Thoughts about gnuplot license and copyrights

[<tim...@en...>]

Hi gnuplot developpers and users !

Recently, I have been thinking about the way gnuplot is licensed. As I
have worked a lot on the wxWidgets terminal (and am still working on
it), and as I am a strongly in favor of "libre" software, I wonder about
the future of this work...

On this mailing list, in september 2005, while talking about a possible
release of a 4.1 version, Hans-Bernhard Broeker explained :
" "People have got used to a release early --- release often" strategy.=20
Sorry, but with the licensing issues as they are, that's a non-option
for gnuplot. "

Well, "release often" is probably not the best idea, and I agree that
bugs have to be eradicated as much as possible before. However, I fear
that the situation of gnuplot tends to be : "rare release, cope with
CVS". And what will this evolve to when the people who hold the
copyrights cannot be contacted ?



With "grep -C 5 Copyright *" in the gnuplot source, I found that most of
the copyrights belong to Thomas Williams and Colin Kelley.

Some files are public domain (fit, matrix).
Most terminals don't mention who holds the copyright ("Copyright [a
year]") but specify an author later (should the reader understand that
they own the copyright ?).

Other files belong to :

Robert K. Cunningham (datafile),
Daniel Sebald (bin_hook),
Petr Mikulik (with the mention : "as open source as possible"),
Craig R. Schardt, Maurice Castro, Russell Lang (windows),
Hans-Bernhard Broeker (dynarray, windows),
Roger Fearick (pm),
Jos van der Woude (statistical functions, specfun),
Phil Type, Bruce Ravel, Gershon Elber (lisp),
Stephen L. Moshier (some code in specfun),
Lars Hecking (tables),
Ronald Florence (cgi.trm),
P. Klosowski (excl.trm),
James Darrell McCauley (grass.trm).




The copyright text mentions that it is not possible to distribute a
modified source code. So, who can give the permission ? Who gave it for
gnuplot 4.0 ? Who will give it for gnuplot 4.1 ? Who will give it later ?



I wonder if this situation has a future... What are the possibilities ?
To my (humble) mind, they are the following :

* Nothing changes concerning copyrights and license. CVS will be filled
by new (useful, powerful and innovative) code, and probably new files
with new people owning the corresponding copyrights. But one day, nobody
will be able to release officially and legally anything. gnuplot will
vanish in some backup disk at sourceforge.net   ;-)

* Something is changed on gnuplot copyright. What for ? Probably to
allow easier distributions. How ? Either by modifying the current
license, or by choosing a well-known and well-understood license
commonly used in open source programs instead. What does this imply ?
Probably to contact all developpers who own their respective copyrights,
and ask for their approval on the proposed change. Some of them disagree
or are not joinable ? Their code should probably be rewritten from scratc=
h.

Well, it's a huge decision, but what do you honestly prefer ? The first
case ? Probably not. But that might happen, really.
I don't see a valid reason to not *protect* the work of great value that
gnuplot represents.
Of course, gnuplot can also vanish with a more adapted license. But
that's not a reason.
Of course, in the early days of gnuplot, the gpl (as an example) did not
exist. But that's not a reason for not changing now.



What do you think about it ? What do Thomas Williams and Colin Kelley
think about it, as they own the copyrights for most of gnuplot code ?
Do you agree that something should be done to protect gnuplot ?

I hope that I have not raised a "too-often raised" problem, already
answered many times. I really don't want to provoke anybody. Please
excuse me for any inconvenience and for the noise.

Best regards,

Timoth=E9e Lecomte