XLIFF TC Telelconference Summary

[Yves Savourel <ysavourel@translate.com>]

XLIFF TC Meeting Summary
Date:  Tuesday, 01 June 2010

=== 1/ Roll call

Present: Asgeir, Bryan, Yves, Andrew, Dimitra, David, Rodolfo, Doug.

Regrets: Christian.

Andrew is official voting member starting today.

 
=== 2/ Approve Tuesday, 18 May 2010 meeting minutes:
   Accept, reject, or amend.
  (http://lists.oasis-open.org/archives/xliff/201005/msg00011.html)

- Asgeir moves to accept the minutes
- Bryan seconds.
- No objections


=== 3/ Current XLIFF business
 3.1. Updates on the 1st International XLIFF Symposium 

Lucia out. Dimitra?
Dimitra: 2 or 3 more submissions. More than 7/8 yet. Web site updated.
Bryan: sounds like we'll have good speakers.
Dimitra: will work on the rooms after we have all submissions.


=== 3.2. June/July/August availability poll (Lucia) 
    (http://www.doodle.com/68fye3zmktwcuuzs)

Maybe August would have less participants, but overall things look ok for all summer.


=== 3.3. "Some Interesting Comments" Thread (Yves, Dimitra)
    (http://lists.oasis-open.org/archives/xliff/201005/msg00012.html)

Bryan: interesting blog entries.
Rodolfo: Saw a record of MemoQ presentation where XLIFF was presented. Some concerns noted there, when comparing tools' support.
Agreement was that it's good that things moves forward. Example of issue: extensibility.
Exchange between different tools is the part to watch out for.
David: Conformance clauses is the important thing we need to do, so compliance is certifiable.
Rodolfo: no clause for now. That's a problem.
David: for 2.0 then we need to make sure we have conformance. Need to keep this in mind while working on the features.


=== 3.4. Conformance clause

Bryan: non-trivial, but needed. Agree with Rodolfo, David, Dimitra.
Should help to get "true" interoperable format.
For 1.2 we thought we escaped the OASIS policy, but now I agree that it's very needed.
Maybe in some cases extensibility can be restricted for better interoperability.
Dimitra: did some tests: interesting results.
Rodolfo: we need to work only with XLIFF file, not tools.
Extensibility is useful, but for example we could ask to have the schemas of the extension available.
David: we talked about that. Did we had a solution?
Asgeir: we use lax for validation. For 1.2 we have both: strict, transitional uses skip, where we should have used lax.
Bryan: lax is the way to go.
Asgeir: Conf clause: we should limit us to 2.0. Too much work otherwise.
David: agree
Dimitra: maybe with the version before, not two.
David: we should do conformance clause only for 2.0.
Bryan: 2.0 cannot really map back to previous version.
Rodolfo: agree
Yves: agree too.
Bryan: need to look if we have existing clauses. Maybe we could define conformance clause independently (like inline), so for charter.
Asgeir: should conformance clause be part of the requirement?
Rodolfo: not fine grain level, but we need to restrict features if they complicate validation.
Bryan: standalone conformance clause is needed (OASIS policy).
But having impact on conformance per feature is good idea too.
David: that's a technicality. Verifiability/test conformance per feature is important. Then that fulfills OASIS requirement.
Rodolfo: can't do that: there is a strict template for the conformance.
Bryan: maybe we could do both as long as the conformance section is there.
Rodolfo: would probably be ok. But text inside the doc are not "conformance clauses".
Conformance clause is very specific. E.g. valid XML and valid with provided schema.
David: So example refer to another part of the specification. Maybe could do the same for other parts.
Rodolfo: Work if we talk about XLIFF documents not tools. Could be written by hand.
David: yes we speak about docs.
Asgeir: what about processing? Can we have conformance clause talk about that?
Rodolfo: before/after processing should be conformant.
Cannot talk about processing.
e.g. one tool generates valid XLIFF, but if you add blank line, not working anymore, even if it's still valid XML/XLIFF.
Do we judge tool or document?
Asgeir: of course. But what about process based on textual definition in the specification?
Rodolfo: processing doesn't matter, just the document before and after. (e.g. translator adds translation).
Asgeir: agree, but certain behaviors could be validated/prescribed.
e.g. segmentation (today) is not prescribed. It could.
David: example means that we would allow them to ignore seg-source.
Rodolfo: can't judge the tool, just the document.
Bryan: sounds like we are starting to discuss different types of conformance (doc/software/etc.)
Rodolfo: what doc vs software would change for the document itself?
Bryan: meant 'features': for example does a tool support it or not?
Rodolfo: then you judge the tool, not the document.
David: look at different groups of transformation. Basic, advance, etc.
Bryan: very good discussion, but need to move on to other topics.
Let's continue on future meetings.
And by Email!


=== 4/ XLIFF Inline text SC report (LISA/OSCAR proposal to re-establish a partnership 
   in developing common text markup)
  1. Yves' SC report

Yves: Meeting next week. Asgeir has created new wiki pages for 'final draft' vs 'working area'.
We are now starting to discuss the requirements
I also contacted Arle to get status on OSCAR input. No answer yet.
Rodolfo: Arle was very busy, but now he has the task to spend some time on standards.

  
=== 5/ XLIFF 1.2 Errata Status Update
  1. Rodolfo's report

Rodolfo: Lucia will work on errata with my help.
Not started yet. Hope to be back on track in July.

Bryan: phone down. Won't be back.

Rodolfo: Time to adjourn.

-end-