Re: [all] Question on mapping best practices

Hi Dave, all,

2012/10/29 Dave Lewis <>

> On 26/10/2012 19:12, Felix Sasaki wrote:
>> Wrt to your comments and ITS mechanisms: why use them at all? Why not
>> specifying the mapping in general, e.g. in a separate profile of ITS "how
>> to use ITS in XLIFF"? We then won't need to use any ITS mechanisms at all -
>> an ITS implementation can use the mapping or not.
>> Above answer may be not enough, let's take it from where.
> Hi Felix, Yves, David, all,
> I think having a separate profile has some distinct advantages. As we've
> discussed in:
> lt/2012Oct/0307.html<>
> and
> lt/2012Oct/0310.html<>
> it seem for some data categories (QualityIssue, Quality Precis,
> transAgentProvenance, disambiguation, text analysis annotation/confidence
> and mtconfidence), the main and possibly only use case for (ref)pointer
> attribute is the XLIFF mapping case.

For disambiguation the mapping to RDFa is another important use case, see
so here we would need pointer attributes.

> See yves example in:
> lt/2012Oct/0357.html<>
> So if the  _only_ use case for pointers in these data categories is
> supporting the XLIFF mapping in this declarative manner, then supporting
> the same in a separate ITS profile could help simplify these data
> categories considerably. What do you think?

It seems that a lot of this depends on decisios in the XLIFF TC, as Yves
wrote here
So not sure - do we have to wait?



> cheers,
> Dave

Felix Sasaki
DFKI / W3C Fellow

Received on Monday, 29 October 2012 05:09:23 UTC