lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gregor Heinrich" <>
Subject RE: The PrecedenceQuery Class
Date Fri, 25 Apr 2003 09:02:21 GMT
Great idea. Moreover, wouldn't that be a quite general approach to search
engine usability? As a matter of style in written texts, itemisation order
does often imply "first things first" (with respect to importance). I wonder
if Google et al. havn't been using this rule implicitly...

Best regards,


Dipl.-Ing. Gregor Heinrich
Fraunhofer Institut für Graphische Datenverarbeitung
Fraunhoferstrasse 5
64283 Darmstadt
Tel: ++49.6151.155x209
Fax: ++49.6151.155x559
PGP: public key on request.

-----Original Message-----
From: David Medinets []
Sent: Thursday, April 24, 2003 8:30 PM
Subject: The PrecedenceQuery Class

The code for this class can be found at

As I was working with Lucene the other day, I envisioned looking for a
document about a terrorist (call me XXX) and how the resulting list of
results should be ranked. So I created the PrecedenceQuery class so that I
could ask for all document containing XXX and a series of tokens such as
/iraq/syria/israel/usa/united states. The document should be ranked higher
(more relevant) as the list moves from left to right. This type of search is
applicable to many different situations. In the realm of computer
applications, this technique could be used in configuration management so
that global values are overridden by environment which in turn can be
overridden by machine-specific information.

Adding support for prefixes, wildcards, regular expressions, and ranges
inside the delimiters should be trivial if anyone needs it.

Let me know if this classes is helpful.

David Medinets
Quality = Resource Multiplication

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message