While everyone has been struggling with toolkits and namespace declarations for a long time now, it is perhaps time to pay some more attention to SRU. This week we try to get as compatible as we can with the specs but we have applications that use the current implementation. We will probably make the The European Library's SRU test interface and test-service also available. But I really would like to see more links to existing SRU implementations on the ZING site.
In relation to this: Rather than trying to explain the benefits of SOAP for searching an author or a title in a bibliographic database (which we could do even before there was SOAP) and solving toolkits problems, please try to find out how much time it will take you to get an URL-based query converted to a database query and convert the response into XML according to the WSDL specs. If we do that, we can make a big succes out of SRU because end-users can benefit from this directly. IMO it is just to early for SRW as there are no generic SRW clients.
Theo
>>> [log in to unmask] 18-11-02 15:00 >>>
At 13:28 18-11-2002 +0000, Robert Sanderson wrote:
>Apart from these unnecessary last minute changes to the WSDL (and [my]
>XCQL!!), nothing has changed since Friday. All this proves to me is that
>WSDL is not supported in many TKs because it's just not that useful. The
>SOAP sample is still the same apart from some trivial and effectively
>decorative namespace changes.
But the whole idea of using SOAP in the first place is that people should
be able to plug the WSDL into their TKs and boom! -- get an API. If we
don't have this, then my vote remains that we drop SOAP because SRW will
fare better as a straight XML protocol over HTTP that'll be much simpler to
document/understand/implement.
So that makes the WSDL important IMO. I don't know if the changes were
necessary or not, and haven't paid particular attention to the rationale
behind them. What's important, again IMO, is the fact that many key
documents have been changed over the weekend without much of any
verification, and that's just no way to manage a serious release of a spec.
I dont want us to run this like OSI, or even the IETF, but things were just
getting a little ridiculous there, as Mike so aptly pointed out. A fairly
broad range of serious programmers and serious managers paying for their
time (sob :-) are waiting for this spec. I think that with the range of
organisations and people putting their name behind it, people have a right
to expect something a little more considered.
--Sebastian
--
Sebastian Hammer, Index Data <http://www.indexdata.dk/>
Ph: +45 3341 0100, Fax: +45 3341 0101
|