XLIFF TC Meeting

When:  Oct 7, 2014 from 11:00 to 12:00 (ET)
Description:

Please get the dial in information from our private Action Item here:
https://www.oasis-open.org/apps/org/workgroup/xliff/members/action_item.php?action_item_id=3663



==========
Agenda:

I Administration (0:00 - 0:10)
  A. Roll call  

  B. Approve meeting minutes, 16 September 2014
     https://lists.oasis-open.org/archives/xliff/201410/msg00000.html  

  C. XLIFF 2.0 as an ISO standard


  D. SOU Report (David)
     https://lists.oasis-open.org/archives/xliff/201410/msg00001.html


(* indicates new since last meeting)

II XLIFF 2.1 (0:10 - 0:45)
  A. New XLIFF 2.1 Tracker page - Bryan     
  https://wiki.oasis-open.org/xliff/FeatureTracking

  B. *Advanced Validation Support - XLIFF 2.1 Feature Proposal approved by electronic ballot
     https://www.oasis-open.org/apps/org/workgroup/xliff/ballot.php?id=2673

  C. *ITS Mapping - XLIFF 2.1 Feature Proposal to be approve by electronic ballot
     https://www.oasis-open.org/apps/org/workgroup/xliff/ballot.php?id=2674

  D. *Error in spc 2.0 in example 5.2.6 (Yves)
     https://lists.oasis-open.org/archives/xliff/201409/msg00033.html


III Sub Committee Report (0:50 - 0:55)


IV Current and New Business (0:55 - )



==========
Minutes:

Agenda

I Administration (0:00 - 0:10)
A. Roll call

Attendance: Bryan, Yves, Asanka, DavidF, David Walters, Fredrik, Joachim, Lucia, Michael, Peter R., Tom, Uwe.

DF: Michael [and Peter] will become voting members after the end of the meeting.

 

B. Approved meeting minutes (DF seconds), 16 September 2014
https://lists.oasis-open.org/archives/xliff/201410/msg00000.html

 

C. XLIFF 2.0 as an ISO standard

This topic is being handled by the P&L Subcommittee until ready for TC consideration.

D. SOU Report (David)
https://lists.oasis-open.org/archives/xliff/201410/msg00001.html

DF: It is now exposed publicly. This topic is also with the P&L Subcommittee.

 

Bryan: the frequency of our meetings, we can discuss this again, as we do not have seem to have so much work going on currently.


(* indicates new since last meeting)

II XLIFF 2.1 (0:10 - 0:45)
A. New XLIFF 2.1 Tracker page - Bryan
https://wiki.oasis-open.org/xliff/FeatureTracking

B. *Advanced Validation Support - XLIFF 2.1 Feature Proposal approved by electronic ballot
https://www.oasis-open.org/apps/org/workgroup/xliff/ballot.php?id=2673

Bryan: The motion has passed.

C. *ITS Mapping - XLIFF 2.1 Feature Proposal approved by electronic ballot
https://www.oasis-open.org/apps/org/workgroup/xliff/ballot.php?id=2674

Bryan: The motion has passed.

B: We have not decided yet how exactly these new features will be implemented, if those features can and should be modules.

DF: It might seem that there is less work going on as we have very slow start after summer, but we really need to move forward and working hard and continue meeting twice in a month if we want to meet the release frequency proposed by Kevin and everyone was in favor. Thanks to all voters for taking part in the electronic ballots and accepting the new work items.

I have some connected issues that I would like to discuss with the TC,

URN prefixes (connected both to the ITS and advanced validation features):

------------------------------------------------------------------------------------

One of the issues is about namespaces. In 2.0 we define XLIFF-defined namespaces and it is important for our constraints (e.g. when we define which attributes are allowed on the inline elements). We use URN prefixes in this definition. I will start this technical discussion also on the mailing list. But we can see now if there are any opinions. OASIS has URN identifiers for all namepaces but W3C stopped caring for them back in 2005 or even earlier.

DF: We also have the advanced validation feature for 2.1, Felix has volunteer to make an NVDL schema, which is important for validation of a multinamespace document, such as XLIFF.

DF: if anyone wants to add any other feature to 2.1 it is now the time to do it.

DF: The URN issue is also connected to the ITS feature. The mapping developed in the W3C WG and subsequently IG: http://www.w3.org/International/its/wiki/XLIFF_2.0_Mapping. I am currently transferring those categories to the 2.1 spec. The categories can be grouped as to how they can be supported in XLIFF 2.0, through core, partially XLIFF and partially ITS, pure ITS as extension, and also non-metadata that just inform extraction or similar. I have added an informative appendix to the 2.1 working draft. It is useful to have them all on one place. It seems to me that we will need to add two more modules, one will be working with the its namespace and the other one with the itsxlf namespace. This is connected with the issue I mentioned before about the URN. Both namespaces are W3C hosted and hence have no URNs defined. I guess OASIS could host the itsxlf namespace or just assign an URN to it, or to both?

Fredrik: Do you want to include a subset of the ITS namespace or do you want to add it completely?

DF: It will always be a subset.

DF: The decision is not only ours, the w3c might need to have a say. I wanted to reuse the W3C namespace where possible. We should talk to the ITS interest group and see if it is possible to take over one of the namespaces or assign URN etc.

Yves: It might be easier to have one module.

Fredrik: OASIS can host a subset of what we need.

ITS in slr:

-----------

Fredrik: we would need some kind of clarification if the ITS uses some normalization. ITS allows for special text restrictions which makes sense in a monolingual context. But when we have two languages like in XLIFF might be different.

Yves: that is a very good question, I have no answer for that one.

Fredrik: I will start the discussion on this topic on the mailing list.

Df: In general, you do not expect to have impact outside the slr module?

Fredrik: No, I do not think so.

Preserve Space data category

------------------------------

DF: custom namespaces including the xml namespace are prohibited in all inline elements, including mrk. If different portions of text have different whitespace handling they should be extracted as different units. I would comment on that on the mailing list. One can also use originalData to preserve whitespaces in text that has otherwise default whitespace handling.

Fredrik: the default for originalData is preserve

DF: In fact the whitespace handling is restricted to “preserve” in original data

Fredrik: that means it could not be flipped..

DF: sure, this could be only used to preserve white spaces.

Yves: whitespace handling can also be set on source and target.

Yves: I have a question what do we do with errors found in 2.0

DF: We agreed in Dublin that 2.x will be restating core, so any corrections to existing specification can be done in 2.1, we do not need to produce extra errata. Of course we do not want to change core materially, we should avoid touching the core. The only error that has been so far reported is the error on the example reported by Yves (https://lists.oasis-open.org/archives/xliff/201409/msg00033.html), that can be considered as an errata type of non-material fix on the new spec, I have it on my todo list. We will deal with other error reports as they come.

Advanced Validation feature:

-------------------------------

DF: Souroush is working on Relax NG and Schematron schemas for the Advanced Validation Support. Schematron is needed for processing requirements. It can also handle some restrictions that cannot be expressed using Schema 1.0 (XSD).

Fredrik commented on the ballot that we might need some discussion why Schematron and NVDL. We may as well have the discussion now.

In Dublin, Felix demonstrated that Relax NG can express some of the advanced contraints that we could not express in our 1.0 xsds. Jirka has a good point that Schema 1.1 can do it. But the industry still largely is with xsd. So one of the imperatives is to keep having 1.0 xsd, as everyone expects to have them and know how they use them. These cover about 80% of the spec, which is good. Now relax NG could cover the constraints. But Relax NG only provides a subset of expressivity that can be covered between Schema 1.0 and Schematron. So we can as well ignore Relax NG. Also Schematron allows for better error messaging than Relax NG.

NVDL is clearly needed for multinamespace validation. This allows us to state what is the relationship between core and module schemas. We used to have modules referenced from the core xsd but had to remove them due to violation of uniqueness constraints.

Soroush is working on a Relax NG schema, but that is fine, we can either discard it, include it as informative, or include as normative with the caveat that in case of conflict xsd + Schematron prevails.

Bryan: Is the purpose of the advance validation support to replace the third party validators? If we introduce a Schematron what would be its role?

DF: It can be a good thing to continue having thrird party validators, e.g. for peple who are less technically savvy and would not use Schematron or Relax NG. It will be up to you or Yves if you want to continue supporting your current validator. But this new approach provides standardized validation that will cover the whole spec. It will be also a way to double check that all validators behave in a same expected way.

Bryan: I have no problem being replaced with Schematron schemas, looking forward to it..

Fredrik: There are two ways of working with schematron: including error context or human readable error text.

DF: Yes, Felix proposed that specification text is loaded as error messages for the schematron rules.

Bryan: Let’s stop the discussion here as we need to conclude the meeting. I invite you to continue the discussion through the mailing list.

Bryan: New business?

DF: We should keep the ITS discussion and the validation support on the agenda.

DF: We will have our 5.5th Symposium at Localisation World in Vancouver on the 29th October. There will be discussions more than presentations; we want to have a good full day working session.

Register for the free FEISGILTT preconference (P05)
It is easy to add to your current registration at the main registration page:
http://www.localizationworld.com/lwvan2014/registration.php

Nevertheless, main conference registration is not required to attend the Symposium

The main topics covered will be:
- Business case for supporting open standards
- Object model and API for XLIFF 2.x
- Interoperability of XLIFF 2.x with TBX, ITS, RDF etc.
- Advanced validation techniques for XLIFF 2.x
- Progress report on XLIFF 2.0 adoption and XLIFF 2.1 development

Bryan: Meeting adjourned.



==========
Attendance:
Meeting Statistics
Quorum rule 51% of voting members
Achieved quorum yes
Individual Attendance Contributing Members: 12 of 32 (37%)
Voting Members: 12 of 13 (92%) (used for quorum calculation)
Company Attendance Contributing Companies: 8 of 13 (61%)
Voting Companies: 8 of 8 (100%)