We recently purchased phDbase for free text retrieval with hebrew. The product is incredibly fast but the results are all wrong...
What we discovered is that it maps the hebrew characters to their ancient dos counterparts (ascii value - 96). These values (128 and above are used for european diacritics , and during the search it lumps all the symbols with a common root (e.g. e, e accented, ), upper and lower case together.
I tried to contact Korenthal but they seem to have dropped from the face of the earth.
If they are monitoring this, or if anyone else had this problem - help would be appreciated.

TIA,


Eylon