discuss about new license : PAL

cityhunter x-y-z at laposte.net
Tue Aug 5 13:48:37 UTC 2003


Don Jarrell wrote:

>First, let me mention that I am not an attorney, so I
>take only a broad, practical view of licenses such as yours.
>
>After reading your license, I would strongly urge you to 
>use one of the existing licenses, such as those written
>and published by those "experts" on this list (OSL, AFL, 
>ASL, etc) or the GPL.
>
>My gut reaction is that you are deviating from the Open 
>Source Definition *considerably* by trying to keep control
>of the later development by licensees, and by putting
>restrictions on roles and operations of maintenance. 
>
when you say libpthread will be a posix compliant thread library you're 
then against opensource, idem for glibc is a POSIX_C_ librairy and not a 
coffee maker
I only try to prevent project overgrowing, so developpers can take a lib 
there, another one elsewhere... this a quality assurance
I've a project called libwavelet (a wavelet is a mathematic transform 
such as fourier that can be used in image processing and sound, there 
are a lot of wavelets) Is it too much to ask that this librairy do only 
wavelets transform?
take the exemple of imlib :
once ago I was looking for a library to make only IO to image files (ie 
a library like imlib, imlib2, imagemagick.... but without all the 
processing ) I can't find one of them.... finally I used imlib but was 
very sad because using a library that was 10 times bigger that my own 
project.....
do you think this is quality insurance?
imagine that imagemagick do IO faster that imlib but imlib as better 
quality.... if imlib creator has said that imlib should only do image IO 
as imagemagick creator then we can choose the imagemagick IO and the 
imagickprocessing beacause these huge libraries would have been splitted in
imlib : do only IO
imagemagick : idem
imlib-imageprocess : do image processing (resize, FX...)
imagemagick-imageprocess:

licensees are not limited in further developpement since they can 
porvide additionnal libraries that come and enhence my project, but 
those library are _independant_, so end user can choose the best, if 
someone comes with a faster and better wavelet processing libraries, 
user can simply forget about my work, keeping ALL THE REMAINING of their 
project intact.... so everyone is happy :
-> the creator because some project has overcome their own project, and 
his own project was usefull
-> the "integrator" (the user that use the package to make a bigger one) 
since their sofware is easely updated and rotten part of their work removed
-> the end user because it has small libraries that will keep memory 
usage low, small packages such as gif2png won't oblige them to install 
the 10G lib required only to make IO (so a use of 0.000000001% of the 
library), the programs run faster (less time to read the package, libs 
and do linking)
-> the end user again because the XXXX package use only required part 
and not 1% of lib 1 (10Mo), 4% of lib 2(4Mo), 50% of lib 3(100Ko)....
simple calcul 100Ko+80Ko+50Ko=230Ko against 10Mo+4Mo+100Ko=14.1Mo.... 
all because once overgrowed a project can't slim.... so keep it sleem 
keep it fast keep modularizable.

QT don't suits me because it redo all the OS work, as gnome do....  I 
find fluxbox, openbox and other friendly smal WM that don't give me the 
impression that my pc is outdated. QT,GNOME,KDE all are overgrowing 
projects, you can't use a part of them, it's all or nothing....
conclusion NOTHING.

note that I don't prevent people forking and improving my work (although 
I prefer people joining or, if i'm unreachable, retake my project....)
take this exemple :
I've improved swapd (the linux swap deamon) which as some major bugs and 
wich is no longer maintained (can't reach the author), but I don't want 
to create another package with another name : this is not my ideas, 
neither my code at the beginning, and user won't know of my package and 
then continue the old buggy one.
so I asked sourceforge to create a project of the same name and will 
submit my modification there. no imagine the old author sudenly want to 
improve the work, come back, see that the project was taken away and 
when asking to have the project back I say "fuck you this is now my 
work, you where not there you have no rights and the new one", the old 
author can sue me ok  but he has no money to do so and since i'm not in 
his country he won't.... he can retake the things after a long painfull 
way....
I don't want such things because it can harm the enduser :
take the exemple of sorcerer (linux distro) and sourcemage : this is the 
best exemple of what I want to avoid (this is a exemple of the 
maintainer being ejected whithout reason) when I begun using sorcerer I 
asked kyle why sorcerer isn't Public Domain and what is sourcemage. 
after some discussion with boths parties, what kyle said was entirely 
true (even his predictions :) : he told me that they will say that he 
was too restrictive, don't listen....) I use sorcerer for a year now, 
making addition writing spells.... and kyle is one of the best project 
manager I know about : listen, improve....

>If 
>you really need to establish these working responsibilities 
>and conditions, you may want to consider a formal (negotiated 
>& signed) agreement.
>
cool I don't even know the people that use my soft, their mothertong, 
and I DON'T WANT
I try to improve the things make them better for the best.

>Also, I have issues with some rather unclear items that
>I would defer to the attorneys on the list to clarify, 
>such as the apparent confusion between code that is licensed 
>under these terms and that which is in the Public Domain.
>
? don't understand, please clarify : for me all is in Public Domain, the 
liscence try only to make the  succession maters clearer.

>
>Lastly, your provision that licensees work with each other 
>in a "friendly fashion" or lose their rights _may_ 
>violate the non-discriminatory requirements of the OSD, 
>but it certainly would prevent me from using the project
>code and license, since what you might think is not a 
>"friendly fashion" I might call normal behavior.
>  
>
I don't know I thought it as a good idea.....1

--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3



More information about the License-discuss mailing list