ju***********@hotmail.com wrote:
>
Are there specific search-tools that allow custom searching e.g., for
search-queries to yield search results within the same
paragraph-block of content?
Yes, there are; I have been familiar with a product which has a very
rich syntax for expressing search requirements. It allows you to express
proximity, 'between-ness' and order (before and after), and the syntax
covers boolean logic and sub-expressions. You can therefore search for
something like (d 09/01/06) AND ((x NEAR y) BEFORE z). The syntax also
accomodates search targets that have arbitrary fields - not just Subject
and Content - so searching on date-last-modified is certainly possible.
In fact this product could be used to express arbitrary SQL queries
against database tables; so the result-set for a search could in
principle have consisted of a mixture of documents and table-rows.
The product in question has only ever been used to index more-or-less
private collections of data and documents, and I believe it has now been
withdrawn as a standalone product. There may be other similar products;
I'm not familiar with Autonomy in any detail, for example, but they
claim to offer rich searching. Verity have now been acquired by
Autonomy, and their product also provides a measure of rich search
functionality.
http://www.autonomy.com/content/home/
[Warning - this page is pretty useless without Macromedia Flush :-(]
Unfortunately I'm not aware of any public WWW search engine that offers
anything like the power and versatility that I've described. To my
knowledge, none of the tools I've mentioned is used to offer public
searching of the WWW. Public XML search engines that I've looked at
seemed to index tiny numbers of documents, and the quality of their
indexes appeared to be poor.
Supposedly this is the promise of the semantic web:- to be able to
perform global searches using a rich syntax, with terms tailored to the
domain of knowledge with which the search is concerned. But it's my
belief that markup alone is never going to deliver this kind of rich
searching. Apart from any other reason, the majority of documents will
never be marked up to facilitate rich searching based on semantic markup.
--
Jack.
http://www.jackpot.uk.net/