In article <5gosup$bc6 at news.u-strasbg.fr>, pingouin at crystal.u-strasbg.fr (Francois JEANMOUGIN) writes:
> Regarding the survey we start some times ago on GCG performances,
>we can say that most of sites are running GCG8 because they are waiting
>for EGCG9 before installing GCG9, so that makes 4 months of licensing for
>an unuse software suite...
Today I'm writing the letter in which we officially stop GCG maintenance.
We will continue to run 8.1 under our existing license agreement, but never
go to 9.0.
It was unpleasant to receive 9.0 and not be able to install it (no local
programs, no EGCG), but GCG hemmed and hawed and promised an SDK, so we
waited (im)patiently, with 9.0 sitting on the desk unused. After several
months the SDK license agreement finally showed up, and, excuse me folks,
but you better not be planning on publishing anything that you develop with
the SDK, because you will not be able to redistribute the code or the
binaries under section Cii of that agreement My reading suggests that you
could not even distribute diffs. We simply cannot sign such an agreement
since most journals require that programs used in papers be made available
to other sites - and what is the point of developing programs using the SDK
if we cannot publish the results that derive from them???
This is section C of the SDK agreement:
C. Use of SDK Source Code; Use and Distribution of Derivative Works. GCG
grants to Customer a nonassignable, nonexclusive perpetual license:
(i) to adapt the SDK Source Code to create derivative works containing the
SDK Source Code (the "Derivative Works");
(ii) to install and run a copy of the Derivative Works on computer systems
licensed under the Customer's Standard Agreement and Customer's License
Extensions but Customer may not distribute, or allow the distribution of,
any Derivative Work to any third party.
(Emphasis mine). There is no wiggle room in there at all. Apparently the
EGCG folks have a different SDK license agreement that allows them to get
by Cii and distribute EGCG, but GCG is not extending this to anybody else.
I would not have been able to submit the programs that I did to EGCG under
this license, indeed, it is pretty clear that the EGCG would not now exist
if this license had been in effect in previous versions.
I can understand that GCG doesn't want their code to show up in other
commercial products. However, their handling of this whole affair is simply
1. They changed the license agreement without prior notification and
shipped what was essentially a nonfunctional product (9.0), since any site
with local code or EGCG could not upgrade.
2. They took months to produce an SDK license agreement, and when it
arrived it contained a clause that was contrary to the general ethics of
research, that is, that research material, in this case programs, be made
available to others as a prerequisite for publication. (In previous GCG
versions there was the quite reasonable situation that any GCG site could
ship a locally developed program to another GCG site, or at least, nothing
in previous licenses specifically excluded it. Shipping such a program
to a nonGCG site was pointless - they couldn't build it or run it without
the rest of the GCG package.)
3. Much of GCG is itself a "derivative" work. I wonder if Zuker, Pearson,
and the many others who contributed programs to the original University of
Wisconsin GCG would have done so if the current set of GCG license
agreements had been in place at the time. I suspect they would have, at
the very least, requested significant remuneration. Perhaps they will do
mathog at seqaxp.bio.caltech.edu
Manager, sequence analysis facility, biology division, Caltech