On Wed, 21 Apr 2004, Matthew J. Dovey wrote:
> > > How can you have a generic XSLT client as described which will
> > > interoperate with all SRU servers, if the echoed XCQL were optional?
> > You can't, it would be the choice of the implementor/database
> > owner to make the target available for that type of client.
> That viewpoint breaks some of the fundamental philosophies behind
Not really. There are many things which will be important for different
types of clients which servers won't implement. For example, not all
servers will be capable of proximity searches, but that doesn't mean that
they're breaking fundamental parts of the protocol.
Pick any other optional field in the requests or responses and the same
argument applies. This is the scope for profiles and communities to say
what must be supported, rather than the protocol which should provide a
solid base and the means of expressing the extras.
> > In fact it might be desirable to exclude webb-browsers since
> > few servers could stand the load of being Slashdotted.
> Well, if you want to exclude web-browsers then you probably want to
> implement SRW rather than SRU (a server can support either or both of
Not necessarily. If this were the case then there would be no arguments
about SOAP vs REST. While there are additional hooks for terminally
braindead clients in SRU (echoedRequest etc) that's not to say that it
should -only- be used with stylesheet based clients or that it is intended
for these clients.
,'/:. Dr Robert Sanderson ([log in to unmask])
,'--/::(@)::. Special Collections and Archives, extension 3142
,'---/::::::::::. Nebmedes: http://nebmedes.o-r-g.org:8000/
I L L U M I N A T I