RE: [all] ITS to XLIFF Mapping

Hi Felix, all

> That would be useful - we can then use the issue name 
> in the mail subject, and people who are not working on this 
> can skip it.

I will raise one then.


> Wrt "how to proceed": although this is not a normative features of ITS 2.0,
> having test files (generic XML / HTML5 / DocBook etc. in > XLIFF+ITS out) 
> seems to be quite useful. Maybe also for the roundtripping, though it 
> seems there is a n:1 mapping from the source format to XLIFF, 
> e.g. all of these <span its:translate=no">...
> <code its:translate=no">...
> would end up in <mrk mtype="protected">
> So should this be part of the or a different "real life usage" 
> test suite?

The only mapping --but maybe that is not the proper term-- we can do is making sure a same content is assigned the same ITS information in both the original data and XLIFF. 
In XLIFF you wouldn't necessarily know to on which element the "do-not-translate" information was set, just that the same content is labeled "do-no-translate" (that is because the original codes are 'abstracted in XLIFF).


> On the "how to proceed" part: do we need to involve the XLIFF TC 
> formally here? By no means I am pushing for that (less formal = faster progress),
> just asking. In terms of the actual work being done we already have 
> many people in both TCs, so that shouldn't be a problem. 

I think we should make sure this effort is visible to the TC. So anyone can participate if they want.
But I don't think the TC needs to be formally involved until we have something that can be a module or some "representation guide".


> Nevertheless a timeline might be good (with milestones like mapping 
> definition, mapping test case dev, mapping testing, etc.).

+1

I also think this needs to be done soon, as it may possibly provide feedback to the ITS WG to fine tune ITS itself. 

Cheers,
-yves

Received on Sunday, 14 October 2012 11:26:43 UTC