🚀 go-pugleaf

We are  looking for an intelligent and powerful text
comparison  algorithm  or application which would be
used  to  compare a text (10 to 100 pages) against a
large database of texts (1 Gbyte) in order to locate
identical  or  highly similar  portions of  the text
(paragraphs, sentences, or phrases) in the database.

The algorithm should be able to spot non-exact similarities,
eg. paragraphs where certain phrases have been omitted or
added, where words have changed, etc. One could either
imagine a two step algorithm which preselects first a
number of similar texts against which the pattern matching
algorithm is applied, or a powerful algorithm which
compares the text against the whole database. Eventually,
such an algoritm might have been developed for other
applications which could be reused, such as pattern
matching in databases of genetic code.

If we could find such a pearl of an algoritm that can be
build into our own applications, we'd be very happy; if
it would be available as a ready-to-use product,
we'd be delighted.

Any suggestions ?

Since USENET access is a little difficult for us could you
please reply by email to the address below.


======================================================================
Tim Wilson,  Le Cour de Justice des Communautees Europeen - Luxembourg
twilson@cix.compulink.co.uk : Tel: +352 4303 3620 : Fax +352 4303 2600
======================================================================