Re: [ISSUE-55] Re: updates provenance mapping and best practive in ITS-XLIFF mapping

Yves,
Thanks for that, we'll look to update those examples without the global rules. 

We then need to capture some of this at some point as xliff- its mapping best practice. One issue here on which i'm unsure is that we have three main use cases, mapping its from a source document to xliff, adding its to xliff for consumption by another its+xliff aware component and the mapping of its in xliff to its in the source format. Do you think that the use of its in xliff will be the same in each case, and more to the point, is this something we should be encouraging in the best practice guideance we give.

Cheers,
Dave

On 19 Feb 2013, at 03:45, Yves Savourel <ysavourel@enlaso.com> wrote:

> Hi Dave,
> 
>> So:
>> a) we might have to support LRQ for mrk in (1) but could 
>> rule it out of scope for (2)
>> b) allowed characters for target would be in (2) and could 
>> use global rules, but perhaps isn't the case for (1) (unless we 
>> include the case where we are importing from another bi-text format.)
> 
> I still think using global rules in any use case is bad: adding information to an XLIFF document means it may be read by another XLIFF tool, so the case (2) can lead to a case (1).
> 
> Let's put it this way: what can we do with global rules we can't with local markup?
> 
> This also goes back to the discussion about using only global rules to define general mapping and avoiding in for specific instance of data. For example you do <its:provenanceRule selector="//source" etc.../> on a file with only trans-unit/source elements and all is well. Then another tool (not ITS aware) adds alt-trans entries, and suddenly you have also all trans-unit/alt-trans/source elements affected by the provenance.
> 
> I would try to stay away from global rules for XLIFF. (actually try to stay away from global rules that are instance-specific in general).
> 
> cheers,
> -yves
> 
> 

Received on Tuesday, 19 February 2013 07:51:57 UTC