Enable Word Breaker and Stemmer for Finnish and Overcome the KB Bug

March 10 2010 No comments yet

TechNet has a good article about the effect of word breakers, stemmers and noise word lists in MOSS search. It was known to some that the word breakers and stemmers were not enabled by default in some languages including Finnish, although the components themselves were shipped in MOSS binaries. It was not until last spring when Microsoft documented that in a KB article 929912. The article contains the instructions to enable the components by editing the registry. In our experience, enabling these components improve the search results significantly. Although some customers have felt that the number of results increases too much.

However, there is a caveat. The KB article inserts the registry values containing paths to DLLs as hex-encoded ASCII characters. The path in the registry value is always the Office Server default directory, for example¬† “C:\PROGRA~1\MI54E7~1\12.0\Bin\mir_fi.dll” (meaning C:\Program Files\Microsoft Office Servers\12.0\Bin). If you have installed MOSS to drive other than C, you must change the registry values accordingly. Otherwise the indexing will fail and you will get event log errors about missing DLLs.

Popularity: 1% [?]

Leave a Reply