RE: notes for FEISGILTT discussions

Hi Dave,
Thanks for sharing.
DL> We need to identify overlaps and overlaps between XLIFF and ITS
Possibly, the following material (presentation to which Yves Savourel, Bryan Schnabel, Felix Sasaki and myself contributed) provides good input

http://www.tekom.de/upload/2913/LOC12_Sasaki_Lieske.pdf
                An explicit section (THE RELATIONSHIP BETWEEN ITS AND XLIFF) starts at
http://www.tekom.de/upload/2913/LOC12_Sasaki_Lieske.pdf#page=37
Caveat: The presentation revolved around ITS 1.0 and XLIFF 1.2.
Cheers,
Christian
From: Dave Lewis [mailto:dave.lewis@cs.tcd.ie]
Sent: Freitag, 14. Juni 2013 13:22
To: public-multilingualweb-lt@w3.org
Subject: Re: notes for FEISGILTT discussions

Hi,

Thanks to Pedro and apologies for fogetting to include his presentation, which included remote input from Karl and Mauricio for an impressive international demo: "Interoperability Frankfurt-Madrid: ITS 2.0 CMS to TMS showcase from London"

We will put all the MLW-LT slides from FEISGILTT up on the WG wiki.

cheers,
Dave

On 14/06/2013 00:18, Dave Lewis wrote:
Hi Guys,
Below are some notes from tuesday discussion session at FEISGILTT. We welcome you thoguhts on some of these issues.

Kind Regards,
Dave
FEISGILTT: Day two discussion
12 June 2013
CMS Interoperability Session
Presentations:

*  David Lewis: CMS Interoperability Overview: identifies challenges

*  Bryan Schnabel (Tektronix): Integrating XLIFF into Drupal for complex enterprise multilingual web content

*  Jesús Torres Del Rey, Experience in CMs based localisation with Joomla

*  David Filip: CMS-LION/SOLAS: CMS-XLIFF roundtrip workflow

CMS Interoperability Issues:
The following issues were discussed:

1.       Post localisation changes: how to deal with annotation or changes to content after it can completed a localisation roundtrip, e.g. arising from quality review or feedback from content consumers or content strategy managers

2.       In general, they are see as complementary, and are so by design. We need to identify overlaps and overlaps between XLIFF and ITS:

a.       Overlap in translate/protect, term annotation

b.      XLIFF has competences in the following areas not addressed in ITS:

                                                               i.      segmentation/extraction,

                                                             ii.      bitext exchange and management

                                                            iii.      TM leverage

3.       Similarly need to ITS competences not addressed by XLIFF.

4.       Source segmentation and immutability/changability of segments and their identifiers. Need to articulate the difference between XLIFF (1.2 and 2.0) segmentation structure; xml:tm segementation structuring and NIF URL recipes

5.       Enriching the target content, with meta-data, e.g. from XLIFF or ITS

6.       Key issue is persuading content creators to annotate source:

7.       Explain how ITS source annotation can help with more consistent extraction and segmentation, and therefore to leverage and consistency benefits across (XLIFF-based) localisation workflows.

8.       Does it make sense to start promoting ITS to content management community and then use this as the wedge to promote XLIFF?

9.       Need to consider how to leverage the growing interest in HTML5 to promote ITS (and thereby XLIFF and their mapping)





ITS Session
Discussion focussed on harmonisation/collaboration opportunities.
This was in addition to discussion on Linport-ITS-XLIFF alignment on the first day, where issues included:

*  Location of external ITS files in LinPort container

*  URL conversion on ITS Ref attributed when referencing a resource in the same container, or another container with a known resource.

*  What specific external resources mentioned and referenced from ITS could be included in LinPort

Common processing classifications
Define common processing agent classification. XLIFF already defines:

1.       Extract

2.       Merge

3.       Modify

4.       Enrich
ITS doesn't include any such classification in the spec (through this was discussed during requirements gathering) We should create a table mapping possible ITS use cases against ther classifications. To be complete for ITS we should add perhaps two other complementary classifications:

5.       Internationalise

6.       Post merge processing (enriching and perhaps annotation stripping)

XLIFF-ITS
Current effort on ITS IG to be finalised.
ITS Module in XLIFF
ABsed on the above mapping an ITS module for XLIFF 2.0 should be developed.
Co-evangelization
There seems good potential in evangelising ITS2.0 and XLIFF2.0 in concert. Common messages to target at potential adopters, in particular in localisation clients/content generators and content management technology sector:

1.       What do different ITS/XLIFF features empower specific content creators/managers to do?

2.       What annotation can be automated and how?

3.       What are the benefits of these use cases for the clinet localisation department

4.       Promote ITS and XLIFF combination success stories accessible with usable test cases and examples

5.       Identify and integrate with best-in-class HTML5 editors

6.       CMS integration in particular:

a.       We need to understand why L10n integration is not more of a priority for CMS vendors

b.      Need to understand possible conflicts of interest, e.g.

                                                               i.      System integrators concerned with loosing work to standards based solutions

                                                             ii.      CMS vendors interested in lock-in

7.       In general, making the use case accessible for CMS clients is probably the most direct route to persuading the vendors to include features. Concretely: collaborate on developing  a multilingual content check list of features that purchasers of CMS could reference. This could provide drill down to test suited that could be used in procurement processes. Tie this into a reference implementation that satisfies these features.

8.       There is a potential to integrate Brian XLIFF drupal plugin and Cocomore ITS plugin to provide a single drupal plug-in that could act as a reference CMS implementation for multilingual CMS procurement checklist.

9.       Investigate development of a version of procurement checklist that could be included  in government procurement guidelines  were adherence to open standards, use of open srouce and avoidance of lock-in is an important requirement.

XLIFF Session
David to provide summary