of little interrelated alternative models - i.e. something simple and elegant. But it might not always be strange or obfuscated or small. Of course, I wouldn't go out of my way to make it normal or bloated, either :) A few things are fairly apparent. People won't be happy unless it's a real OS, i.e. you can run it on your computer without running any other OS under it. (That doesn't mean it can't also be emulated, obviously.) The big problem that presents itself there is that there are no runtime libraries available that don't assume a certain OS is running under them. For example, every time you compile a C program, it links in crt1 and libc functions; those functions are specifically tailiored, usually in asm, to the OS you're currently running (unix, windows, whatever.) So we'd need to basically re-implement libc, or rather, the parts of libc that we like/need, at a level that works under ESO. Now this raises a fairly interesting question to me, why not have the OS provide all those common services as services, rather than as functions that are linked statically into the executable? That removes a layer of abstraction, but that may not be a bad thing in an esoteric OS - it would help keep it smaller and simpler. So in this spirit I would propose a very simple virtual machine (that doesn't really resemble the machine languages Befunge, Brainf*ck etc) where essentially the only functionality that is exposed (besides branching & stuff) is the ability to call an operating system function. That's right, even incrementation would be defined in an API somewhere. (Of course this could always be optimized [inlined] in a "serious" implementation of ESO.) Russell Bornschlegel wrote: > (Argh. Missed reply-all again) > Chris Pressey wrote: > > OK... my point here, assuming I had one, which I didn't, was that, > > whereas in computer programming languages we go to great lengths to > > avoid namespace collisions, in spoken languages we actually encourage > > them, presumably to reduce the number of "terminal tokens" > > (phoneme-strings that you have to be able to pronounce and distinguish.) > > I'm not sure what the natural-linguistic _benefits_ are to allowing > namespace collisions, [...] Fewer words. I'm not sure why that's a benefit exactly - fewer "keys in the dictionary", I suppose. Phonetic laziness? > > [keywords instead of directories] > There's some implicit support in modern OSes for what you're talking > about here. Consider the increasing reliance on find-file capabilities > on modern machines with big hard disks where it's easy to lose stuff. > Consider web search engines and the integration of filesystem browsers > with web browsers. I know what you're getting at, but it just isn't quite the same to me. Searching for something (by name, or properties or content) is an intensive process in comparison to simply locating a file by a "fully qualified name". > Consider symbolic links and the desktop -> my > computer -> filesystem -> desktop-file circularity you mentioned a > while back, turning the filesystem tree into a cyclic graph. Yeah, but - yuck :) For one thing, it's directed; (cd place ; cd ..) could leave you in some unexpected, unexplored, mysterious part of the filesystem, if place is a symbolic link. That ain't none too good... it would be nicer if there was something between a symbolic link and a block-link, something that is consistently referred to so that the graph isn't directed. Anyway, that's a bit of a digression, 'cause with keywords you don't really need symbolic links either, I don't think. :) > An > underlying tree representation is kept mainly to provide start and > end points for enumerations, and for compatibility. True; they're not very strong reasons to me, though, certainly not wrt ESO. > > Adding wildcards to this approach lets you specify "virtual > > directories"; for example, asking to view the files for "games&*" would > > list files in games&foolang, perl&games, whatever&games&stuff, etc. > > This obviates the need for a search path; you just look in > > "executables&*" and lo and behold, there are all your executables. > Yeah; dealing with the new ambiguities thus created, and keeping > the system indexed efficiently, become the new problems to solve > in this case. Yeah, efficient indexing might be a bit of a space-hog; each keyword is a key in a multi-keyed database table. With each key there was be an additional index, and that's a lot more overhead than just a directory. As for further ambiguities... well, files could be stored within their "virtual directory" as a bag, so that usr&local&share&foo&READMEcould point n>1 different files. usr&local&share&foo&README would actually point to the newest (by filedate) of the files named README with those keywords, and a syntax like usr&local&share&foo&README#2 or usr&local&share&foo&README@19990701 could be used to reference older versions of README. Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Tue, 10 Jul 2001 17:16:29 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > Russell Bornschlegel wrote: > > Consider symbolic links and the desktop -> my > > computer -> filesystem -> desktop-file circularity you mentioned a > > while back, turning the filesystem tree into a cyclic graph. > > Yeah, but - yuck :) For one thing, it's directed; (cd place ; cd ..) > could leave you in some unexpected, unexplored, mysterious part of the > filesystem, if place is a symbolic link. That ain't none too good... In the world of the web browser, the back button takes care of this. You could also, whenever a symlink is created, create new bidirectional links between the directories/namespaces enclosing both ends of the symlink, somehow... > As for further ambiguities... well, files could be stored within their > "virtual directory" as a bag, so that usr&local&share&foo&READMEcould > point n>1 different files. usr&local&share&foo&README would actually > point to the newest (by filedate) of the files named README with those > keywords, and a syntax like usr&local&share&foo&README#2 or > usr&local&share&foo&README@19990701 could be used to reference older > versions of README. Hmm, sorting by timestamp is the Right Thing in some cases; not sure if it's the right general answer. _R ------------------------------ Date: Tue, 10 Jul 2001 17:26:31 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > So this raises the question, am I still interested in building ESO, the > esoteric operating system? [...] > > A few things are fairly apparent. People won't be happy unless it's a > real OS, i.e. you can run it on your computer without running any other > OS under it. (That doesn't mean it can't also be emulated, obviously.) > > The big problem that presents itself there is that there are no runtime > libraries available that don't assume a certain OS is running under > them. For example, every time you compile a C program, it links in crt1 > and libc functions; those functions are specifically tailiored, usually > in asm, to the OS you're currently running (unix, windows, whatever.) Actually, the C runtime library and the POSIX specification have co- evolved in such a way as to minimize the amount of asm you need to write the CRTL, provided that you can write your own linker and that you have equivalent functionality in the OS to a POSIX subset. Almost all of the CRTL is normally written in C ("the portable assembler"). CRTL implementations typically has more code to deal with cross-CPU portability than with cross-OS portability, for example. (See Plauger's "The C Standard Library" for more details) For a number of practical reasons, assuming that the OS supports a POSIX subset has a lot of advantages. Even if you think POSIX is a bad OS API, it's a very useful one to bootstrap from, in much the same way as having a C compiler on a new platform is a useful precursor to porting your favorite language to it. -RB ------------------------------ Date: Thu, 12 Jul 2001 01:25:57 +0300 (EEST) From: Esoteric languages community Subject: Re: pretentious rant about operating systems (fwd) Originator: fvdp@decis.be (Frederic) Message-ID: <3B417CBD.DCE90183@decis.be> References: <3B3FA3E9.42C03905@catseye.mb.ca> Content-Type: text/plain; charset=3Diso-8859-1 Content-Transfer-Encoding: 8bit X-MDRemoteIP: 192.168.0.7 X-Return-Path: fvdp@decis.be X-MDaemon-Deliver-To: misc-bounce@esoteric.sange.fi Reply-To: fvdp@decis.be Chris Pressey wrote: >=20 > I'll get right to the point. WHERE'S THE METADATA? >=20 > The answer: when it is there, it is spread across umpteen configuration > files that don't know about each other, OR, it is lumped willy-nilly > into an obscure, convoluted centralized database... but it's usually not > there. Configuration management is a hard thing anyway. I agree the current state of affairs is unsatisfying. >=20 > The fact is that system objects (files, devices, etc) have relationships > with each other. Relationships are nothing new in computer science, > we've been using relational databases for years. So why hasn't the > rather simple concept of "Object-A somehow-relates-to Object-B" filtered > down to the operating system level yet? >=20 > We have dependencies in 'make' scripts, and we have 'install' and > 'uninstall' scripts, and we have 'file type detection' in the form of > either file suffix or 'magic'... but these are all loose and scattered > functions. In the meantime, you can rename a file and none of these > measures are the wiser. =20 Magics resist renaming. (So each binary format should contain a magic IMO). Relying on file extension is both bad and good: under Windows, I rather like the way it works; if I want file F to be opened by=20 program X by default (=3D by double-clicking of F's entry), it suffices to rename F with a suitable extension -- and I can define my own extensions and the way they behave. > (And when an OS does have enough seatbelts to > warn you, the warning is seemingly never specific enough: "file may be > required by some installed packages" - well I'd love to know which > ones?!?) >=20 > There are some other little things that are just shell/filesystem > issues. I'd love for a shell to be able to change to a directory just > by typing its name.=20 In Windows you can already open a file just by typing "start myfile.ext" or even "myfile.ext" (just tested !) (same extension -> program association as described above.) > It should be obvious that "enter" is the verb to be > implied when you just name a directory. I'd also love for the FS to be > able to treat archive files (.tar, .tgz, .zip, .bz2 etc) as directories, > so you could "enter" them and copy files in and out with the usual > commands (not to mention having programs like "find" able to search > within archives automatically :) You can do that at least partially under Windows with a commercial program like ZipMagic (http://www.ontrack.com/zipmagic/): .zip files appear as ordinary folders in Windows Explorer. (I never tried to "explore" them using the DOS command prompt or the "find" command though. Apparently, things like that should work.) Fr=E9d=E9ric vdP "Windows is not an OS" -- but then I wonder what operates my system ? --------------- Error message done ------------------------- --- Listar v0.129a - job execution complete. ------------------------------ Date: Thu, 12 Jul 2001 01:19:23 -0500 From: Chris Pressey Subject: Re: ESO (was Re: [www] Re: various) Russell Bornschlegel wrote: > Chris Pressey wrote: > > Russell Bornschlegel wrote: > > > Consider symbolic links and the desktop -> my > > > computer -> filesystem -> desktop-file circularity you mentioned a > > > while back, turning the filesystem tree into a cyclic graph. > > Yeah, but - yuck :) For one thing, it's directed; (cd place ; cd ..) > > could leave you in some unexpected, unexplored, mysterious part of the > > filesystem, if place is a symbolic link. That ain't none too good... > In the world of the web browser, the back button takes care of > this. Well, in the world of the shell shouldn't there be a 'bk' to complement 'cd'? > > As for further ambiguities... well, files could be stored within their > > "virtual directory" as a bag, so that usr&local&share&foo&READMEcould > > point n>1 different files. usr&local&share&foo&README would actually > > point to the newest (by filedate) of the files named README with those > > keywords, and a syntax like usr&local&share&foo&README#2 or > > usr&local&share&foo&README@19990701 could be used to reference older > > versions of README. > Hmm, sorting by timestamp is the Right Thing in some cases; not sure > if it's the right general answer. Even better, save new files as diffs of old files. And apply diffs to make new diffs when old files are deleted. Kinda complicated, and probably a bit slow, but very, very cool :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Thu, 12 Jul 2001 09:10:42 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > > Russell Bornschlegel wrote: > > In the world of the web browser, the back button takes care of > > this. > > Well, in the world of the shell shouldn't there be a 'bk' to complement > 'cd'? Sure. Get on it, man! Actually, historically, people have added push-cwd and pop-cwd functionality to shells, via aliases, batch files, and other such, then used push-and-cd instead of cd, which makes pop act like the "back button" or "bk" command. Like other things I've brought up here, people have been putting this functionality into filesystems quietly, without talking about what it means in terms of topology. > Even better, save new files as diffs of old files. And apply diffs to > make new diffs when old files are deleted. Kinda complicated, and > probably a bit slow, but very, very cool :) Do "journaling filesystems" do this? I haven't had a chance to learn about them. -RB ------------------------------ Date: Thu, 12 Jul 2001 13:54:31 -0500 From: Chris Pressey Subject: Re: ESO (was Re: [www] Re: various) Hmmm... sorta taking this all together... we have POSIX-based OS'es that stemmed from a time when resources were expensive, so the most graceful way to build them is to bootstrap them through C. At the same time, we have lots of resources these days, so we want to use them. To do things like support "undo" actions (undo cd, undo write-file, undo setenv, etc, is what we're talking about here.) And other stuff. The result is sort of that these features have been added piecemeal - they're not packaged, they're not in the core, and they're certainly not standardized - I mean a piece of software like ZipMagic is pretty cool, but what if you want to support the next big compression format (like bz2?) It would be nice if the OS "knew what compression was," i.e. provided an API, so the functionality would be more general, and the same could be said for the other features. Russell Bornschlegel wrote: > > Even better, save new files as diffs of old files. And apply diffs to > > make new diffs when old files are deleted. Kinda complicated, and > > probably a bit slow, but very, very cool :) > Do "journaling filesystems" do this? I haven't had a chance to learn > about them. Possibly - I was thinking like CVS, although probably not as complicated as CVS. Actually, closer to Bluetail's Wikie, a wiki written in Erlang; it saves all additions as diffs to existing entries (I don't know if other wikis do that.) Hmmm... come to think of it, CVS itself could be abused as a chat protocol, with full history keeping no less :) Although it has very little to do with what we're talking about here, I have to say that telnet is RIDICULOUSLY complicated for what it does... I don't know for sure, but I have to imagine that 90% of the negotiation it does could be completely ignored by 90% of the present-day implementations and it would still work fine. Sometimes I think linemode and character-at-a-time mode should be two entirely different protocols. Like, why don't more telnet clients allow you to force the local echo on or off? Oh well... Chris -- IAC DON'T WORRY IAC I WON'T BITE ------------------------------ Date: Thu, 12 Jul 2001 12:09:02 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > Although it has very little to do with what we're talking about here, I > have to say that telnet is RIDICULOUSLY complicated for what it does... Well, again, it wasn't designed that way, it "just growed." Like the whole TCP/IP suite and friends, it's big, unwieldy, starting to fragment, relied on by many, and now becoming very hard to replace. And any possible replacement will suffer from second-system syndrome... But you knew that. -R ------------------------------ Date: Thu, 12 Jul 2001 12:28:06 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > > Hmmm... sorta taking this all together... [snip] > The result is sort of that these features have been added piecemeal - > they're not packaged, they're not in the core, and they're certainly not > standardized - I mean a piece of software like ZipMagic is pretty cool, > but what if you want to support the next big compression format (like > bz2?) It would be nice if the OS "knew what compression was," i.e. > provided an API, so the functionality would be more general, and the > same could be said for the other features. Well, let's see... you could write a filesystem driver that accepted compression systems as codecs in DLLs/shared libraries. If you write the APIs correctly, the codec DLL could support standalone use as well as in-filesystem use. (Since multimedia player software evolved long after filesystems did, they generally work this way already.) The default filesystem codec is essentially "memcpy". The cool thing is you could, possibly, use the same API for filesystem compression and for network compression... Same logic applies to many aspects of an OS; essentially you're moving to a sort of OOP model (encapsulating functionalities behind DLL interfaces, treating plug-in features polymorphically). The overhead of the dynamic-link indirection is assumed to be affordable on a modern system. Pretty soon, you're gonna find yourself arguing in favor of COM and reinventing Win32. -RB ------------------------------ Date: Thu, 12 Jul 2001 16:09:53 -0500 From: Chris Pressey Subject: Re: ESO (was Re: [www] Re: various) Russell Bornschlegel wrote: > Pretty soon, you're gonna find yourself arguing in favor of COM and > reinventing Win32. Well, let's not get carried away, here :) In a sense, yes. But with two qualifiers: such functions shouldn't really be DLL's as much as they should be part of the kernel; and that such linking shouldn't necessarily be dynamic. Of course, "part of the kernel" is sort of a vague way to say it... in a microkernel system, everything would be some sort of library, and Win32 is actually sort of close to that idea, assuming you can imagine win.dll (or whatever it's called) as "the kernel". And the linking need only be dynamic the first time - like a JIT compiler. Statically linked copies could be cached for subsequent use. I don't know how much this would improve system performance, but I think it would be worth trying. I guess I mean that these functions ("compress" and "undo" to take the most overt examples) are so basic, they should actually exist somewhere below the file system, rather than above it. For example, what if you store a DLL in a folder which is actually a compressed archive? (Can you store the compression codec DLL itself in such a place? :) Making them optional components might actually be a drawback on the entire system, as for any given system you couldn't be guaranteed they'd actually be there... To tie this in with the licensing debate... I don't see proprietary software as a big evil. Not nearly as much as proprietary "standards" (of course, with friends like telnet, who needs enemies? :) And an OS is essentially a bunch of standards rather than a bunch of software - it doesn't matter how foo is implemented as long as foo does what foo is documented to do. I'm not sure about second-system syndrome. I mean I'm sure it exists, but to what extent it's unable to overcome the "barriers to entry"... I don't know. Desktop OSen certainly seem to be sloooowly merging towards POSIX... I mentioned that Micros~1 could take FreeBSD and relicense it, but I totally forgot that Apple has basically already done exactly that for MacOS X. POSIX is far better than nothing, but it's got it's own problems. I've seen software (I forget what exactly) that mentions in it's docs that "This program complies with POSIX except where POSIX is clearly brain-damaged," and that about sums up the usefulness of POSIX; as nice as it is to have something like it, it's *still* a stopgap measure pieced together by a committee... about as gracefully as HTML 3.2. Mmmm nummy nummy yummmm. It might be nice to know what the hell we want to use computers *for*, though, instead of just exalting in the throes of "wow, that's neat" hype. The problem with these machines is that they're just too damned general-purpose :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Thu, 12 Jul 2001 14:36:15 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > POSIX is far better than nothing, but it's got it's own > problems. I've seen software (I forget what exactly) that mentions in > it's docs that "This program complies with POSIX except where POSIX is > clearly brain-damaged," and that about sums up the usefulness of POSIX; > as nice as it is to have something like it, it's *still* a stopgap > measure pieced together by a committee... Yeah, which is why I kept saying "POSIX subset" the other day. Wasn't it Linus who said that Linux was as POSIX-compliant as he could make it without looking at a POSIX spec? > It might be nice to know what the hell we want to use computers *for*, > though, instead of just exalting in the throes of "wow, that's neat" > hype. The problem with these machines is that they're just too damned > general-purpose :) Don't worry, the industry is very shortly going to tear away at least one purpose (media authoring). Pretty soon they'll have these babies to a point where they're just as good as television. -R ------------------------------ Date: Thu, 12 Jul 2001 16:52:48 -0500 From: Chris Pressey Subject: Re: ESO (was Re: [www] Re: various) Russell Bornschlegel wrote: > Don't worry, the industry is very shortly going to tear away at > least one purpose (media authoring). Well, that Final Fantasy movie ought to demonstrate that well enough. Out of curiously, anyone know what kind of operating systems those Hollywood effects-artists are running? > Pretty soon they'll have > these babies to a point where they're just as good as television. But... where I come from, a television is not a piece of media authoring equipment. :) Chris -- This sentence will be false in 20 seconds starting..... NOW! ------------------------------ Date: Thu, 12 Jul 2001 15:23:29 -0700 From: Russell Bornschlegel Subject: Media Authoring (was Re: ESO (was Re: [www] Re: various) Chris Pressey wrote: > > Russell Bornschlegel wrote: > > Don't worry, the industry is very shortly going to tear away at > > least one purpose (media authoring). > > Well, that Final Fantasy movie ought to demonstrate that well enough. > Out of curiously, anyone know what kind of operating systems those > Hollywood effects-artists are running? I can't say for the people who did Final Fantasy, but I would guess that on the whole, computer graphics for movies are done on NT and/ or IRIX (SiliconGraphics Un*x), the former gaining market share and the latter losing. I know that the mid-to-high-range 3D packages like Maya are coming to Mac OS X as well. > > Pretty soon they'll have > > these babies to a point where they're just as good as television. > > But... where I come from, a television is not a piece of media authoring > equipment. :) I think you misunderstood; I was dripping sarcasm and bile, saying that the computer industry is going to _remove_ the media-authoring functionality from our computers (as a side effect of increasing copy protections); the end result of this progression will be to make computers as useful as televisions. Because, of course, consumers have no need, desire, or ability to create. -RB ------------------------------ Date: Fri, 13 Jul 2001 02:40:02 +0300 (EET DST) From: Panu A Kalliokoski Subject: Re: ESO (was Re: [www] Re: various) > Russell Bornschlegel wrote: > > > Even better, save new files as diffs of old files. And apply diffs to > > > make new diffs when old files are deleted. Kinda complicated, and > > > probably a bit slow, but very, very cool :) > > Do "journaling filesystems" do this? I haven't had a chance to learn > > about them. > > Possibly - I was thinking like CVS, although probably not as complicated > as CVS. Actually, closer to Bluetail's Wikie, a wiki written in Erlang; > it saves all additions as diffs to existing entries (I don't know if > other wikis do that.) CVS does not save new files as diffs of old files, but old files as diffs of new ones. That's much more sensible. In a typical programming project, where you save a file 100+ times before even making anything run, using diffs from old to new soon becomes impractical. > Hmmm... come to think of it, CVS itself could be abused as a chat > protocol, with full history keeping no less :) Well yes, except that CVS' locking is ill-suited for chatting. :) > Although it has very little to do with what we're talking about here, I > have to say that telnet is RIDICULOUSLY complicated for what it does... Those are just add-ons; telnet is mostly a straight pipe over the network. (Tough guys send their e-mail by telnet localhost 25) Panu ------------------------------ Date: Fri, 13 Jul 2001 20:05:38 +0300 From: Mooz Subject: [lang] [lang][Befunge] Dividing large numbers, floating point result This has got to be the silliest and most inefficient way I've seen division performed on a computer. Just as it is done with pen and paper. The program takes keyboard input in the form of two integers and a forward slash in between eg. "17/7" or "31337/42". The result is a floating point number, as accurate as it can fit in funge-space. The fun part? Take a debugger that can show the playfield, run it and laugh :) -----BEGIN BEFUNGE CODE BLOCK----- >1+>:~::"0"-:v>$:v>>21p :0#v_1-::::6g "0"-:0`!8v p +v_:#`0 #<_|_ " |-+77:+1g12_!v8!-"!":+gg12\+*+8< 6v 3>#$"9"-:v#<|_/ >$11g3+:6g4v->8+*+:"9"`"0"\-76*0p:v \| p#_!:`"#0_ ^"|" vp3p04+1:g0<^+2g11p+1g12\-*+55`"9"< ^< ^6+1p11:\_!-6#<^>1+:6g:84*#^-#4 _$$40g:"+"\:"|"v v59:+1p3+1\g6p12:<1:p2p3\< >*\2p:21g1+:11g-#^_$$31p70p11gv 2x v06:g22p124 p42 -\p22+1:g04p00::-3-< 3x >p80p >22g::3g:" "-!88#v+#*< 4x v g42_^#!:-1p22+1pg12\+<>+90p650p11v 5x >:23p 3g60g:1+60p21ggv ^77p08p06:,:v# >0`!#v_22g:1+22p :3g:" "-^++ >v 8x ^ _! @ >$$>#_ > 22g00g1^: 0g 9x $ v_v#`0_v#:-gg12g06 gg05 : < 51 v_|#-* 48 :g <_v>60g1+60p1+:70g- #^_^+ > #\ $^$ >50g1 +50p50g90g-1- !#v_^ 5 >v>50p:!#^_50g:vv6 8-6g05< # < !p0< :2^+1\-1\p1\"0"<>* +22g1-1p!#v_ :50g1-g> 80g:1+80p21g1+pv +1^g22:: < v_ #` #g13g2$#2!#<^ # ^_v#-" ":g-1g05: +1< 1g >64*-#^_$| >$ 60g21g2+p v >$$80p22g>1-::60p:21v p2 ^p12:+2g12p 08p06:\-# #< ^vg+1g12\gg< +>+g"0"-!#v_:22g\- 00g\^>g21gg-:!#^_0`90g\-90p^ >:" "-!#v_v ^2g12\" ":<$!_^#!-+88-gg12-1g06 :+"0"< p080_v#<`0+1:-g08- < > #2g#\1#<-:1g\>1-:31g-"."\#|_\:< ^++55p081< ^10 :$$ Subject: Re: ESO (was Re: [www] Re: various) On Thu, 12 Jul 2001, Chris Pressey wrote: > Russell Bornschlegel wrote: > > Pretty soon, you're gonna find yourself arguing in favor of COM and > > reinventing Win32. > Well, let's not get carried away, here :) > In a sense, yes. But with two qualifiers: such functions shouldn't > really be DLL's as much as they should be part of the kernel; and that > such linking shouldn't necessarily be dynamic. If I may be so bold, it seems like the way to do this is to use "DLLs" that are...how to phrase this...only dynamic on startup. Kind of like the OSF/1's "Pseudo-Static" runtime linker. Basically, you've got yourself a global lookup table. If an API happens to be supplied by the kernel, its address gets loaded into the lookup table at startup. If not, it enters the table when linked. Applications don't know the difference, because external function calls are hashed into the table and the link followed. Of course, everyone knows what happened to OSF/1, I guess... [...] > And the linking need only be dynamic the first time - like a JIT > compiler. Statically linked copies could be cached for subsequent use. > I don't know how much this would improve system performance, but I think > it would be worth trying. If you mean "between sessions" caching, usually not worth it. Caching "to the end of the session" usually works well enough, though. > I guess I mean that these functions ("compress" and "undo" to take the > most overt examples) are so basic, they should actually exist somewhere > below the file system, rather than above it. For example, what if you > store a DLL in a folder which is actually a compressed archive? (Can > you store the compression codec DLL itself in such a place? :) Making > them optional components might actually be a drawback on the entire > system, as for any given system you couldn't be guaranteed they'd > actually be there... I suspect Russell described it that way, not for the purpose of not installing codecs one doesn't want, but rather for future expandability. Which is good, since it'd be inconvenient to have to install a new version of the operating system every time some fool publishes a new encryption algorithm. > To tie this in with the licensing debate... I don't see proprietary > software as a big evil. Not nearly as much as proprietary "standards" > (of course, with friends like telnet, who needs enemies? :) Quick aside: Telnet is technically simple as one would imagine it should be: Send data, get reply. However, there are a million and two optimizations (for example, Nagle's Algorithm) that make it excessively complicated. [...] > It might be nice to know what the hell we want to use computers *for*, > though, instead of just exalting in the throes of "wow, that's neat" > hype. The problem with these machines is that they're just too damned > general-purpose :) It's always been my understanding that "wow, that's neat" *is* what the things are for. Well, that, and for convincing people to spend thousands of dollars to replace smaller, more convenient devices around their desk and home... ------------------------------ Date: Fri, 13 Jul 2001 14:55:07 -0700 From: Russell Bornschlegel Subject: Re: ESO (was Re: [www] Re: various) John Colagioia wrote: > Quick aside: Telnet is technically simple as one would imagine it > should be: Send data, get reply. However, there are a million and two > optimizations (for example, Nagle's Algorithm) that make it excessively > complicated. It's not the optimizations that Chris was whinging about, it's the plethora of silly options. It's easy enough, of course, to implement telnet in such a way that it says WON'T to all the DOs... -R ------------------------------ Date: Sat, 14 Jul 2001 09:10:33 -0400 (EDT) From: John Colagioia Subject: Re: ESO (was Re: [www] Re: various) On Fri, 13 Jul 2001, Russell Bornschlegel wrote: > John Colagioia wrote: > > Quick aside: Telnet is technically simple as one would imagine it > > should be: Send data, get reply. However, there are a million and two > > optimizations (for example, Nagle's Algorithm) that make it excessively > > complicated. > It's not the optimizations that Chris was whinging about, it's the > plethora of silly options. It's easy enough, of course, to implement > telnet in such a way that it says WON'T to all the DOs... Hmmm...Let's see...Telnet is RFC...854, which you can find at: http://www.graphcomp.com/info/rfc/rfc0854.html for those that feel like reading along. Blah, blah, blah, NVTs, skidah, skidah, skidah, half-duplex, yadda, yadda, yadda, "Do, don't, will, won't" handshake, some technical details that shouldn't be specified, ah. Options. Nope. "Control characters." IP, AO, AYT, EC, EL. SYNC, DM, STOP (for non-telnet interaction). Character set specs (also not good in the spec, in my opinion), BRK. Command sequence for setting options: "Interpret as Command" (IAC= 255), command (240-254), optional third byte if command was one of {DO/253, DON'T/254, WILL/251, WON'T/252}. Options aren't listed here, oddly enough, though. OK, on to RFC855 (finding the webpage, based on the above data, is left as an exercise for the reader). "Telnet Options Specification." That seems promising. Uhm...apparently, there ain't none at this time. Conclusion: All "silly options" are nonstandard, and no telnet implementation is required to support the, except insofar as knowing how to send a {IAC-WON'T-