[Minutes] MLW-LT working group call 2012-08-02

... are at http://www.w3.org/2012/08/02-mlw-lt-minutes.html and below as



      [1] http://www.w3.org/

                               - DRAFT -


02 Aug 2012


      [2] http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0057.html

   See also: [3]IRC log

      [3] http://www.w3.org/2012/08/02-mlw-lt-irc


          Arle, davidF, declan, des, dom, felix, milan, jirka,
          shaun, leroy, pedro, philr, michael, tadej


          Arle, Felix, tadej


     * [4]Topics
         1. [5]charter change
         2. [6]call for consensus on special requirements
         3. [7]Named Entity/Disambiguation
         4. [8]mtConfidence
         5. [9]issue-29 progress
         6. [10]issue-34 (quality)
         7. [11]house keeping
         8. [12]aob
     * [13]Summary of Action Items

charter change

   david: I see impact on two levels
   ... good to be official part of HTML, good for our prestige
   ... but it will also force us to be more consicse , more
   ... I agree, we need to make the cut really soon
   ... and what we put out of scope in the first wave

   Felix: Some background why this happened. In the original
   charter we said we would define metadata for HTML5. We would
   use RDFa and Microdata. This approach is difficult. Jirka, in
   discussion with the HTML group, was pointed to a solution, to
   define its- attributes.
   ... This mechanism was not created by us, but was advocated
   because namespaces are not possible in HTML5, but this is a
   replacement for that.
   ... W3C international discussion did not say it was a wrong
   approach, but rather that we need to coordinate this work with
   the HTML5 working group, that we keep them aware and are OK
   with it.
   ... But what does it mean that they are OK with it? One thing
   is that it does *not* mean we are adding attributes to HTML5
   ... Rather we need a review from the HTML5 working group that
   they are OK with our approach. It sounds like a minor
   difference, but to come back to the process, adding attributes
   to HTML5 would be adding to the work done in HTML5. We cannot
   do that. HTML5 is in last call and nothing can be added. All
   that we are doing is defining attributes and getting the
   blessing of the HTML that we are following the right approach.
   ... We need to change the charter for this because we said we
   would not invent our attributes, but instead use RDFa and
   microdata, but we are inventing our own attributes. The change
   in charter is to make them aware that we are doing this.
   ... In terms of timing, it is important that we do this now
   before we finalize the draft so that we can move forward with
   out plan.
   ... Last point: as David said, this has good parts and bad
   parts. The good part is that we now have more interest from the
   HTML community and working group in our work. That interest, I
   know from experience, is not easy. This is all public, btw.,
   you can and should let people know.
   ... The bad aspect is what David said: we can be motivated to
   be as web content-producer digestible/understandable as
   possible. We need to be careful that what we describe and
   define and keep that perspective in mind: we need to make it
   understandable to people outside of localization. Look at what
   Arle did in changing attribute names to make them more
   ... Like David said, it also means we need to close the set of
   data categories we want to deal with.
   ... We may still add mtConfidence, but aside from that, it
   makes sense ton concentrate now on how to sell what we have
   agreed upon to the web content people.


     [14] https://www.w3.org/2002/09/wbs/33280/mlw-lt-charter-2012-update/

   Felix: One admin detail: to make this work, we will need a
   review of the charter. Everyone representing an organization,
   please fill in this form or get your rep to fill it in.

   David: I think this is important and that we took the time is
   good. But let's keep the discussion short.

call for consensus on special requirements

   david: comments that Yves made were made before
   ... the category as specified now contains two to three
   different categories
   ... I think the contents of this category, at least the
   display-size should be taken further
   ... is there anybody who wants to take this further?

   pedro: you mean to split this into several ones?

   felix: propose that micha takes an action item to split this
   into several ones

   micha: sure

   <scribe> ACTION: michael to split special requirements into
   several data categories [recorded in

   <trackbot> Created ACTION-189 - Split special requirements into
   several data categories [on Michael Kruppa - due 2012-08-09].

   micha: it would be just two categories

   <Arle> Apologies for having to drop out. I'll look at the
   minutes later, but someone came to the door and I can't put
   them off. I may be able to jump back in later.

   david: display size, storage, band characters
   ... this should be split I think
   ... should be quite easiy

   pedro: so summary is: we split special requirements into three:
   display, storage size, forbidden characters

   david: yes, band characters are the least stable part
   ... the reg ex thing needs to be resolved

   pedro: the current attributes of storage size and display size
   are part of one category?
   ... for me it is fine

Named Entity/Disambiguation

   david: a lot of discussion about this during last weeks call
   ... action item for tadej to implement this, tadej, what's your

   tadej: I went through the minutes
   ... mostly things were around good terminology to fit all
   ... right now I have a version that integrates all suggestions
   ... I still work on the one with different variants of pointer,
   refpointer etc.
   ... I will send a new version of the draft, this time on google

   david: I thought it should be final?

   tadej: thought it would be necessary
   ... I have enough information from everyone

   felix: no need to have too many call for consensus for a data
   category, if it is ready, we will put it into the draft

   david: have a task force or just post it?

   tadej: from my perspective I think this is ready
   ... just want to have another review from the people on the

   felix: that's fine

   pedro: for this data category
   ... we should involve piek vossen, he can give great input on

   felix: agree, if we send this to the list, piek hopefully can
   join the discussion


   david: had good discussion about mtConfidence
   ... mtEngine self evaluation
   ... Chris Wendt said that this would serve their purpose

   declan: understand the difficiulties ms is mentioning in the
   ... the parts MS was talking about could be hard to implement
   ... would propose to jsut implement mtConfidence score
   ... the automatic metrix are hard and may not be that useful
   across the automatic workflow

   david: in the august list, I responded to jan nelson
   ... declan and chris wendt made similar points
   ... other pointers are needed to produce the score, but not
   needed for a content attribute
   ... agree it would be messy to try to implement this with
   reference implementations
   ... agree with Declan and Chris that self evaluation order and
   confidence would be more useful and stable
   ... happy to drive only mtConfidence
   ... human evaluation does not suffer from this
   ... many people do this
   ... not error checking, but people using simple scale
   ... this evaluation gets more importance
   ... would be good to be able to encode it

   felix: I hope that we can postpone this discussion since we
   have too much stuff to do, we should focus on that

   pedro: for post editing you need a lot of other information
   ... score itself is not enough

   david: think post editing is out of scope
   ... it would be messy if we try to map score and post editing
   ... not sure if this is what you meant
   ... I'm happy to continue just with mtConfidence
   ... this needs to move forward on the ML

   <scribe> ACTION: dfilip to draft a section about mtConfidence,
   based on the discussion [recorded in

   <trackbot> Created ACTION-190 - Draft a section about
   mtConfidence, based on the discussion [on David Filip - due

issue-29 progress

   david: maxime is working on this
   ... prominent in the light of recent changes

   <tadej> scribe: tadej

   fsasaki: The current status is that output to RDF is already
   done and independent of RDFa or Microdata, we are at the point
   of needing a chapter for the standard and defining the RDF

   <fsasaki> rdf representation here
   lt/2012Jul/0065.html and

     [17] http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Jul/0065.html
     [18] http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Jul/att-0065/nodelist-rdfxml.xml

issue-34 (quality)

   <fsasaki> phil: we are very close to being able to issue our

   <fsasaki> .. had various naming and implementation details, we
   are very close

   <fsasaki> david: I discussed with arle that he would submit a
   speaking proposal for seattle

   <fsasaki> .. what's the time line for closing?

   <fsasaki> phil: need to check with Arle

   <fsasaki> david: on track for closing this within august

house keeping

   <fsasaki> see overdue actions at

     [19] https://www.w3.org/International/multilingualweb/lt/track/actions/overdue

   <fsasaki> action-158 - jirka, will do editorial work next week

   <fsasaki> jirka: might make sense that yves edits this

   <fsasaki> action-164 discussed during the call today

   <fsasaki> felix: see editing plans for HTML5 and query language
   attr. here

     [20] http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2012Aug/0062.html


     [21] http://www.w3.org/International/multilingualweb/lt/wiki/PragueSep2012#Objectives

   <fsasaki> felix: this is just a start about the prague f2f,
   feel free to comment

   <fsasaki> david: short update on my action item - seattle event

   <fsasaki> action-34

   <fsasaki> david: we extended call for papers

   <fsasaki> .. felix and arle, can you promote the event on the
   social media setup

   <fsasaki> ACTION: felix to promote seattle event on mlw setup
   [recorded in

   <trackbot> Created ACTION-191 - Promote seattle event on mlw
   setup [on Felix Sasaki - due 2012-08-09].

   <fsasaki> david: we have a strong pc

   <fsasaki> david: lot's of interesting submissions on the way

   <fsasaki> david: on good track with this event


   <fsasaki> pedro: felix asked me to present in prague
   implementation, things of what we use for our showcase,
   progress indicator and readyness

   <fsasaki> felix: everything you have available, if possible
   just show us on the list


     [23] http://www.w3.org/International/multilingualweb/lt/wiki/Best_Practices

   <fsasaki> pedro: agree to focus on this next year

   <fsasaki> david: thanks, think we did good progress today,
   thanks all for your hard work

   <fsasaki> bye everybody

Summary of Action Items

   [NEW] ACTION: dfilip to draft a section about mtConfidence,
   based on the discussion [recorded in
   [NEW] ACTION: felix to promote seattle event on mlw setup
   [recorded in
   [NEW] ACTION: michael to split special requirements into
   several data categories [recorded in

   [End of minutes]

    Minutes formatted by David Booth's [27]scribe.perl version
    1.136 ([28]CVS log)
    $Date: 2012/08/02 15:10:14 $

     [27] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
     [28] http://dev.w3.org/cvsweb/2002/scribe/

Received on Thursday, 2 August 2012 15:14:21 UTC