[Biocorba-l] Using the Naming Service as a top level object

[Juha Muilu]

Hi Brad, 

The use of naming service is very good idea! To make it more useful we
need also consistent naming schema as you wrote. Your naming schema
looks good. Please have a look what we have:

http://corba.ebi.ac.uk/cgi-bin/ns/ns.start

E.g:
databases/EMBL/nsdb/Embl
databases/EMBL/taxonomy/TaxonServer
databases/RHdb/Proto/RHMap
databases/RHdb/Info/InfoTrader <- name of the corba server
           ^    ^
           |    |
           |    part of the db??
           name of db
We probably can agree on first two fields already. I remember that we
had bit of argument about the rest, so it may not be prefect :-).  

AppLab servers:

analysis.AppLab/BioApps/Clustalw
analysis.AppLab/BioApps/Blastn
analysis.AppLab/BioApps/Cap3

Perhaps it would make sense to build some sort of categories
(alignemnt,similarity search) so that discovery of services comes easier
(or this can be left to dedicated servers... there should be actually
some other standards for this already)

Any other comments, examples etc.?

        Juha

Brad Chapman wrote:
> 
> Hello all;
> While in class last week I got to thinking about how the current
> BioCorba spec doesn't really have a high level interface, like the
> BioEnv interface from BioCorba 0.2. Then I got to thinking about
> using the standard Cos Naming Service to serve the same purpose.
> 
> Anyways, this weekend I hacked up a python implementation of my
> ideas, and I really like what turned out so I thought I would bring
> it up here. Basically, what I did is:
> 
> 1. Set up a Naming Server (omniNames from omniORB) on one of my
> computers.
> 
> 2. Come up with a standard naming scheme for how to bind high-level
> BioSequenceCollection items to the naming server.
> 
> 3. Write a python implementation around this which makes it easy to
> register, unregister and retrieve Database-like
> BioSequenceCollection servers from the naming server.
> 
> Basically I just chose a simple naming scheme in which the
> NamingContext "directory" for Database-like objects is:
> (id = Database, kind = BioCorba)
> and the individual databases are bound inside this with names:
> (id = "The Database Name", kind = "Version")
> 
> This Naming Server based scheme works good for me because:
> 
> => If you have a up-to-date Name Server implementation in your ORB
> (which ORBit is not, unfortunately), then you can connect to the
> Name Server just knowing a URL and port, which I think is more
> natural to people than the CORBA IOR strings.
> 
> => It is easy to bind/unbind objects, so it was easy for me to set
> up my stuff to all be bound to a central server. I like this _a lot_
> :-).
> 
> Anyways, I just wanted to report on what I've come across with this.
> If people think this is a good idea, I'd like to maybe hash out a
> simple naming scheme for storing top-level objects so that we could
> use this approach from different languages. Also, we could
> potentially install a Naming Server on the new Bio* hardware (with
> Chris' permission, etc) which would give people learning
> CORBA/BioCorba a ready made place to hold there objects.
> 
> So, I like the Naming Service :-). Anyone else think this is a good
> idea?
> 
> Brad
> --
> PGP public key available from http://pgp.mit.edu/
> _______________________________________________
> Biocorba-l mailing list
> Biocorba-l@biocorba.org
> http://www.biocorba.org/mailman/listinfo/biocorba-l

-- 
 +--------------------------------------------------------------------+
 |Juha Muilu, Ph.D., EMBL Outstation| Email:  muilu@ebi.ac.uk         |
 |European Bioinformatics Institute | Phone:  +44 (0)1223 494 624     |
 |Wellcome Trust Genome Campus      | Fax:    +44 (0)1223 494 468     |
 |Hinxton, Cambridge CB10 1SD, UK   | http://industry.ebi.ac.uk/~muilu|
 +--------------------------------------------------------------------+