- From: Alan Kent <ajk@mds.rmit.edu.au>
- Date: Tue, 13 Mar 2001 10:19:02 +1100 (EST)
- To: www-zig@w3.org
> (ignorant question warning) > >Do you need a schema to use XPATH? I had the impression that you could use >XPATH style access to XML data independent of knowing anything about the >DTD/schema/or similar structural concepts. You do not need a schema to use XPATH. Schemas relate more to how to parse and validate the XML. The same is true of DTDs. XPATH effectively works on the result of the parse, so is a separate issue. Note, the world is a little murky here still in my opinion. Its still shifting (although settling). I mean it was not that long back since they sorted out all the details on namespaces (Eg: could a namespace URI be relative? Was it a string? etc.) But we have certainly used an XPATH engine here without schemas. XML query languages are actually a very interesting topic. My current feeling of the XML query language group (mainly from what people have told me) is that its got a pretty relational feel - that is, its more oriented towards tabular data, E-business and all that. We use XML to encode documents and so need good textual query operators. This is where Z39.50 really is. This is why we use Z39.50. It is not clear to me yet that the XML query group will be able to develop a query language that does a good job of supporting both documents (textual data) and tabular (forms, e-buisness, relational etc) data. Has anyone tried to do the sorts of queries that Z39.50 is used for today with XPATH? For example 'fuzzy match of "centre" within three words of "earth"'. Last time I looked, XPATH does not address how to parse text into words for indexing/searching. ajk
Received on Monday, 12 March 2001 18:20:01 UTC