Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther wrote:
> Ok, find attached the new ocaml licence proposal, which will go into
> 3.08.1 release, which is scheduled for inclusion in sarge.
> As said previously, it fixes the clause of venue problem, and the
> 6c problem.
FYI, I think this is fine.
I'm still uncertain whether QPL 3 is a freeness problem; but at the
moment I'm inclined to say "it's free, but it still sucks".
There is a sense in which it does impose a fee on people who distribute
modifications. But the "fee" is strictly a restriction on distribution
of modifications. Distribution of modifications is illegal unless a
license is granted, so it isn't actually taking away anything which the
modifier could otherwise do -- so in that sense it seems like it's not
really a fee.
It doesn't actually fail DFSG 3, since the modifications can be
distributed under the same license as the initial code -- however, that
license is discriminatory.
It gives more rights to the "initial developer" than to anyone else.
Now we allow copyright holders to discriminate between recipients in
dual-licensed materials and so forth, as long as everyone gets their
"DFSG rights". But this goes further -- it effectively requires the
creators and distributors of modifications to license them in a
discriminatory manner, giving the "intial developer" -- but nobody else
-- relicensing rights. However, such relicensing rights are rights
beyond the "DFSG rights".
I don't like it one bit, since it's "grabby". But I'm currently
thinking that it is free.
(It's also definitely not GPL-compatible, of course.)
Let's go abstract for a moment.
We accept no restrictions on use, and very few restrictions on creation
of modifications. Those areas have been pretty well hammered out.
We accept more restrictions on distribution (modified or unmodified).
This is a restriction on distribution of modifications. Specifically,
it is a restriction on the licensing of distributed modfications. The
real question here is "Is this an acceptable restriction on the
licensing of modifications?"
We accept some such (copyleft), have one explicit requirement (DFSG 3:
must allow modifications to be distributed under the same license), and
quite frankly haven't encountered many others, which is probably why
this is a contentious issue.
The history of DFSG 3 might help in understanding this. It prohibits
two sorts of licenses:
(1) a license which requires modifications to be distributed under more
restrictive terms, or requires extra permissions
(2) a license which requires modifications to be distributed under
*less* restrictive terms
Was (2) deliberate and clearly thought about? If it was deliberate, the
motivations behind it might indicate that forcing modifications to be
licensed in funny ways was objected to in general, which would be a good
argument that the QPL chicanery was intended to be considered non-free.
If it wasn't deliberate and everyone was thinking about (1), then it
doesn't indicate anything. :-(
The old Netscape Public License had a similar issue.