Re: Translation Technology at Sun [ was Re: (no subject) ]

[Tim Foster <Tim Foster Sun COM>]

On Thu, 2003-05-22 at 20:15, Gudmund Areskoug wrote:
> and here's how to make it easier to spot "false matches", fuzzy 
> or 100% ones: marking the strings up in the database with 
> hierarchical/relational subject information

Yep - that's dead right, and is a fairly standard thing in translation
memory. We usually do a search of the form : 

"1st, search for translations from the last release of this product"
"Then search for translations in <some category>"
"Finally, search the whole database"

(of course, the question then is, do you rank a fuzzy match from the
first criteria above an exact match from the last criteria ?)

My personal feeling, is that concentrating on .po files is missing the
bigger issue of trying to help automate docs translation, which has a
higher volume (and is less "scriptable", since you don't have the nice
key-value structure of po files with, say docbook)

That said, I could imagine that getting the UI translated sooner could
have a  higher priority : thus bringing as many people as possible to
the GNOME platform...

	cheers,
			tim