Re: Thoughts about gnuplot license and copyrights

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

Lars Hecking wrote:
> Dave Denholm writes:
> [...]=20
>  =20
>> It may prevent the gnuplot terminal drivers from being
>> distributed with another program. (Maybe not : it says you cannot
>> distribute complete modified source tree, but doesn't see you cannot
>> distribute a partial set of unmodified sources..?)  =20
> This only serves to demonstrate how badly thought out the license is
> in the first place. Along with your cvs example further down.=20
Agreed.
>
>  I must add that getting permission for a release, while it sounds like
>  a big thing, was never a problem - my ipression was rather that tw is
>  happy enough to see ongoing development. There waws a discussion about
>  backwards-compatibility before the 4.0 release, due the syntax changes=
,
>  but that was resolved quickly (the 4.x branch will continue to accept
>  the old syntax).=20

>  Once this discussion is settled, I will write up a summary and raise
>  all the valid points with tw. It would be nice if at the same time I
>  was in a position to tell him we're closing in on a new release :) =20
That is a wise decision.
So, it would be nice to have the point of view of other developpers.

...

Dave Denholm a =E9crit :
> Lars Hecking <lhe...@us...> writes:
>
>  =20
>> Dave Denholm writes:
>> [...]=20
>>    =20
>>> It may prevent the gnuplot terminal drivers from being
>>> distributed with another program. (Maybe not : it says you cannot
>>> distribute complete modified source tree, but doesn't see you cannot
>>> distribute a partial set of unmodified sources..?)
>>>      =20
>>  This only serves to demonstrate how badly thought out the license is
>>  in the first place. Along with your cvs example further down.
>>    =20
Agreed.
>
> Actually, another program could include the *entire* unmodified
> gnuplot source distribution, and compile just the terminal drivers...
>  =20
Ok, it's possible, but to remain pragmatic, it's far from being practical.
>>> Maybe one way out is to persuade Thomas Williams to grant permission
>>> to a trusted nominee to release as and when required. Thomas would
>>> retain the option to revoke that permission at a time of his choosing=
,
>>> but it does mean that if he does vanish, there is one other person ca=
n
>>> release..
>>>      =20
>> =20
>>  I must add that getting permission for a release, while it sounds lik=
e
>>  a big thing, was never a problem
>>    =20
>
> One day, Thomas Williams *will* become unavailable, one way or
> another. One hopes that he will be around for a long time to come, but
> no-one is immortal.
>
>  - my ipression was rather that tw is
>  =20
>>  happy enough to see ongoing development. There waws a discussion abou=
t
>>  backwards-compatibility before the 4.0 release, due the syntax change=
s,
>>  but that was resolved quickly (the 4.x branch will continue to accept
>>  the old syntax).
>>    =20
> I trust no-one thinks that he is trying to block development. It's
> just unfortunate that this is probably the only way to avoid a fork.
>  =20
I am happy to know that Thomas pays attention to gnuplot, even if he=20
doesn't participate actively anymore. Unfortunately, Dave is right, he=20
can someday give away with his "responsibility" to bless new releases,=20
just by being unreachable for example.

> [...]
>>> could it be said that
>>> having the CVS available on the internet is a violation of
>>>
>>>  * Permission to modify the software is granted, but not the right to
>>> * distribute the complete modified source code.
>>>      =20
>> You're probably right. I assume that cvs is used with the private
>> agreement of Thomas.
>>    =20
> I'd assumed not, but maybe it gives a way out...
>
> gnuplot predates sourceforge and all the other hosts of open-source
> projects.
>
> If Thomas has (or will) sanction the cvs on sourceforce, then that
> gives us a working definition of which is the official master set of
> sources for gnuplot. So I wonder if we can open things up a bit, yet
> still prevent forking, by defining something in terms of that CVS.
>
> ie a released version of the gnuplot sources corresponds to a labelled
> version in cvs, Modifications can only be distribed in the form of
> patches to a specific labelled version. etc, etc.
>
>
> Hmm - it may not help if it ends up that only Thomas can do the
> labelling :-(  Maybe only he can label a major release, but the
> maintiners can independently label a minor release ?
>
>
> Not sure how that can be put into legalese, but...
Ethan wrote :

> Actually, I believe that distribution by CVS is exactly in line
> with that requirement. A CVS server, internally, is in fact a large
> set of sequential patches applied to a base version.  That is what
> lets you back-track to any previous date's version.  The fact that
> the server is capable of applying the patches for you on the server
> side, as well as handing you the base code and the separate patches,
> is a matter of convenience rather than substance. =20
Ok, so in some way cvs can be "legally" considered as a valid release way=
. IMHO it's not in the "spirit" of the copyright, and anyway it still pre=
vents from doing a standard package... Moreover, defining the license in =
the terms of a technology such as cvs is dangerous : for example, AFAIK s=
ourceforge will in the future switch to subversion...


Regards,

Timoth=E9e