[Biojava-l] Heads up: packages moving

Jason Stajich jason@chg.mc.duke.edu
Mon, 22 Jan 2001 13:26:03 -0500 (EST)


Mike - 

We'd certainly appreciate you donating the code.  We have been discussing
about how to make the OMG LSR => BioCORBA mapping work since last summer,
there are a couple of places where things might be difficult, but a number
of the proposed strategies seem to be doable.  If you are willing to at
least take a look at this, it would be very beneficial for the project.

I do think separating the biocorba-java cvs module out is a good thing, we
have actually separated bioperl-corba-client and bioperl-corba-server into
two separate cvs modules to not force a client-only user to install any of
the bioperl corba code to run the server.

I'm bogged down in getting the bioperl 0.7 release ready - once we get our
feature / locations issues settled and fuzzy feature/ multi-location
features are handled I'll bring the bioperl corba code up to the current
0.2 IDL spec.  

One key component to the biocorba project is to provide code that is
supported by all the bio language project (java,perl,python) so any bridge
that is done to support OMG LSR should be implemented in each language's
corba client/server code if possible.  We should probably start a wiki
page to this effect at some point...

Jason
Jason Stajich
jason@chg.mc.duke.edu
Center for Human Genetics
Duke University Medical Center 
http://www.chg.duke.edu/ 

On Mon, 22 Jan 2001, Matthew Pocock wrote:

> "Dickson, Mike" wrote:
> 
> > Regarding maintaining the corba work, I'd been considering donating a bunch
> > of the code we've done for the OMG LSR BSA to BioJava.  It would of course
> > be based on different IDL's than the current corba work.  My idea was to
> > write adapters that map the BSA Impl code to BioJava under the hood (for a
> > server implementation for instance) and also an adapter  that would map the
> > BSA interfaces to BioJava interfaces. This would allow a client to use the
> > BioJava interfaces and still talk to remote objects.  I believe something
> > similar to this approach was used in the current corba code.  Anyway, I'd
> > like qet an idea of interest in the work I'm proposing.
> 
> If you would like to donate the code, I can set you up with CVS access. The
> adapters would go into their own CVS packages. I guess you would be responsible
> for maintaining them. Interoperability with BSA has to be a good thing - what do
> the industry people think?
> 
> I don't see BioCorba and BSA as rival CORBA services for BioJava, esp if neither
> one is in the core repository. I think that if both are available, then people
> can chose the one that is most apropreate for a given task.
> 
> Matthew
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
>