XML Search Engine

[Michael Kay]

Tim Bray:
>Can anyone else substantiate this?  Last time I looked at BRS/search,
>it was a very traditional atomic-document thing; it had some fielded
>search, but it could only *find* documents.  Obviously for XML you
>need to find elements.

Not obvious at all. Search engines have always primarily been in the
business of finding documents, I can't see why XML changes this. Of course
it is necessary to present the search engine with "documents" at an
appropriate level of granularity, which is not necessarily the original XML
source document; a filter can do this.

>>Own experience is that relational vendors are completely uncapable of
providing a
>>good solution for text retrieval.

My view is that the relational products are quite useful in hybrid
environments, where the data is mostly structured but includes some lengthy
text fields, e.g. product descriptions in a product database. Their main
limitation is that they take a strictly boolean view of the world: as with
SQL, each record is either a match or it isn't. That's been discredited in
information retrieval research for 25 years, though it is still found in
many products, especially at the low end of the market.

Mike Kay


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)