Re: wgo i18n

[Gergely Nagy <greg gnome hu>]

On Mon, 2006-11-06 at 11:53 +0100, Quim Gil wrote:
> On Mon, 2006-11-06 at 01:43 +0100, Ramon Navarro Bosch wrote:
>>
> > * Can we define the exact workflow to use on translations ?

I also think it is a good idea to think and talk about the i18n work
flow.

The reason for setting up a CMS was to avoid having to author content
inside CVS. Currently AFAIK the xml2po work flow is still "optimized"
for this CVS based work flow.

It is also clear that the out-of-the-box support in Plone is not
adequate. So the question is, how do we integrate?

Things that are important:
- what format does Plone export to?
- how is data exchanged? (webdaw, http get / post?)
- how are the change notifications syndicated?
- authoring work flow

Format
------
Some format requirements were already discussed in [1], stating that the
markup should reflect structure. This of course goes for the served
XHtml. 

But this is especially for import/export, because all relevant
information must be conveyed in the markup, since the content will be
exported in one language (English), and re-entered into the system as a
new resource in another language.

It is also up for discussion whether we exchange XML sources, or PO
files ready for translation?

How is data exchanged? / Change notification
--------------------------------------------
This is a technical detail, but possibilities should be explored, and
agreed upon. Both sides are "smart", so Plone can actively push out
changed content, or the translation tools can poll for changes. E.g.
Plone could provide RSS feeds for updates (even per language, per page,
per priority, per author, etc).


Authoring work flow
-------------------
Last but not least the authoring work flow should be discussed, and if
possible, improved. One aspect is that the existing translators are
comfortable with their translation environment. The other is that we are
looking for more translators. Since with the integration we are
expanding the capabilities of the whole authoring system, changes to the
current work flow should at least be considered. Maybe we can offer
alternative work flows? 

cheers,
Greg



[1] http://live.gnome.org/GnomeWeb/WebPolicies

> I'll also leave the details of this one. We should aim to integrate as
> much as possible with http://progress.gnome.org (aka
> http://cvs.gnome.org/viewcvs/damned-lies/ ) and the workflows they have
> defined there.
> 
> Some tips while the translators define their requirements/wishlist:
> 
> - Translators love stats like
> http://progress.gnome.org/releases/gnome-2-16 They help them catch the
> remaining work and go for it.
> 
> - Maybe it would be useful to have notifications of new/updated
> translatable content to be sent to a mailing list (or the lists of each
> language).
> 
> - FWIW, well supported languages have a reviewing process of the
> translations committed done by the language coordinators. Languages not
> that well supported have their translations committed directly.
> 
> - It would be useful to have at least 2 priority levels: pages that MUST
> be translated and updated (most if not all the static pages) and
> news-alike pages (i.e. case studies, gnome products...) 
> 
> - We need to define when a language becomes publicly available (i.e.
> having >95% of priority pages and >50% of the rest. Maybe it's worth to
> offer some kind of non-supported visibility to the languages with lower
> levels of translation? Perhaps this would help them get more
> volunteers...
> 
> > On this part we will probably need some new code to adapt and I would
> > like to know how much work is needed to estimate how many help we will need.
> 
> I believe the tough bone is the integration with Damned Lies. If this is
> achieved... translators shouldn't care if the source texts are in a CVS
> or a CMS.
> 
> _______________________________________________
> gnome-web-list mailing list
> gnome-web-list gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-web-list