Picking a license

[Patrick Maupin]

On May 14, 2:26 pm, Paul Boddie <p... at boddie.org.uk> wrote:
> On 14 Mai, 20:36, Patrick Maupin <pmau... at gmail.com> wrote:
>
>
>
> > That statement was made in the context of why Carl doesn't use GPL-
> > licensed *libraries*.  He and I have both explained the difference
> > between libraries and programs multiple times, not that you care.
>
> Saying that GPL-licensed applications are acceptable is a minor
> concession to the use of copyleft licensing if one advocates
> permissive licensing for all things which are not perceived to be
> finished products: things that one isn't looking to re-use somehow.

I am only "advocating" to the extent of explaining why I license stuff
permissively, and why, whenever I incorporate other stuff, it has to
be licensed permissively as well.  How you license your stuff is your
business.

> Saying that one likes Octave and that it uses the GPL

I don't recall saying I liked Octave.  I have no opinion.  I have
never used it.  Just threw it out there as an example of how people
rewrite proprietary software to counter your indignation that somebody
would rewrite PyQt.

> damning it with faint praise if one were then to say that its parts
> should be permissively licensed so that one can incorporate its
> functionality into something else.

I'm not operating a chop shop.  Never even had cause to look at the
source.  I really don't care about it.

> No, I don't care if you have a
> problem with GPL-licensed libraries because it is, as we have
> established repeatedly, your problem not mine.

Sure, the problems that I see with the GPL lead me to choose non-GPL
solutions for libraries.  And I never asked you to care.  I had a
brief moment of hope that you could see that my concerns were valid,
if personal, but apparently you can only concede that if you attribute
some sort of selfish ill-will to me.

> > I personally can't see any realistic chance of detriment.  How could a
> > proprietary clone hope to compete against free software on one side
> > and real matlab on the other side?  That's a no-win position, so I
> > wouldn't expect to see any proprietary clones.
>
> Well, only permissively licensed software would encourage such clones.

See, there you go with choice of language again.  Remember, we're both
biased with different viewpoints.  You say "encourage"; I say
"allow."  I further argued that it's immaterial that it's allowed,
nobody sane would do it.

> At that point, there are incentives for people to develop
> functionality for proprietary deployment instead of for the upstream
> project.

What incentives?  The incentives that the original matlab team will
keep outcompeting you from the top, or the incentives that the open
source octave team will keep outcompeting you from the bottom?

> [PySide and proprietary software]
>
> > No, PySide is about non-GPL software, and is released under a license
> > that even RMS recognizes as "free", and it is certainly not of
> > marginal utility.
>
> No, PySide is about permitting the development of proprietary
> applications by providing a solution to the all-important "ISVs" which
> lets them develop and deploy proprietary software.

That's an interesting viewpoint.  Originally, both Qt and PyQt were
available dual-licensed under the GPL or proprietary licenses.  For
anybody serious about proprietary development, the proprietary
licensed versions were actually quite reasonably priced.  Really the
major advantage I see in PySide licensing is for somebody like Ed or
Carl or me, who simply wants to be able to deliver programs with no
strings attached.  That was not possible under the GPL-licensed
version (because of the strings attaching to the customer that Ed has
talked about) or the commercial version (because then Ed couldn't even
give his customer the source to PyQt).  PySide is LGPL, which Ed still
might not touch, but at least any "linking" required between that and
code that uses it is really just an import statement, so then again,
he might be OK with that.

Of course, the fact that Qt and PySide are now both free of cost for
non-GPL customers certainly helps Nokia in their push to get people to
take them.

> Do you really think
> a platform vendor whose "ISVs" routinely ship proprietary software on
> their platform and on other platforms, and who will demand the ability
> to continue to do so, now expects all these "ISVs" to provide their
> applications under the modified BSD licence?

Not at all.  But they have now *enabled* ISVs to do that.  Before,
with QT and PyQt, it was GPL or proprietary.

> Sure, other developers
> can use the software - even people releasing GPL-licensed software -
> but that is highly unlikely to be the primary business motivation.

I think the motivation was to remove all impediments to using it on
the platform, and I see nothing wrong with that motivation.  They
already spent a lot on Qt, and they really want to leverage that.

> If
> you think the mobile telephony vendors are a bunch of fluffy bunny
> rabbits playing with each other in sugary meadows of niceness, I don't
> want to be present when someone directly and finally disabuses you of
> this belief.

I don't recall writing anything that would give you that impression,
but whatever.

> It's all about people selling stuff to "consumers" over
> and over again, preferably with the "consumers" rarely if ever being
> able to opt-out and do things their own way.

Nokia is locked in a tight battle with multiple players.  Developments
like Nokia using Qt and PyQt (which give small developers a chance to
play easily on that platform) and Google handing out Android to
whoever wants it are going to make the selling over-and-over
increasingly harder to do.  (Not that they won't do if if they can.  I
just think they are increasingly going to have a hard time playing
that game.)

>
> > > (And PyQt is not "marginally useful" - it is a widely-used and widely
> > > well-regarded library.)
>
> > Well, we agree on that.  But I don't know why you're trying to claim I
> > said PyQt was only marginally useful.
>
> Because you followed on from writing about PyQt by introducing the
> topic of "marginally useful" libraries, thus giving the impression
> that you regarded PyQt as "marginally useful".

Well, it's pretty obvious to anyone paying attention that PySide was
written for the sole purpose of creating something with similar
functionality as PyQt, but under a permissive license, so when I wrote
"nobody's going to waste time rewriting a marginally useful GPLed
library just to put a permissive license on it, either." I *obviously*
was explaining that projects which *aren't* marginal, such as PyQt and
MatLab, are the *only* kinds of projects that would be rewritten for a
simple license change.  You really should slow down and read a bit
more carefully.

Regards,
Pat