shin at comeng.chungnam.ac.kr
Mon Aug 31 07:40:58 BST 1998
Robert Hanson wrote:
> As a programmer though, I see a few ways on how it could be implemented...
> 1. Matches could be made on the XML as a stream., and return it as such. So
> you invoke a method on the Query control which asks for a single match. The
> application would that find the first match and return only that one. From
> there you could recursively ask for the next match.
> 2. Return all of the matches in a single action as a single XML document.
> The problem is that if the XML document is very long, this could take some
> time (even days... if the XML document was big enough).
> 3. The query application acts as a search engine where is indexes the XML
> documents before a query is made to increase the speed for the search. This
> could be combined with the interface I mentioned in #1 or #2.
> Does anyone have any thoughts on this?? I was actually thinking about
> creating a search application based on the current XML-QL note, and have not
> decided on how the interface should work. For those familiar with
> Microsoft's ADO control, I was thinking something like that. In which case,
> the search would be return a single match for each call to the query engine.
> The query would also be optimised based on what the options the user
> wanted... like forward only, 1 match at a time (with a READ command, then a
> MoveNext command) -OR- backwards and forwards ( involved caching previous
> matches) -OR- forward, backward, first, last ( involves caching all of the
> matches before allowing reads ).
> Robert Hanson
XML-QL is quite interesting, but I thing it should allow IR queries as well
as database queries. It is because XML documents can be viewed as a set of
collections as well as a database, where a user can retrieve relevant documents
as well as get exact ones.
Look the SQL/MM Full text part. It allows Boolean operators and other operators
frequently used in information retrieval systems.
In this sense, I prefer the approach (3) that Robert Hanson proposed. It would
be better to index XML documents and serve the queries fast. Without the indices,
it would be difficult to compute the similarity between XML documents and user
queries on the fly.
As I posted earlier, BUS is an efficient indexing and retrieval engine for
this purpose. See the page http://savage.comeng.chungnam.ac.kr/~sgml
Department of Computer Engineering, Chungnam National University, Korea
E-mail: shin at comeng.chungnam.ac.kr FAX: +82-42-822-4997
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;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev