Re: New version of articulate available

[Graham Percival]

On Sun, Mar 20, 2011 at 04:23:12PM +0100, David Kastrup wrote:
> Graham Percival <address@hidden> writes:
> 
> > The suggestion that a .ly file would somehow be a derivative work
> > of lilypond is ridiculous.
> 
> Depends on how interlocked and crossdependent it is with internals of
> Lilypond and whether or not stuff has been cross-copied.

If there's no allowances for "interoperability", and if the amount
of "interlocked-ness" (how do we measure this?) of articulate.ly
means that it's a derivative work, then any serious use of scheme
functions in lilypond would automatically mean that the music must
be GPLv3 or later.

That's crazy.  If that's actually true -- which I doubt -- then I
would argue in the strongest possible terms that we should add a
"using the lilypond scheme API does not require that the music is
placed under the GPLv3", similar to our font exception.

My personal stake is that I'm using scheme to extract music events
from lilypond for Vivi.  I was planning on placing Vivi under the
GPLv3, but I *don't* like being forced to do so.  If using
lilypond scheme actually means that -- and if we don't add an
exception to allow the use of scheme code and calling
ly:music-functions in our own .ly files -- then I'll seriously
look at dropping lilypond input and use musicxml instead.


> A .ly file that represents some score certainly is separate from
> Lilypond.  A .ly file that is intended to run as an integrated part of
> Lilypond when typesetting, however: I would not be able to call that a
> separate work without analyzing the source.

Please examine articulate.ly in detail and give your opinion about
whether it is legal for Peter (and NICTA) to place that work under
the GPLv2.

This is a very serious allegation, and I think we should clear it
up immediately, before even thinking about any patches.


I will admit that one comment in articulate.ly says:

% Gradually speed up a piece of music. Stolen from the feather
% code in
% the Lilypond base.
% Overflows moment and causes infinite Lilypond loop, or segv
% --- DON'T USE
#(define (ac:accel music factor)

Since articulate.ly was primarily written in 2008, I would expect
this to have come from the GPLv2 version of lilypond (as it was
until Fall 2009... actually, the stable 2.12 version is still
under GPLv2).  And it that "DON'T USE" comment is accurate, then
perhaps the entire function should be removed to avoid confusion.


> > articulate.ly is a 668-line .ly file containing a bunch of scheme.
> > It is absolutely not a derivative work of lilypond.
> 
> That is not the question.

Isn't that precisely the question?  You wrote:
"It is not even clear that Peter can release/distribute it under
GPL version 2.0 unless it will work unmodified with a version of
Lilypond released under GPL version 2.0"

If articulate.ly is not a derivative work, then he (and/or his
employer) are free to choose any license they wish.  If, for some
reason, articulate.ly *is* a derivative, then he (and/or his
employer) are *not* free to choose any license.


At the moment, I don't care about the patch.  I'm shocked at the
suggestion that Peter's research might be illegal, and I would
like you to clarify this as soon as possible.

Please example articulate.ly in detail, and give your opinion as
to whether it consititutes a derivate work.

- Graham