[MOBY-l] further thoughts on atomic services & service description
mwilkinson at gene.pbi.nrc.ca
Mon Nov 25 15:38:13 EST 2002
Martin Senger wrote:
> But you need AN agreed way. So pick up one... And because the services
> registered by Moby are Web services, the WSDL is more than reasonable. Or
> say loudly that it is not the case, that the services do not need to be
> web services.
Can you define "web service" for me, as it seems that we might be
chasing our semantic tails here...
>>In MOBY (current architechture) the interfaces are described in
> What do you mean by that? AFAIK the current Moby is defined by Perl
> interfaces. I do not know aht you mean by ontologies in this context.
I mis-spoke somewhat there. I mean that the *service* interfaces are
defined by ontologies, not MOBY Central itself (which, as I have said
numerous times, is NOT a MOBY service...right now...).
We have two ontologies, one which describes every valid input/output
data type, and one which describes every valid service "transformation"
type. between the two of them, we hope to combinatorially define every
possible MOBY service (it sound ludicrous when I say it that way... but
that is the intention...). The service definition, therefore, is a
combination of input + transform + output, and may be represented in
WSDL or any other appropriate language; WSDL being the most obvious one
for many reasons. However, the service definintion is not *defined by*
the WSDL, it is *represented by* the WSDL. I cannot (currently) create
a willy nilly WSDL service definitions and force them back into my
ontology, it is a unidirectional flow - the ontologies define the
service, and the WSDL describes it to the world.
"Speed is subsittute fo accurancy."
Dr. Mark Wilkinson, RA Bioinformatics
National Research Council, Plant Biotechnology Institute
110 Gymnasium Place, Saskatoon, SK, Canada
phone : (306) 975 5279
pager : (306) 934 2322
mobile: markw_mobile at illuminae dot com
More information about the moby-l