Foreign language indexing

valery
Posts: 26
Joined: Thu Mar 15, 2001 9:24 pm

Foreign language indexing

Post by valery »

Hi,

We are building a search engine for a set of web-sites spread internationally.
For this, we need to index foreign-language sites and be able to present results in _English_. Does Webinator/Texis have such translation capabilities?

If not, we could put the translation script between webinator and the foreign-language sites (like the one from SYSTRAN technologies..). But then there is another question: will the performance of Texis Metamorph engine drop significantly when searching through computer-generated text as opposed to human-written? (I suppose the engine has been tuned to the human language constructions to extract meaning etc...)

Alternative to the latter method would be to index the sites in foreign language (there are foreign-language versions of webinator, right? do I have to have separate licenses for each of them?) and apply translation to the user queries (ENGLISH -> foreign language) and webinator output (foreign language->ENGLISH).

Could you please comment on which approach is better.

Thanks a lot!
Valery.
Commercial Webinator user.
User avatar
John
Site Admin
Posts: 2623
Joined: Mon Apr 24, 2000 3:18 pm
Location: Cleveland, OH

Foreign language indexing

Post by John »

Webinator/Texis does not have any translation capabilities. The performance of the search should not drop significantly with computer generated text unless it is very badly generated.

Translating twice, both the query and the results sounds as if it would most likely not return as good results since there is less context to aid the translation of the query, and perform worse due to the extra translation of every result.
John Turnbull
Thunderstone Software