Re: Software relicensing, how is it done ?

["Tristan Van Berkom" <tvb gnome org>]

(apperently my other email just now missed the list due to mailing from my
ordinary email address, here it is...)

Hi Guys,
  Theres obviously been some scrutiny concerning our decision to finally
relicense Glade or primarily, libgladeui - so I will try to do my best
to address your concerns and then share a little where I'm coming from
(remember I wouldn't be here in the first place if I didn't love you guys).

On Fri, Oct 31, 2008 at 3:21 PM, Hubert Figuiere <hub figuiere net> wrote:
[...]
> No yet another BitKeeper-like situation. We have seen what it does.
>
> BTW, there are already 6 IDEs that are Free Software: Anjuta, KDevelop,
> CodeBlock, Eclipse, MonoDevelop and Emacs[1]. So why wasting time to
> allow a 7th one that could be non-free instead of making sure the
> existing one rock even more.
>
> I'm very skeptical about the whole process of relicensing Glade to allow
> non-Free derivative of it.

I'm not 100% schooled on what exactly happened with BitKeeper, but my
base understanding is that the developers found ways to work around the
license in order to base a new work on free work, drop the free one
and only support proprietary extensions of BitKeeper ?

Without jumping to conclusions about the above statement all I can say
is that it deeply saddens me to think that its possible that I could be
suspected of such a treasonous plan, by people I respect and have come
to consider as my peers; as specially when I stand here practically
single-handedly responsible for delivering you freely a Glade 3 that
was little more than a prototype and a dream years ago.

If these are indeed the trust issues we are faced with in our
community, there's obviosly nothing I can say to put your worries
at ease.

On Sat, Nov 1, 2008 at 10:33 AM, Richard M. Stallman <rms gnu org> wrote:
>    I dont see how I can agree that entering in direct competition
>    with anyone who wants to make a dollar from a software solution is
>    going to bring us to that long-term goal.
>
> The GNU Project has a history of competing successfully with
> proprietary software.  For instance, GCC competed directly with
> non-free C compilers, and has done quite well against them.  And the
> GNU operating system as a whole has done pretty well against Unix.
>
> Any free IDE almost surely competes directly with non-free IDEs, but
> that is no reason to give up developing them, and I am confident our
> community will not.

Richard,
  We obviously dont share the same goals as a big picture, so I wont
try to pretend to.

While "They" may be playing a game of keeping secrets in an attempt to
cripple free software so that theirs is perceived as "better" - I cannot
sit and play the same game. My weak attempts to get corporate users of
free software to give back to the community will fall on deaf ears for
my obvious hypocrisy.

> While you may be most concerned with who makes how much money, I'm
> more concerned with advancing our freedom.

While I am sincerely greatfull that we have guys in the political
sphere and the PR world as well, I've prefered to stay silently
patient and write Glade, for exactly free, and so I will not indulge
in a meaningless argument about the above statement.


On Sat, Nov 1, 2008 at 4:48 AM, Naba Kumar <naba kumar gmail com> wrote:
[...]

Naba,
   I was not expecting, albeit not completely surprised by your
reaction, and even a little flattered that anyone would think that
Glade gave you a competitive "edge". I believe your success in Anjuta
and my own with Glade is based on patiently doing things correctly
and getting it right, never in a hurry to make a release for the
public eye, and with closed ears to criticisms and other momentarily
more popular or more successful projects.

> I am fully with Richard here. LGPLing libgladeui is essentially
> LGPLing 'the glade application'. Being a library doesn't change that
> fact, because it's mostly a means for free IDEs to integrate glade
> application, like Anjuta does.

We have never seen it this way - and no matter how hard we've tried
to express ourselves as a core library for the editing and serialization
of GObjects, obviously nobody is catching on, for instance - how come
there is *still* no Glade plugin to edit gstreamer pipelines ?
(*really* no offense to the gst-editor authors, I tried using that
tool a number of years ago and always asked myself, if I wrote a
tool to do just that, why dont they use it ?).

The plugins distributed with the full glade package *are* Gtk+ interface
specific and in your terms could be considered an "application of
the libgladeui library" i.e. applied usage of libgladeui in the context
of Gtk+ interfaces and Gtk+ widgets, I would prefer to think of these
plugins as the all important use case that libgladeui was invented for;
historically.

The license of those plugins dont really concern me, but I also dont see
why someone would want to create a Gtk+ interface editing program using
libgladeui, when such an application of the library obviously exists,
I also dont see much point in barring competition all together, Ive been
doing alot of work lately on the gtk+ plugins and I would love to see
someone challange me to do better, which is why I also think it would just
be all around simpler to just LGPL the whole thing.

> I am against going anything less than GPL for free software
> 'application'. My reasons being exactly what Richard pointed out -- it
> allows extending your applications with proprietory solutions (with
> plugins for example) making the application 'as a whole' not free
> software anymore. This has complicated results later. You won't be
> able to influence/see/change the non-free portions of your application
> and could mean significant resistance in controlling the direction of
> your application. Your 'free software' could merely become an
> instrument to replace itself with proprietory solution.

There is something importantly incorrect about that statement
that people need to understand, when you say that "You won't be
able to influence/see/change the non-free portions of your application"
you are forgetting that it is not *my application* that we are
speaking of, we are speaking of a hypothetical user of *my library*,
in which case I couldnt fathom perceiving that as any kind of threat.

As for the "replacing itself", these ideas I have to consider as
completely paranoid, I hereby single handedly challenge any think
tank to write a tool that uses libgladeui and rewrite a core to
replace it, without my own help. The libgladeui maintained by
myself, if even by virtue of being public domain, will still
rock like it always did, and I'm betting it will still rock more
than the hypothetical competition.

[...]
> Again I can't agree more. As Anjuta developer, I might feel
> biased/selfish here, but I think it's far more important to have it
> working well with free IDEs than it promoting proprietory IDEs. Right
> now glade+Anjuta integration isn't anywhere close to what I want. This
> stance will get worse if suddenly all the glade integration attention
> goes away to a proprietory IDE (if it was another free IDE, then
> nobody loses). You need to think of general 'eco system' as well, not
> just glade.

libgladeui/Anjuta integration is really something I always wanted
to improve, and regretfully I've always had to make glade work instead.

While it may sound rediculous, All I/we would have needed is one hacker's
weekends for one lousy year, obviously you were busy with Anjuta, I was busy
with Glade, and nobody else thought it would be cool enough to make the
effort - and it sure is sad.

All that being said, for years now we have had a vision of libgladeui as
a library, not rediculously bound to the creation of "interfaces" or "widgets"
but basic serialization of GObjects in general, we never got around to doing
the work to get it properly LGPL'd, as we've intended in the past and the day
has finally come that I need an LGPL libgladeui, using a LGPL'd libgladeui
will actually give me the opportunity to sit in the seat of the embedding
application of libgladeui and give some good attention to the highlevel
libgladeui api.

My alternative here is so ugly I would rather not even pronounce it. My aim
is to use libgladeui as LGPL, with the goal in mind that I can
continue to safely
share all my libgladeui related work with the community, so that users, plugins
and embedders such as Anjuta can continue to benefit from my work.

With that, I sincerely hope that this humble ghetto dawg has succeeded in
aliviating the worries of at least some.

Best Regards,
                -Tristan