[Freecats-Dev] Bilingual file format

[Marc Prior]

The subject of a bilingual file format came up recently. I have just 
discovered that there is in fact an open industry standard for a bilingual 
file format, XLIFF. Perhaps other members of the group are aware of it, in 
which case: my apologies! If not, a very interesting article on the subject 
can be found at:

http://www.sun.com/developers/gadc/technicalpublications/articles/xliff.html

(The article also has a screenshot of a very interesting application: a 
twin-window XLIFF editor (which I would guess has been written in Java). I 
have already been in touch with Sun, and I am told that the application 
really does exist (I thought it might have been just a "bare" interface for 
demonstration purposes), but that there are no plans at present to make it
commercially available.)

Is it reasonable to assume that in two years' time, XLIFF will be to 
bilingual files what TMX currently is to translation memories? I would think 
yes. Some TM applications already support XLIFF, it seems.

As I mentioned in a recent posting, OmegaT has no bilingual file format as it 
does not make use of a bilingual file. Whatever the relative merits and 
drawbacks of using a bilingual file *within* a TM application, though, it is 
a major advantage to be able to produce one for export. A strong point in 
favour of Wordfast (and Deja Vu) is that they are capable of emulating 
Trados' "uncleaned file" format.

The existence of a bilingual file format also solves another problem that I 
have been considering, namely that a corpora-based memory system is in many 
ways superior to a database of unrelated segments. 

Thoughts, anyone?

Marc