[MOBY-l] further thoughts on atomic services & service description

Alan Robinson alan at ebi.ac.uk
Wed Nov 6 12:13:45 UTC 2002


On Tue, 5 Nov 2002 moby-l-request at biomoby.org wrote:

> Date: Tue, 05 Nov 2002 09:35:06 -0600
> From: Mark Wilkinson <mwilkinson at gene.pbi.nrc.ca>
> Organization: PBI-NRC
> To: "moby-l at biomoby.org" <moby-l at biomoby.org>
> Subject: [MOBY-l] further thoughts on atomic services & service description
> 

[...]

> Sequence > BLAST > Blast/Report

To perhaps elaborate on Martin's position, in bioperl one uses more than
one method to get a BLAST report, c.f. a factory creation, a job
submission & a job retrieval.

  $remote_blast = Bio::Tools::Run::RemoteBlast->new('-prog' =>
			'blastp','-data' => 'ecoli','-expect' => '1e-10' );

  $r = $remote_blast->submit_blast("t/data/ecolitst.fa");

  while (@rids = $remote_blast->each_rid ) {
      foreach $rid ( @rids ) {$rc = $remote_blast->retrieve_blast($rid);}}


I must admit that my initial belief of bioMoby was that the "obj1 ->
transform -> obj2" could describe the overall concept of a service at a
high level in a metadata type of way. However, there would be a separate
negotiation between the client & the service using the associated WSDL to
figure out the actual methods & inputs/outputs for the service as defined
using XSD & not being restricted to MOBY types exclusively.

To clarify - your approach in bioMOBY is to completely hide the
programmatic API required & instead to have a single "doBLAST" method
where data & input parameters are sent together & that blocks until the
results return?

Btw, if you decide to divide up MOBY services into "primitives", you need
to determine if state is to be maintained between the calls. Since web
services don't have call backs, this requires a "cookie-like" approach.


alan

--
============================================================
Alan J. Robinson, D.Phil.             Tel:+44-(0)1223 494444
European Bioinformatics Institute     Fax:+44-(0)1223 494468
EMBL Outstation - Hinxton             Email:  alan at ebi.ac.uk
Wellcome Trust Genome Campus
Hinxton, Cambridge
CB10 1SD, UK                http://industry.ebi.ac.uk/~alan/
============================================================







More information about the moby-l mailing list