RE: New parts of the draft, please have a look (Re: [all] draft agenda August 23, 2012, 14:00 UTC)

Hi Dave,

 

Yes, I think it’s clearer for me.

And “S ‘or’ T” make sense.

 

-yves

 

From: Dave Lewis [mailto:dave.lewis@cs.tcd.ie] 
Sent: Monday, August 27, 2012 3:59 AM
To: public-multilingualweb-lt@w3.org
Subject: Re: New parts of the draft, please have a look (Re: [all] draft agenda August 23, 2012, 14:00 UTC)

 

Hi Yves, Olaf-Michael,
I agree with Yves, surely one could only judge a segment as being 'mistranslated' or 'untranslated' when considering the target segments, albeit with reference to the source. The quality annotation therefore only applies to the target in this case. 
 
More generally, I assume that the content addressed by this data category should be restricted to content regarded either as a single instance of a source text, or as a single instance of a target text, as they would be regarded in a localization process that is translating from a single target source language to a single target language. 

Obviously to make the quality assessment on many of the types, including the ones above, the quality reviewer needs access to the source. But that access could be offered in many ways and its not the job of this data category (or ITS) to define this. In other word, while an assessment of a translation is in fact the assessment of the couplet of source and target, we can't assume this binding is present in the document we are annotating, and other external mechanisms may sometime be needed to resolve this binding for other users.

So for example, this data category should NOT be applied to a bi-text or parallel text element, e.g. an XLIFF trans unit, though it would be applied to individual source and target (perhaps also alt-trans?) elements in a trans unit. 

does that clarify things?
Dave



On 24/08/2012 11:52, Yves Savourel wrote:

Thanks for the feedback.

 

First: I’ve replaced “S+T” by “S or T” has it was in Dave’s original proposal.

 

I’ve also change the scope for ‘length’ from “T” to “S or T” as this issue can obviously be applied to either contents.

 

As for ‘mistranslated’ and ‘untranslated’: I’m not sure I understand why those would be applying to the source. For example, if I process a source document there is no way I can generate such issue.

 

Maybe the problem is that we don’t define what “scope” means?

I assumed that meant: the issue can be detected when processing the source/target content. But maybe I’m wrong. Could you elaborate your idea Dave?

 

Cheers,

-yves

 

From: Olaf-Michael Stefanov [mailto:olaf@stefanov.at] 
Sent: Thursday, August 23, 2012 7:19 AM
To: public-multilingualweb-lt@w3.org
Subject: Re: New parts of the draft, please have a look (Re: [all] draft agenda August 23, 2012, 14:00 UTC)

 

Dear Felix, Dave and all,

In reviewing the "new draft" below, concentrating on sections related to quality, I find an inconsistency in the "scope" values in the table, referred to at 6.18.1 Definition, as list of type values <http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#lqissue-typevalues> :    <http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#lqissue-typevalues> <http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#lqissue-typevalues>.

Specifically, I find that "mistranslation" and "untranslated" and just as much "S+T" (Source and Target) scoped, as, e.g. "omission" or "inconsistency". 

Saying that "mistranslation" and "untranslated" relate in scope only to "T" (Target) doesn't make sense.

"addition" and "duplication" are the only two Values in this table that should be truly scope "T" (Target) only.
All others, i.e. including "mistranslation" and "untranslated" should be scoped as "S+T" (Source and Target).

Kind regards, 
olaf-michael

On 2012-08-23 09:39, Felix Sasaki wrote:

Hi Dave, all, 

 

as an input to the "new draft" publication, please have a look at the draft at

https://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html
esp. the new sections related to quality and disambiguation. These are here with a warning about stability and a request for comments for people outside the group.

 

Best,

 

Felix

 
Am Donnerstag, 23. August 2012 schrieb Dave Lewis :

Dial in details:
1.  Please join my meeting.
https://global.gotomeeting.com/join/682416317

2.  Use your microphone and speakers (VoIP) - a headset is recommended.  Or, call in using your telephone.

Austria: +43 (0) 7 2088 2169 <tel:%2B43%20%280%29%207%202088%202169> 
France: +33 (0) 182 880 932 <tel:%2B33%20%280%29%20182%20880%20932> 
Germany: +49 (0) 811 8899 6930 <tel:%2B49%20%280%29%20811%208899%206930> 
Ireland: +353 (0) 19 036 185 <tel:%2B353%20%280%29%2019%20036%20185> 
Spain: +34 911 23 4170 <tel:%2B34%20911%2023%204170> 
United Kingdom: +44 (0) 207 151 1816 <tel:%2B44%20%280%29%20207%20151%201816> 
United States: +1 (626) 521-0015 <tel:%2B1%20%28626%29%20521-0015> 

Access Code: 682-416-317
Audio PIN: Shown after joining the meeting

Meeting ID: 682-416-317

Agenda
----------
Topic: Minutes Approval
http://www.w3.org/2012/08/09-mlw-lt-minutes.html

Topic: Agenda Approval

Topic: HTML working group interaction
Quick update on charter change from Felix and a discussion of:
a) how the need to adapt the specification or provide supporting material to maximise a positive outcome
b) what happen where there isn't agreement or concensus takes too long

Topic: Felix: Issue 42 on relationship between tools and confidence scores
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0149.html
as it impacts on on several of the data categories to be discussed in the next topic, namely:
Text Analysis Annotation: http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0139.html
mtconfidence:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0132.html
translation and translationRevisionAgent:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Jul/0256.html
and http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0234.html
and quality

Topic: Data Category Specifications: status, inclusion in next version, update actions for:

special requirements:
now split into separate  forbiddenCharacters, storageSize and displaySize data categories: see thread starting at:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0184.html

MTconfidence score: David, Dave, Declan
thread starting at:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0132.html

Entity/disambiguation: Tadej:
thread starting at call for concensus
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0067.html

transaltionAgent and translationRevisionAgent: Dave
thread starting at:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Jul/0256.html

standoffProvenance: Dave
thread starting at:
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Jul/0278.html

quality: quick roundtable assessment readiness for publication as Arle and Yves won't be on call

Topic: Publication schedule: Felix proposes to publish a new draft next week, see
http://www.w3.org/International/multilingualweb/lt/wiki/Main_Page#Draft_documents_and_time_line

http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0075.html

need agreement to publish based on above data category discussion and outstanding actions

Topic: Test Suite - Dave
Draft test suite using ITS1.0 common format available at:
http://phaedrus.scss.tcd.ie/its2.0/its-testsuite.html
and a there is a separate mail list for discussing this. Discussion needed on:
a) is everyone able to use this common format for their conformance testing, and is it deficient for conformance testing in any way?
b) how can checking of conformance via this common format be maximally automated (and where can't it be)
c) would another format be preferable, XLIFF has been suggested
d) how should we manage the distinction between conformance testing and functional component/product and product integration testing (needed for LT-Web and general promotion but not strictly for conformance)

Topic: Open Action Items
ACTION-34: David - quick FEISGILTT update
https://www.w3.org/International/multilingualweb/lt/track/actions/34