[GMOD-devel] Tagging chado in cvs

[David E. <em...@mo...>]

Hi GMOD,

Colin and I checked a few chado schema changes into cvs late last week
after
conversations between ourselves and also with Chris.  Refer to the
Changes
document for details.

As many of you may know, we (FlyBase) have been working on migrating
the
D. melanogaster release 3.1 sequence and annotation from the gadfly
database
into chado.  After six months of work, both refining the schema and
developing
implementation conventions, we are satisfied that the chado schema as
it
presently stands is sufficient for this purpose.   We've put a cvs tag
on
the current chado schema, "flybase_migration_build", and our intention
is
to use this version for our final migration of gadfly data to chado.
Colin
will be producing a dump of gadfly using this tagged version in the
coming
days.  For those who may have been using Colin's last dump of gadfly
data
into chado  - chado_gadfly6 - we've produced a list of important
changes
which will be seen in the new dump, and we'd be glad to share this.

At the CSHL GMOD meeting in May, we discussed making a "freeze" of the
schema
for GMOD developers to use in developing the first GMOD system.  We
would like
to propose that the "flybase_migration_build" be tagged additionally as
"chado
version 1_01", and that this version be used for the first GMOD system
implementation.  The advantage of doing this would be that the gadfly
dump
which Colin is making could serve also as a reference implementation of
version 1_01.

Unless there are objections, we'll plan on creating tag "chado_1_01" at
the
end of this week.  I encourage people to review the production modules
(listed
in chado/modules/idb-full.modules) and let us know if there are any
problems.

We should note that while we all look forward to continuing chado
development
in the coming months (the genetic and expression modules at least are
slated
for review and probably major revision), developers need to be aware
that the
schema is in production at several sites, that schema churn produces
very real
expenses downstream, and that therefore any changes to core modules
need to
be discussed and agreed amongst core developers before they are
committed.

Best,

-Dave