[ISSUE-55] XLIFF - ITS mapping

Hi Felix, all,


I think the “hurry” part comes several reasons:


a)  If we discover some issue on the ITS side when doing the mapping it may be good to discover that before the last Call Working Draft. I doubt such issue would arise at this point, but we never know. For example, we came up with the stand-off notation for the Localization Quality Issue when trying to map it to XLIFF.


b)  The XLIFF TC set itself a deadline of Dec-1 for a Committee Draft, so, like for ITS it would be good to discover any problem that could affect the draft before the CD.


c)  Implementors that use XLIFF should probably have some mapping done before the January face-to-face.


To me, it’s simple: I don’t really want to have a data category in LC stage that I’m not sure I can map to XLIFF. Sure we can always modify ITS to accommodate necessary changes later, but it would be nice to avoid major changes. 


All that said, obviously, anyone with a stake in a given data category that need work should make that their priority, if they don’t have time to do both.


And since I do have a bit of time to do both, I’ll work on putting the Disambiguation changes in the specs, since no one else volunteered yet.






From: Felix Sasaki [mailto:fsasaki@w3.org] 
Sent: Tuesday, October 16, 2012 1:00 AM
To: Dave Lewis
Cc: public-multilingualweb-lt@w3.org
Subject: Re: [ISSUE-55] XLIFF - ITS mapping


Hi Dave, all,


the effort in general is of course very good. However, we had discussed before that we would work on such "best practice" topics in 2013. Why the hurry?


I am asking since as you know from "provenance and agents"


or the tool discussion we have a lot of open issues in the group; and we haven't even started the editorial sections yet. 

I want to avoid the situation of not enough time for and finally (= end of November) dropping data categories because we spend the time on other topics.





2012/10/16 Dave Lewis <dave.lewis@cs.tcd.ie>

Hi all,
First thanks to Yves for starting the mapping page at:

We discussed this in a breakout session at the public section of the XLIFF PC today. the aim was to flesh this out further by mid Nov, just to make sure there were no hidden problems with the important use case that might impact the ITs draft.

Concretely we will break down the the mapping work per data category to named individuals just to make sure we get proper coverage and to save Yves having to do all of this important work itself. Please put you name against some of the data categories if you'd be willing to help.

It was pointed out that mapping to both XLIFF 1.2 and XLIFF 2.0 are important. Also, it was noted that while some ITS tags map to exisitng its meta data, some do not and some also map differently if they have a structural scope rather than being related to a fragement of text within an element. So therefore the mapping will be a mix of mapping from TIS to XLIFF meta-data and also use of meta-data with XLIFF elements, withing the constraints of the specficiations. It is therefore unlikely that ITS tags can be applied to XLIFF using the rules for default, inherience and overriding.



Felix Sasaki

DFKI / W3C Fellow


Received on Tuesday, 16 October 2012 12:20:02 UTC