Rosetta One Point Zero

[Sean Wheller]

On Monday 25 April 2005 06:53, Mary Gardiner wrote:
> Rosetta discussion and specifications will be continuing for a few days,
> but a basic outline of our use of Rosetta, our needs from Rosetta and
> any other comments would be useful for them soon: as in, in the next
> 12-48 hours.

Much of the idea is here
http://www.ubuntulinux.org/wiki/DocteamStepByStepl18n

However, in general we would like Rosetta to be a black box from our 
perspective. The main task of the Docteam is writing of docs in English. 
However we recognize the need for translation and so facilitate and support 
i18n processes. Our documentation process therefore has 3 loosely connected 
touch points at which we give and get information to and from Rosetta. I'll 
try outline this thinking.

1. We write an XML document in English.
2. We create a POT file.
3. We get it into Rosetta (touch point 1).
4. We monitor by way of a script the progress of translations (touch point 2).
5. As PO files are completed we automatically collect them and add to our 
repos (touch point 3). We discussed the possability of receiving email to the 
doc list notifying about completed PO files. This can be on a per file basis 
or a single digest containing the status of all our files.
6. Then we create a new XML instance from each of the translated PO files we 
have added to svn.
7. As we update the English XML we merge changes into the POT file
8. At this point the cycle starts again from step 3.

The exchange of POT and PO files should be automated between our SVN 
repository and Rosetta.

Finally at release we do the packaging. Ideally we would like to release 
documentation in language packs.

Overtime we will need this same functionality to be available to tagged 
versions of our docs. Since we tag after release of the distro this should 
not be a problem. Saying this, I think that we will manage this from our side 
by way of our scripts.

With this implimented our major focus can switch to fixing bugs in the 
toolchain. Rosetta was adding strange additional characters to the PO files. 
These caused our XML instances not to be valid and well-formed. A bug is 
filed for this. 

In poxml we also have problems. We are getting superfluous line feeds at the 
end of start tags. This is a problem of kdesdk and poxml and not Rosetta.

This is the current process. Naturally, we would appreciate all efforts that 
make the touch points as automatic possible. From the time we fire up xml2pot 
and add the file to our repos to the time we fire up po2xml, we would like 
the interaction to be automated.

>
> Here's Corey's comments from IRC:
> <Burgundavia> 2. Be able to specify what language the pot file is in,
> and those that language would be automatically translated

I don't know about POT files in languages other than English. Perhaps in 
Rosetta, but In our repos I would only like to see English POT files. They 
are the core drivers of our interaction with Rosetta.

Hope this helps,

-- 
Sean Wheller
Technical Author
sean at inwords.co.za
http://www.inwords.co.za
Registered Linux User #375355
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/ubuntu-doc/attachments/20050425/d7525e1b/attachment.pgp>