Basically some ifilters will respect embedded language tags for some
document types (word, xml, html). These documents may be broken by different
language word breakers than the default one for your server.
The words will be broken according to language rules and stored in your
catalog as such.
Then when you search the default language rules will be applied at query
time (or overridden if you use the language predicate).
Consider a word doc tagged as German. The words will be broken according to
language rules - so wanderlust would be broken and stored in your catalog as
wanderlust, wandern, and lust.
If you search on it using the English language options you will only get
hits to this document. If you search on lust using the English language
options you will get hits to this document. If you search on wanderlust
using the German language options you will get hits to documents in a
variety of languages containing wanderlust, wandern, and lust.
Watch out for false friends/false conjugates and wander words/wanderworts.
Director of Text Mining and Database Strategy
RelevantNOISE.Com - Dedicated to mining blogs for business intelligence.
This posting is my own and doesn't necessarily represent RelevantNoise's
positions, strategies or opinions.
Looking for a SQL Server replication book?
Looking for a FAQ on Indexing Services/SQL FTS
"Radrizzi Gilles" < XXXX@XXXXX.COM > wrote in