Re: GPL scripts with a GPL-incompatible interpreter
On Sat, Dec 14, 2002 at 02:37:26AM -0500, Anthony DeRobertis wrote:
> Section 2 says you have the option of changing that source code, and
> then distributing that changed code as in section one (i.e., in source
> form), so long as you note your changes, license under the GPL, and some
> stuff about what the program must display when it starts up.
This seems inconsistent. You deduce that section 2 is really talking only
about source code, because it says "under the terms of section 1".
But when section 3 says "under the terms of section 1 and 2", you interpret
that as meaning that non-source works can be distributed under those terms.
I think that section 2 extends the permissions of section 1 to "a work
based on the Program". It places no limit on the kind of work.
To interpret it otherwise, you'd have to explain why section 1's
"source code" requirement remains, even though "verbatim" is overridden.
> Section 3 says you can also distribute binaries, given that you mark all
> your changes, etc. (the terms of section 2), include source code with
> the relevant notices (the terms of section 1), and, in addition do one
> of the famous three things [machine readable with it, 3yr offer from
> you, 3yr offer from someone else].
You mean, mark all the changes in the binary files? I never thought of that,
but it does indeed say that. What happens if the binary file format has
no provision for prominent notices?
> > I read section 3 as being an additional restriction on modifications
> > that are also "object code or executable form". The open question is
> > then whether an executable script is in "executable form".
> It truly does not matter if an executable script can be called "object
> code" or "executable form." Notice that section three says "You may",
> not "If you distribute ... executable form..."
> If some part of the license has already given me permission to
> distribute, there is no reason I would use section 3's permission to
> redistribute. The section that already gave permission is section 2,
> because a script is source code.
Well, it also says "provided that" and imposes restrictions in addition
to sections 1 and 2, without offering additional permissions. So with
this reasoning I don't see why anyone would ever need section 3.
(The exception in the definition of "source code" only relaxes part of
the additional requirement to accompany it with source code; it doesn't
help if you are unable to distribute "the Program in executable form"
"under the terms of Sections 1 and 2" in the first place.)
> This doesn't open any holes in the GPL; only section 3 give me
> permission to distribute stuff that isn't source code. So your embedded
> system would either have to run completely off source code(!) or follow
> section 3.
It's clear that our basic disagreement is here. I see nothing in
section 2 that would limit it only to source code.
Hmm, maybe I'm misinterpreting "under the terms of". Does the GPL
mean that these terms should be _extended_ to recipients, rather than
that they are _imposed_ on the distributor? I.e. if you use section 2,
then you must follow section 2 when preparing the modified work, and
recipients get redistribution rights as specified in section 1. That
would explain why section 2 needs to explicitly state 2b.