> -----Original Message-----
> From: Matthew Dovey [mailto:[log in to unmask]]
> Sent: Monday, October 28, 2002 5:20 AM
>
> I suggest we go for namespaces of the form URN:ZiNG:SRService:Types,
> URN:ZiNG:SRService:SRW, URN:ZiNG:SRService:SRU, URN:ZiNG:ZeeRex,
> URN:ZiNG:SRService:Indexes etc.
I like this form.
> I've erred towards String for encoding records, whereas Poul has opted
> to XML fragments - as Sebastian points out we haven't a consensus on
> this, although I thought the arguments for escaped XML as a string was
> strongest. Can we come to an agreement on this?
I also prefer string encoding.
> I included an initial schema for XCQL within the types.xsd schema,
> whereas Poul has seperated these out (using Rob's updated one) - any
> feelings on which we should opt for (a separate schema for
> XCQL or just
> a single schema for all the bits)? If we separate out XCQL it may be
> worth separating out diagnostics so that we have a well defined
> namespace for those.
If the decomposition doesn't come with a similar increase in complexity, as
demonstrated in the next item, then I prefer it.
> I opted for including all of the message parts in the WSDL,
> whereas Poul
> has opted for moving them into the types schema. This has the
> following
> consequences:
>
> i) The XML will change from:
> <SOAP:Body>
> <query>...</query> etc
> To
> <SOAP:Body>
> <SearchRetrieveRequestPart>
> <query>...</query> etc
>
> ii) Any stub code produced (certainly by Axis and Java tools) will be
>
> SearchRetrieve(SearchRetrieveRequestPart arg)
>
> Rather than (what I think is more intuitive)
> SearchRetrieve(String query, int startRecord, ...)
>
> iii) The SRU WSDL will have to remain including all of the
> message parts
> in order for the URL-Encoding binding to work (as I understand it).
>
> It was to maintain as much similarity between SRU and SRW plus produce
> logical looking code stubs that I opted for including all the message
> parts in the WSDL. Any other opinions on the matter?
The more intuitive version is my preference. There will be people hand
coding these things.
Ralph
|