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

[Martin Senger]

> 
> Only if that is the service-providers responsibility.  Under the 
> existing MOBY architechture it is not... which is *precisly* why it was 
> designed this way.
>
   Perhaps I miss something: but WSDL will be always responsibility of the
service provider, not? The open question is whether the service WSDL is a
part of the Moby Central (whether a service is registered by sending its
WSDL), or if there is only a "pointer" (to a real WSDL) registered. But in
both cases there is, somewhere, a WSDL. And for this WSDL is greatly
important to know how "atomical" it should be, to know what Moby Central 
expects.
   Here is a more concrete example: I have an sequence analysis web
service with two methods "run-blast" and "get-results". And here are two 
scenarios you mentioned:

   1) Moby Central registers only atomic services.
      In this case I (as a service provider) need to create two WSDL 
files, one containing "run-blast" method and one containing "get-results" 
method (and document that they should be used in this order).

   2) Moby Central register "pathes" of methods and allows to register 
services with more methods.
      In this case I can have only one WSDL, and I need to register 
(additionally) a path "run-blast -> get-results" (using, for example, 
WSFL).

   You may guess that my favorite would be the scenario 2. It will make
possible to use Moby Central for complex services, but also it would allow
to composite simply transformation-like services into complex tasks (as is 
the main goal of BioMoby).

   Martin 

-- 
Martin Senger

EMBL Outstation - Hinxton                Senger at EBI.ac.uk     
European Bioinformatics Institute        Phone: (+44) 1223 494636      
Wellcome Trust Genome Campus             (Switchboard:     494444)
Hinxton                                  Fax  : (+44) 1223 494468
Cambridge CB10 1SD
United Kingdom                           http://industry.ebi.ac.uk/~senger