In article <PMR.97Mar19161952 at unst.sanger.ac.uk>, pmr at sanger.ac.uk writes:
>> The standard SDK licence has restrictions on redistributing software
> which have to be amended for EGCG, so that will take a little longer.
I haven't seen a copy of the SDK license yet (though I hope to soon),
but I'm curious about this restriction. Is it operative because EGCG
programs contain GCG source code vertbatim (as applied to copyright
interpretation), or are they requesting that sites which purchase the
SDK accept restrictions on redistribution of their own code which
contains calls to GCG functions?
> As for this mysterious "sources were found in several (at least one)
> commercial package", that may not be quite true. The GCG licence
> restricts distribution of binaries linked with GCG libraries (that's
Does this restriction apply only to binaries linked agains the static
object libraries (e.g. applib), or to code linked against the shareable
image libraries (e.g. genlib)?
> why EGCG needs a special licence). It could be that someone just
> distributed binaries that were linked to their own copy of GCG. That
> does not mean the code would run without GCG. But it would still be
> illegal under the SDK (which allows you to create and run software,
> but not to distribute it).
Hmm. As you've articulated it here, it seems (to me, not a copyright
expert by any means) that it'd be difficult to support such a limitation;
it'd be roughly equivalent to, say, Sun Microsystems contending that
one could not distribute software developed on a Sun because it was
linked to Solaris libraries. One might justify a restriction on static
linking due to content of the binaries, but it seems quite a stretch
to request that sites refrain from distribution of their code based on
GCG's copyright interest in their APIs.
Regards,
Charles Bailey bailey at genetics.upenn.edu
