Re: [ACTION-153] related to [ISSUE-15] targetPointer

2012/7/10 Yves Savourel <ysavourel@enlaso.com>

> Hi Shaun, all,
>
> I guess you are right: creating a multilingual document (case a) and
> pointing to a target (case b) are difficult to address at the same time.
>
> Your solution for (a) with replicating the selected node (and updating the
> language value) would work fine, but it doesn't cover the more complex
> cases (b) needs.
>
> My solution for (b) is just not possible to implement for creating new
> nodes in many cases, and therefore cannot really work for (a).
>
> We can't cover all requirements of both cases, so that leaves us with the
> need for separate data categories.
>
> Since ENLASO/Okapi's main need is for case b (pointing to a target) I'll
> move forward with that.
>
> So the implementation would be something like:
>
> <its:targetPointerRule selector='//source' targetPointer='../target'/>
>
> And the tool would not be expected to create the target element if it does
> not exist yet. (A note to possible implementers: this makes it a lot easier
> to implement).
>
> As Felix pointed out we should see how many real implementations of this
> we have a chance to get. Currently the intents listed are:
>
>  - 4.2 (XLIFF Round-tripping) I think that is DavidF right?
>

David is planning to wrap the Okapi functionality in his tool - not sure
how we should / can count that.



>  - 4.5 (Okapi component) that me, so a definitive yes.
>  - 4.10 (XLIFF deep web training exporter) I don't know who that is, Milan
> maybe?
>

According to
http://doodle.com/vgt6rbv2cgf7ydpk
Milan will be on the call on Thursday - I'll put targetPointer on the
agenda so that we can get more info about this.

Best,

Felix


>
> Cheers,
> -yves
>
>
>
>


-- 
Felix Sasaki
DFKI / W3C Fellow

Received on Tuesday, 10 July 2012 07:15:50 UTC