[seek-dev] Re: GLOBUS and EcoGrid Query wsdl

Matt Jones jones at nceas.ucsb.edu
Tue Sep 16 10:13:45 PDT 2003


Bing,

I agree that using a String type is somewhat easier to program, but the 
problem is that it is too loose for the EcoGrid interface definition to 
make a reasonable *contract*.  The WSDL definition will ultimately be 
the contract that defines exactly what an EcoGrid query system will have 
to implement.  If our method, as defined in the WSDL, says it takes a 
"String" as input, then *any* string should be ok and handled because it 
is not further constrained.  If in fact that String actually needs to be 
an XML document conforming to query.xsd, then this constraint must be 
expressed in the WSDL in order that people implementing the interface 
(outside of our group) know what the EcoGrid requirements are.  The WSDL 
should be our only means of communicating these requirements.

My 2 cents,

Matt


> Dave and Jing,
> 
> Yes. A resultType makes more sense and that's what our XSLTs from
> Metacat and SRB are used for.
> 
> I had also another suggestion that the input of the 'Query' is a string for
> the XML document rather than QueryType. This makes implementation
> much simpler and easier. We can validate the input of XML string against
> our schema, query.xsd, in server side.
> 
> For a QueryType input, a corresponding class should be created. This seems
> to me is an extra overhead considering serializing in and out of an object.
> 
> Jing, where is the code of our grid service generated by wsdl? I'd like to
> take a look of the class since my SRB Parser expects the input is a string
> for the XML document.
> 
> 
> Bing




-- 
-------------------------------------------------------------------
Matt Jones                                     jones at nceas.ucsb.edu
http://www.nceas.ucsb.edu/    Fax: 425-920-2439    Ph: 907-789-0496
National Center for Ecological Analysis and Synthesis (NCEAS)
University of California Santa Barbara
Interested in ecological informatics? http://www.ecoinformatics.org
-------------------------------------------------------------------




More information about the Seek-dev mailing list