Re: XFree86 4.4.0 RC3

[Torrey Lyons <torrey@xxxxxxxxx>]

Since we have been invited to express our opinion, I believe that the 
new XFree86 license does more harm then good. The goal of achieving 
equitable attribution is laudable, but this license is too blunt of 
an instrument to achieve it.

There has been a fair amount of emotion expressed over this issue and 
some of it could rightly be called FUD. The specter of GPL 
incompatibility has substantially muddied the waters due to long help 
political differences. However, there are legitimate problems with 
the new restrictions this license places on binary distributions.

At 3:50 PM -0500 2/13/04, David Dawes wrote:
> This begs a couple of questions that I have yet to see answered:
>
>   1. If other third-party contributions are acknowledged in, say, a CD
>      booklet, why shouldn't XFree86's be?

This is the core of the argument. All reasonable people would agree 
that XFree86 should be given appropriate credit, but this license 
places conditions on how and where this credit should be given that 
make it problematic for binary distributors. I imagine the position 
of the XFree86 BoD is that the conditions on binary distribution are 
minor and easily satisfied. However, I agree with others that there 
are many actual and conceivable circumstances where the burdens 
placed by the new license are overly restrictive. Some examples are:

1. The many contributor problem: Others holding copyrights in the 
tree have been encouraged to consider adopting this license. The 
attribution requirements could soon become unwieldy.

2. The multiple acknowledgements problem: The license presupposes 
that there is one place acknowledgements are gathered. The OpenBSD CD 
case has been pointed out where only graphics designers are credited 
on the CD case. The only way to satisfy the new license requirements 
appears to be to credit XFree86 everywhere 3rd parties are 
acknowledged. This places a burden on documentation and package 
design, which could be substantial combined with #1.

3. The single file problem: What if I only want to borrow part of one 
file with the XFree86 copyright? In the past I could use this for my 
proprietary or other purpose as long as I retained the 
copyright/license notice in the source code. Now I have to credit the 
XFree86 Project in any binaries I produce. In answer to your question 
above, sometimes XFree86's contributions may be too small to warrant 
being specially recognized. The requirement that credit has to be 
given for even tiny contributions has the practical effect of making 
XFree86 copyrighted code unusable on certain projects. This goes 
against the original idea of the MIT/X11 license.

The core idea of MIT/X11 license is that the code can freely be used 
by anyone. The new license makes the "barrier of entry" too high for 
some people and thus runs counter to the spirit many X11 developers 
are working under. The XFree86 Project is recognized in the open 
source community because it produces good code that people use, not 
because it widely requires credit. My view is that we should return 
to producing code that no one has problems using.

--Torrey
_______________________________________________
Forum mailing list
Forum@xxxxxxxxxxx
http://XFree86.Org/mailman/listinfo/forum