RE: Proposal: Searching XML

At 14:40 21-04-2002 +0100, Robert Sanderson wrote:

> > In order to model the Z39.50 abstraction in XML terms, you actually need
> > two XML structures - one containing the XML nodes used during the
> > search, the second using the nodes use during the present and some
> > However, XPath also specifies not only the query but also the
> > information returned i.e. also forms the role of the e-specs during a
>
>Right.  Or if you wanted to search a record encoded in XML but return it
>in GRS1, MARC, SUTRS etc, then the second part would get really get messy
>:)

Actually, I would very much like to see a new Espec (Espec-X ?) which 
allows you to give SimpleElement requests using XPATH path statements as 
opposed to the GRS-centric model currently assumed by Espec-1.

And watch the "encoded" above... both the search and the Espec take place 
on abstract records... the record doesn't have to be encoded in any 
specific way as long as the server knows what to do with the request and 
can apply it to whatever weird form the record is actually stored in. XPATH 
is after all just a vocabulary to ask the server to look for certain 
things, "assuming they can be represented according to XPATHs data model... 
the server doesn't have to mess about with DOM trees anymore than it has to 
store USMARC records internally. It just has to act as if it did...

That's the great, crazy beauty of Z39.50. It's dumb and simple enough that 
practically any piece of software can pretend to be a Z39.50 server, which 
is why we have a zillion weird little databases hiding behind Z39.50 
servers. And that's also the big problem with XML Query (can't believe I'm 
doing this again)... try getting a twenty year-old bibliographic catalogue 
to act as an XML Query engine and you'll quickly start to see that Z39.50 
is not really as hard as they say.

--Sebastian
--
Sebastian Hammer, Index Data <http://www.indexdata.dk/>
Ph: +45 3341 0100, Fax: +45 3341 0101

Received on Sunday, 21 April 2002 18:03:37 UTC