[FM Discuss] translation workflow

[adam hyde]

hi,

We are working on the Digital Foundations translation sprint (at
eyebeam, nyc). Its a mini-sprint (1 day) to get a start on translation
to spanish but we also have some people working on .de and .pt and
someone coming in this afternoon to look at Hebrew. We wont get it all
done but we will get a good start.

i thought i would make a few notes here for other translation
teams/projects.

So...we have found that babelfish does an alright translation of the
chapters. Hence our workflow is like this :
1. copy the URL of the chapter in READ (npot write) into babelfish
(doing it this way means you get images in the final output whereas
other workflows drop the images and just leave confusing text
references)
2. copy the translated output (not the view source, just the copy and
paste direct from the browser) into babelfish
3. mark the status as 'machine translated' (a new status we created
today)
4. translators come through and look at the side-by-side view and tweak
5. later someone comes thru and translates the screenshots

this seems to work well. google translation does a better job but when
you copy and paste the translation it has all the english sentences next
to the translated sentences (they are normally hidden in span tags but
come visible when you copy and paste) so we use babelfish which doesnt
have this problem

adam


-- 
Adam Hyde
Founder FLOSS Manuals
German mobile : + 49 15 2230 54563
Email : adam at flossmanuals.net
irc: irc.freenode.net #flossmanuals

"Free manuals for free software"
http://www.flossmanuals.net/about