Request: Summary of GPL Problems
Andrew C. Greenberg
werdna at mucow.com
Tue Nov 13 21:11:48 UTC 2001
On Tuesday, November 13, 2001, at 05:00 AM, goran.hultgren at bluefish.se
wrote:
> A lot of OSS developers like GPL and if they can't use Squeak to write
> GPL programs that is
> a potential loss in the future for the Squeak community I think.
Perhaps, perhaps not. A lot of OSS developers also don't care, so long
as it is open source and generally free (per dictionary, not FSF), as
Squeak-L certainly is. The lack of GPL has not hurt a number of
exceptionally popular and successful open source projects, including
Apache, Python, FreeBSD, and otherwise.
But the real problem is deeper. GPL just doesn't work with monolithic
image systems, for the reasons previously stated -- it has the effect of
writing out of the license the (known to be essential) exception for
independently written programs. At least not as RMS interprets the GPL
(I do believe there are strong legal arguments to the contrary). In
part for this reason, no GPL'd Smalltalk system has been successful to
date, nor can it ever succeed, again, for reasons previously stated.
And there isn't much hope to those who agitate for a GPL changeover,
notwithstanding these problems. Apple certainly will never agree to
broaden limitations substantially beyond, say, the APSL, which FSF
actually changed their definition of "free software" to facilitate
panning it. (I remembered the colloquy after OSI approved APSL, RMS
published his criticisms claiming it was non-free. When someone noted
that the definitions on the FSF website didn't support his argument, do
you think the argument or the published definitions changed? So much
for principle.)
At any rate, so long as Squeak-L remains arguably non-GPL compatible
(and the problem really is predicated more on the fact that the
cohabitation language in GPL works on the OS/application model, using
assumptions that fail for monolithic images), this just isn't going to
happen. Squeak is highly likely to abandon the image model (indeed,
current thoughts seem to be moving along the line of a shared net-based
uber-monolithic model), even if we could shed the license.
It seems to me, indeed, that the GPL community has suffered a loss
because of its inability to use Squeak. But the broader open source
community seems to have accepted it warmly, and it is only a few who
seem to be having difficulties.
Mind you, I do support attempting to get Squeak-L more in line with
traditional open source licenses, shedding the proprietary font
provisions, and modifying the indemnity provisions to the compromises
that have already been negotiated in MPL, APSL and so many others, which
will bring Squeak-L to the level of OSI-compliant.
More information about the Squeak-dev
mailing list
|