[License-discuss] License which requires watermarking? (Attribution Provision)
rick at linuxmafia.com
Wed Dec 19 21:40:46 UTC 2012
Quoting Richard Fontana (rfontana at redhat.com):
> Actually, section 7 of GPLv3 was intended to allow a limited form of
> badgeware (as well as certain other kinds of restrictions). But the
> example cited by the original poster:
> goes well beyond what the FSF intended to authorize.
> Unfortunately, I have seen a number of such misuses of GPLv3/AGPLv3 7(b).
Getting back to what I was groggily trying to say last night: My sense
is that OSI's approval of CPAL back in '07 was motivated in part by a
perception that a modest badgeware requirement was one arguably
reasonable method for giving reciprocal licensing enforcement power in
ASP/SaaS deployment, and that Socialtext's CPAL proposal wasn't so
extreme in its requirements as to preemptively kill third-party
commercial competition the way badgeware licensing usually does (the OSD
#3 concern I cited).
Maybe FSF felt similiarly (that 'a limited form of badgeware' was
sufficiently harmless and worth trying in that sense).
By contrast, the _classic_ badgeware we'd seen up to that point was
typified by SugarCRM's, which invented (or at least popularised)
competition-killing ASP licensing in outraged reaction to
Bangalore-based vtigerCRM's emergence as an independent fork of SugarCRM
1.0, in 2004 when the latter was under MPL 1.0. SugarCRM, Inc. designed
its new licensing (my interpretation) to cripple any subsequent forks'
commercial potential. And a gaggle of imitators then followed them.
As the saying goes, that's certainly their legal right, but it just
isn't open source. Much of the newer ASP pseudo-open source code we're
seeing has eschewed CPAL in favour of far more extreme badgeware of the
'mandatory advertising on every single UI screen' variety. Which, I
conclude, shows that proponents are not merely seeking preservation of
'attribution' as they have tended to claim in this forum: If they had
been, then CPAL's modest and minimally intrusive requirement would have
sufficed even for ASP/SaaS deployments where notices in the source code
may not be visible. Thus, it appears what they are really seeking under
the guise of 'attribution' is spiking of competitors while still
claiming to do open source.
So, you say OSI, FSF, and (possibly) Debian's approval of CPAL was
wrongly decided, and you might be right -- but my point is that its
approval and subsequent non-adoption has at least had the virtue of
making clear that the assertion about merely wanting 'attribution' fails
the sniff test.
Also, a larger point: OSI-approved licence Foo modified by a bunch of
restrictions dangled from it is pretty much reliably NOT open source.
Cheers, Nothing's hotter than having a copyeditor correct your sex scenes.
Rick Moen -- Max Barry
rick at linuxmafia.com
More information about the License-discuss