[Pipet Devel] more on Eazel

Brad Chapman chapmanb at arches.uga.edu
Wed Feb 23 09:04:58 EST 2000

Jeff wrote:
> Talks with Gary (for those who don't know, Gary and I met in person 
over a
> week's time to hash out some design issues) led me to start calling 
Loci a
> 'connectivity broker'.  This means that Loci will be agnostic to the
> communications protocol used by the back-end program.  CORBA, HTTP, 
> pipes, XML_RPC, etc. can all be supported if the support is 'plugged 
in' to
> Loci. 

This makes a lot of sense, and I don't want to make Loci 
CORBA-specific. I just think it might be a good idea to start with 
corba connections to the backend and then work from there to add 
other protocols, etc. I think there are a couple of advantages to 
using CORBA first:

1. There is already a nice iterface (in the idl) as a starting point 
for the connection which will make things easier, I think, when we're 
still feeling around in the dark trying to connect things to Loci.

2. Although any single corba connection to a program is more difficult 
than a single unix pipe connection to a program, there is a lot more 
power in a single program. For instance, connecting with 
biopython/perl/java will give us lots of functions to play with, 
instead of just one. Or think about OpenDX with the corba connections 
Harry will be setting up :-)

3. There is a lot o' love for corba in bioinformatics, so likely we'll 
be able to grab a lot of programs/databases with this approach.

Jeff wrote:
> But, you're right that the first programs to make work under Loci 
should be
> bioinformatics apps.  We're bioinformaticists, right?  However, I 
still want
> non-bioinformaticists to see the capabilities here.  
> programmers outnumber us, how many to one?  10, 100, 1000?  If you 
get these
> guys excited about Loci, it'll really take off!

You're right, of course! I just think with just a few of us now, we 
should gun for something "manageable" but leave things open enough to 
add on later. Otherwise we're going to be swamped.

Gary wrote: 
> And it would be nice if we could the the corba connectivity up and
> running right away.  There's a lot of excitement with bioperl,
> biopython, ebi, etc etc on corba connectivity, so we could possibly
> recruit some help from these groups to get it up and running.

Well, I don't know about help (it seems like all of the projects are 
pretty swamped with ideas but not enough people to help with them!), 
but it would be really cool to get it working!


More information about the Pipet-Devel mailing list