xml-dev - Re: xml search engine?

[Walter Underwood <wunder@infoseek.com>]

At 11:42 AM +0200 3/30/00, Holger Flörke wrote:
>>Someone else wrote:
>>There is a problem I see for xml search engines. How are they going to
>>cope with all the various DTD's?
>
>To cope with this problem I developed the concept of "structural roles".
>Each role refers to a set of elements that have common semantics.

This is essentially the same solution as used in Ultraseek
Server. The "XML Mappings" page in Ultraseek Server is a
bit simpler, because we didn't want to restrict our customers
to "people who can write XPath expressions". We can map from
an element name, or an element name restricted to an attribute
with a particular value. This handles cases like <abstract>
and <div1 type="abstract">, which is a huge subset of 
real-world DTDs.

I think of them as "meta data elements", using the Dublin
Core terminology, but the model is the same. Quick, call
it a Design Pattern and let's publish a book.

wunder
-- 
--
Walter R. Underwood
Senior Staff Engineer
Infoseek Software, part of go.com
1-408-543-6946
wunder@infoseek.com (work)
http://www.best.com/~wunder/

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************