W3C

- DRAFT -

MLW-LT Dublin Showcase

18 Jun 2013

Agenda

See also: IRC log

Attendees

Present
omstefanov, dF, Milan, fsasaki, Ankit, Pablo, dave, Declan, chriLi (for "birds eye view" topic), Stephen, K_Savkov, pnietoca, kfritsche, jirka, Yves, fredH, konstantionSakov, nausicaPopa, sebastianH, sorenBendtsen, stephenHolmes, vineet, Tatiana, danielMcGowan, melHowes, tadej, renat, leroy
Regrets
Chair
Dave
Scribe
daveL

Contents


Meeting start and ITS 2.0 birds eye view

<fsasaki> topic: Simple Machine Translation, Translation Package Creation, Quality Check

<omstefanov> Yves Savorel (ENLASO) presenting Okapi "Simple Machine Translation, Translation Package Creation, Quality Check " using slideshow http://www.w3.org/International/multilingualweb/lt/wiki/images/9/92/ENLASO-slides-Dublin-Jun2013.pptx

<omstefanov> It would be good if Yves could also share the samples (HTML and XML) that he is presenting

<omstefanov> Yves is demonstrating by using the Okapi "Rainbow" program to "test run" his demo file with ITS 2.0 markup!

<omstefanov> Great demo/presentation !

<omstefanov> demo has run successfully.

<omstefanov> has left some "issues", strengthening the demo of using in this e.g. the MS MT engine to translate text not otherwise translated

<omstefanov> Yves is now demonstrating the XML doc output to show how an XLIFF processor could use output

<omstefanov> now showing a terminology match

<omstefanov> We absolutely need Yves' examples! Can't emphasize how effective Yves' presentation is !!!

<omstefanov> Yves indicates that OmegaT is an OpenSource product that currently already understands some of ITS 2.0 data categories

<omstefanov> Yves now demonstrating running Quality check after main translation process

<Stephen> I know the schedule is tight, but can Yves describe the itsxlf namespace that he described as a bridge? Will this required in XLIFF 2.0?

<K_Savkov> Does ITS 2.0 ruleas aalow to map to XLIFF resname attribute, apart from Id?

<omstefanov> Stephen: The answer is yes, it will be required in XLIFF 2.0. Hope you're hearing Yves' comprehensive reply

<fsasaki> hi stephen, k_savkov, here is more information about the mapping: xliff 1.2 http://www.w3.org/International/its/wiki/XLIFF_1.2_Mapping and xliff 2.0 http://www.w3.org/International/its/wiki/XLIFF_2.0_Mapping

<K_Savkov> thanks

<fsasaki> it contains also info about the namespace itsxlf: is a schema prefix for the namespace http://www.w3.org/ns/its-xliff/

<Stephen> Thanks for the links!

<fsasaki> hi K_Savkov, the answer to your quesiton "Does ITS 2.0 ruleas aalow to map to XLIFF resname attribute, apart from Id?" is yes

<dF> Just to add to Yves' answer: In XLIFF 2.0 the its and itsxliff namespaces will get the status of the module, which will guarantee the roundtrip

<dF> in XLIFF 1.2 and XLIFF 2.0 ITS is working as extension, where it does not use XLIFF core methods

<Stephen> Great presentation - huge potential, particularly for content authors!

<Yves_> To answer the question about ID and resname: Yes: The mapping basically says that any ID (e.g. defined using the IdValue rule) is mapped to resname.

<fsasaki> thanks a lot for the good feedback, stephen! Esp. the comment about content authors is very encouraging

<omstefanov> Pedro Díez Orzas (Linguaserve) made a great business case for a CMS 2.0 implementation.

<omstefanov> Now Karl Fritsche (Cocomore) is explaining how his tool allows for annotating a document with ITS 2.0 data categories update.

<omstefanov> Karl has just done a live demo to translate the document, already translated into several other langs to also translate it into French

<Stephen> ** Someone needs to mute their mic! **

<omstefanov> demo is now continuing with Mauricio talking, remotely from Madrid. Can be heard very well, even a bit too loud.

<omstefanov> Maricio is explaining the results of the translate action that Karl had just run at Cocomore, as these results have ended up at Linguaserve

<omstefanov> Karl wraps up presentation by showing how French translation is now in the list, and then going "live" to the translated page

<fsasaki> link to ITS Jquery parser at http://plugins.jquery.com/its-parser/

<fsasaki> small demo with the ITS Jquery parser at http://attrib.org/jquery-its-example/

HTML-to-TMS Roundtrip using XLIFF with CMS-LION and SOLAS

<fsasaki> solas demo link at http://demo.solas.uni.me/locConnect/index.php

<omstefanov> David Filip (UL) and Milan Marasek (Moravia) are starting a presentation of an HTML-TMS roudtrip using XLIFF with CMS-LION and SOLAS.

<omstefanov> David is showing SOLAS component MT-Mapper

<omstefanov> David explains that what SOLAS adds to the process is a "Broker", i.e. it understands ITS 2.0 and can send jobs, segment by segment to tools even if they do not understand ITS 2.0 themselves.

<omstefanov> Next, Text Analitics Broker is being explained. Their only current broker is "Tilde" (will be presented later in detail - by Tilde)

<omstefanov> Next Broker being explained is Target Populater, selects one of several alt translations for next step(s). Meant to be a decision maker.

<omstefanov> Another SOLAS component is LKR, the Localisation Knowledge Repository

<omstefanov> David is now creating a sample job to translated English into Spanish to show all the various steps. Category/domain chosen is geography.

<omstefanov> Uploading HTML5 source file. For HTML5 an external rules file is being used.

<omstefanov> SOLAS is creating XLIFF 1.2 based on HTML.

<omstefanov> Next step is to allow user to create a workflow by chosing from a set of components.

<omstefanov> These can be ordered by picking from a list of Components and ordering them into an initially empty list, Workflow.

<omstefanov> Once the workflow list has been created a job is uploaded

<omstefanov> David explains the job through the first 4 steps selected.

<omstefanov> Now Milan is explaining the steps programmed by Moravia, shows how different translation tools are selected based on markup.

<omstefanov> Next explains how additional markup, relating to quality confidence is injected into document

<omstefanov> David now explains how the SOLAS broker has called Milan's service as well as several MT engines and is now showing how it looks after being enriched by MT.

<omstefanov> One sees a number of Alt Translates from various engines.

<omstefanov> David explains that CMS/LiON controls the projects sent to SOLAS. Therefore they can be reviewed there when completed.

Online MT System Internationalization

<Stephen> The sound is echoing pretty badly for me, I can just about make it out. Sounds like his laptop mic is on :(

<pnietoca> http://its2demo.aeat.es/AEAT.internet/Inicio.shtml

<pnietoca> http://www.agenciatributaria.es/AEAT.internet/Inicio.shtml

<fsasaki> thanks a lot, pnietoca!

Reviewer's Workbench: Harnessing ITS Metadata to improve the human review process

<Stephen> Nice presentation Mel. I think that the analytics component is really usefuL!

<fsasaki> https://global.gotomeeting.com/join/697800869

<fsasaki> https://global.gotomeeting.com/join/697800869

ITS 2.0 Metadata and Machine Translation

ITS 2.0 validation

<fsasaki> /me laughs about its-term=yves :)

<SebastianSK> hi, there is an incredible loop back on the audio - hearing jirka twice....

<fsasaki> thanks, sebastian, just told people on gotomeeting to mute themselves, hope that this will help

<Sebastian_Hellmann> The moderator on Gotomeeting can also mute all participants, if they don't comply

<fsasaki> for ITS2 and HTML5 - XML workflows, see also https://github.com/kosek/html5-its-tools

<Tatiana> the quality of audio is very low :((

<fsasaki> tadej? you are talking?

<Tatiana> we cannot hear you, Tadej :(

<Tatiana> better! :)

<fsasaki> now it's better!

<Tatiana> great!

<fsasaki> now it's better, just go on

<Tatiana> again :(

<Tatiana> ;(

<fsasaki> tadej, audio going down again

<fsasaki> now better again!

<Tatiana> ok :)

<Stephen> Hah, it's perfect when you're not presenting! Go figure :-)

<Stephen> Sounds good!

<fsasaki> github link at https://github.com/tilde-nlp/taws/wiki/TAWS-Technical-Documentation

<fsasaki> (parser is in c#)

remote presentations

<fsasaki> SebastianSK, can you say something about the availibiltiy of the plugin?

<fsasaki> that is, avail. for download?

<fsasaki> thanks a lot, SebastianSK!

<SebastianSK> The availability of the extension here http://extensions.libreoffice.org/extension-center will be announced here: http://www.init.de/en/LibreOfficeWriter?ILO2

<fsasaki> thanks, sebastian

afternoon topics

<fsasaki> http://www.w3.org/International/multilingualweb/lt/wiki/Dublin_June_2013_f2f_and_showcase#18_June_afternoon:_WG_discussion_continued

special issue "localization focus"

<fsasaki> dF & daveL will setup easychair to prepare submissions for special issue

<fsasaki> potential contributors: ankit, pedro / karl, daveL / dF, sebastian / dom

<fsasaki> felix can contribute to SW / NIF

<fsasaki> people who submitted to FEISGILLT can upload to easychair

<fsasaki> jirka might contribute too

<scribe> scribe: daveL

preparing recommendation

felix: will do technical work on preparing for final stages of publication
... what other dissemination opportunities?
... aim to release videos from today's session or similar
... would need to be a shorter if we were putting them on youtube
... also looking for each partner to put an ITs promotional page on their web page
... for video we should aim for a single youtube editors.

Karl: points out that some are harder to shrink down - the cocomore/linguaserve one is ling

davidF: have promised 1000 word article for multilignual magazine

<fsasaki> ACTION: felix to nudge everybody about videos + 1pager (e.g. blog post, news etnry, ...) to be provide by 20 July [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action01]

<trackbot> Created ACTION-544 - Nudge everybody about videos + 1pager (e.g. blog post, news etnry, ...) to be provide by 20 July [on Felix Sasaki - due 2013-06-25].

Source code from Tilde

<fsasaki> https://github.com/tilde-nlp/taws

felix: source code of tilde impementation now posted online

<fsasaki> https://github.com/tilde-nlp/taws/wiki/TAWS-Technical-Documentation

felix: this is a general C# parser
... please pass on any comments

implementation page

<fsasaki> http://www.w3.org/International/its/wiki/Use_cases_-_high_level_summary

<fsasaki> http://www.w3.org/International/its/wiki/its implementations

<fsasaki> http://www.w3.org/International/its/wiki/its_implementations

<fsasaki> categories: parsers (javascript, java, c#,...)

<fsasaki> use case driven implementations (drupal editing), plugins

<fsasaki> ok to have same implementation multiple times in the wiki

<fsasaki> re-use "usage scenario doc" material

<fsasaki> and ponit to "usage scenario doc"

<fsasaki> ACTION: felix to nudge implementers to provide input to http://www.w3.org/International/its/wiki/its_implementations [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action02]

<trackbot> Created ACTION-545 - Nudge implementers to provide input to http://www.w3.org/International/its/wiki/its_implementations [on Felix Sasaki - due 2013-06-25].

<fsasaki> wikipedia page about ITS http://en.wikipedia.org/wiki/Internationalization_Tag_Set

XML Schema for ITS

felix: discussion on list about if we need it, or is it just a nice to have

jirka: have changed own view, to be more in favour of this, to add examples of how to use ITS schema in own schema

<fsasaki> xml schema is avail.

<fsasaki> needs to be tested

jirka: can developed and test this in one week. so should be in time for next stage. but should not be authoritative

dF: only one should be authoritiative, though both could be normative as they are used for validation.
... so in case of conflict the relaxNG one is the authorative

<fsasaki> http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#list-of-elements-and-attributes

jirka: relaxNG needs to be the more authoritative, since it is more fully expressive than XML schema

<fsasaki> have non-normative xml schema as appendix?

dF: so agree XML schema as a non-normative appendix, and keep relaxNG as normative, and therefore the authoritative schema

jirka: so add 'for convenience' note referencing the XML schema, in existing schema subdirectory

<fsasaki> have separate links for the schemas. all under http://www.w3.org/TR/its20/schemas/

jirka: this is a persistent URL that can be referred to from the spec

<fsasaki> ACTION: jirka to create ITS2 (xml) schema and to link them from the spec [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action03]

<trackbot> Created ACTION-546 - Create ITS2 (xml) schema and to link them from the spec [on Jirka Kosek - due 2013-06-25].

review feedback

<scribe> ACTION: jirka to complete and put up schemas to http://www.w3.org/TR/its20/schemas/ [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action04]

<trackbot> Created ACTION-547 - Complete and put up schemas to http://www.w3.org/TR/its20/schemas/ [on Jirka Kosek - due 2013-06-25].

felix: EU project review, some outstanding items in feedback
... in terms of full version of pdf of deliverabes and accessibility of online resources.

<fsasaki> ACTION: daveL to allocate time for 26 june call to discuss comments from kimmo and reviewers [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action05]

<trackbot> Created ACTION-548 - Allocate time for 26 june call to discuss comments from kimmo and reviewers [on David Lewis - due 2013-06-25].

<scribe> ACTION: dlewis6 to put discussion of review comments on next week's call [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action06]

<trackbot> Created ACTION-549 - Put discussion of review comments on next week's call [on David Lewis - due 2013-06-25].

Best Practice Topics: 1) XLIFF mapping 2) PROV-O mapping 3) LQI mappings 4) CMS-interop and Readiness 5) CAT tool UI mapping

daveL: doesn't think CAT tool UI mapping is a tractable topic for best practice, given discussion last week at FEISGILTT

<fsasaki> ACTION: felix to track xliff mapping and assure that it is done including testing - due 30 August [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action07]

<trackbot> Created ACTION-550 - track xliff mapping and assure that it is done including testing [on Felix Sasaki - due 2013-08-30].

felix: apart form XLIFF mapping, rely on self-appointed champions for each to drive these forward.

olaf: tow more comments
... first: interest from JAIMCATT, there is a meeting in strasbourg in april, so we might be contacted to represent ITS there
...second: have to leave now but aim to join us in berlin and madrid

Summary of Action Items

[NEW] ACTION: daveL to allocate time for 26 june call to discuss comments from kimmo and reviewers [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action05]
[NEW] ACTION: dlewis6 to put discussion of review comments on next week's call [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action06]
[NEW] ACTION: felix to nudge everybody about videos + 1pager (e.g. blog post, news etnry, ...) to be provide by 20 July [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action01]
[NEW] ACTION: felix to nudge implementers to provide input to http://www.w3.org/International/its/wiki/its_implementations [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action02]
[NEW] ACTION: felix to track xliff mapping and assure that it is done including testing - due 30 August [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action07]
[NEW] ACTION: jirka to complete and put up schemas to http://www.w3.org/TR/its20/schemas/ [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action04]
[NEW] ACTION: jirka to create ITS2 (xml) schema and to link them from the spec [recorded in http://www.w3.org/2013/06/18-mlw-lt-minutes.html#action03]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2013/06/26 11:21:19 $