Dear all,

I'm a user of CafeTran (www.CafeTran.com) which uses H2 as a database 
engine (http://cafetran.wikidot.com/working-with-external-databases). I'm 
completely unexperienced with this feature of CafeTran. Before I start 
learning it, I'd like to ask a question here.

In another user group 
(http://groups.yahoo.com/neo/groups/help_/conversations/messages/40917) I 
found this quote:

WFS integrates its own database engine. Most other solutions rely on
> standard database systems, like Oracle™, Lucene, or SQL engines. A
> standard database system (DBMS) may sound reassuring to executives who
> don't grasp the nature of linguistic databases. But there are major
> drawbacks. The first is that DBMSes are built for standard chores like
> business management. They are not optimized for linguistic purposes. A
> standard database system is fine for storing TMs, but its results are
> poor when it comes to exploiting them full-speed in real production. A
> TM engine, when looking for a match in a large TM, needs to access
> then score thousands of candidate TUs before making its final choice –
> all in well under a second. No wonder search engines (Google, Bing),
> faced with a similar task of detecting fuzzy correspondances in large
> masses of text, use their own proprietary database and indexing
> scheme.


My question would be: is H2 the best solution to store Translation Memories 
with millions of segments? Translation Memories are a special kind of XML 
files where every segment (e.g. a sentence) in a source language is 
assigned to a segment in a target language: 
http://en.wikipedia.org/wiki/Translation_Memory_eXchange

Thanks for your help!

Hans 

-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/h2-database.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to