how to turn fuzzy off

[Thunderstone]

sometimes looks better to have NO fuzzy searches, esspecialy if texts are
in different languages.

is it possible to turn fuzzy off?

edmunds

[Thunderstone]

Not sure what you mean by "FUZZY"

A: Concept Expansion via the Thesaurus
B: Relevance Ranking
C: The approximate pattern search
D: Word form processing / stemming.


In any event, all of these items are controllable by reading
the manual and adjusting the Texis Web Script accordingly.

The <APICP> tag controls all linguistic aspects of the search
(eg. prefixes, suffixes, thesaurus usage, and delimiters).
See: http://www.thunderstone.com/vortexman/node103.html

The default search method used by Webinator is the "LIKEP" relevance
ranker. There are many adjustable parameters to this clause.
See: http://www.thunderstone.com/texisman/node163.html

The ones specifically germane to this message are:

likepallmatch ( must all terms in the query be matched?)
likepleadbias ( how significant is closeness to document start?)
likeporder ( controls weight of word ordering)
likepproximity ( how important is the proximity of terms?)
likepmaxdist ( controls the maximum radius of ranked terms)

To set these, you precede the database query with <SQL> tag similar to this:

<SQL "set likepallmatch=1"></SQL>

You may also use the other text query methods that are available in the
program instead of LIKEP. These are the SQL LIKE, LIKER, like LIKE3 .
See: http://www.thunderstone.com/texisman/node57.html


The default Webinator search script uses many of these commands, so
it should not be difficult to figure out how to change them to fit your
needs.


Thanks,

Thunderstone