[minutes] Internationalization telecon 2015-11-12


text version follows:

Internationalization Working Group Teleconference

12 Nov 2015



    See also: [3]IRC log

       [3] http://www.w3.org/2015/11/12-i18n-irc


           r12a, Addison, JcK, David

           Felix, Steve

           Addison Phillips



      * [4]Topics
          1. [5]Agenda
          2. [6]Action Items
          3. [7]Info Share
          4. [8]RADAR and Active Work Review
          5. [9]Timed Text WG Charter?
          6. [10]Encoding Tests for Encoding Spec
          7. [11]Possibility for a character to be interpreted
             differently depending on locale
          8. [12]AOB?
      * [13]Summary of Action Items


Action Items


      [14] http://www.w3.org/International/track/actions/open

    close action-467

    <trackbot> Closed action-467.

    close action-479

    <trackbot> Closed action-479.

Info Share


      [15] http://www.w3.org/International/tests/repo/results/default-font

    richard: revitalized the above test
    ... haven't linked to it from anywhere yet
    ... shows how language affects the font selected when there is
    no styling

    <r12a> [16]http://w3c.github.io/tlreq/

      [16] http://w3c.github.io/tlreq/

    richard: working on tibetan layout requirements
    ... one of the actions from tpac
    ... was update with what info I had
    ... do that in english
    ... and pass along to translate to chinese
    ... just completed first pass
    ... may be ready for fpwd
    ... soon-ish
    ... after the chinese is ready
    ... may notice that the styling of the tlreq looks a bit
    ... tried applying the new CR spec styling
    ... but a lot comes over
    ... asked on sped-prod to update respec
    ... comments need to be sent by end of november
    ... (rummages for link to fantasai's sample document)



    <r12a> [18]https://github.com/w3c/tr-design/issues

      [18] https://github.com/w3c/tr-design/issues



RADAR and Active Work Review


      [20] https://www.w3.org/International/wiki/Review_radar


      [21] https://www.w3.org/International/wiki/Project_radar

    clreq needs links

    richard: bidi, contacted by developer working on safari
    ... isolation of @dir
    ... send tests that now break because of isolation
    ... helping him work through those or if they are expected
    ... broken things were stuff like HEBREW<span
    ... isolated span is neutral rather than strong
    ... can look like a "lost space" because of isolation

Timed Text WG Charter?

    <JcK> me may have to drop off again, briefly, but without much




      [23] http://www.w3.org/2014/03/timed-text-charter.html

    richard: worth reviewing the discussion of bidi in webvtt
    ... think they did a pretty good job

    <r12a> [24]https://github.com/w3c/webvtt/issues/227

      [24] https://github.com/w3c/webvtt/issues/227

    <scribe> ACTION: addison: write back to timed text and note
    that we've reviewed their charter [recorded in

    <trackbot> Created ACTION-480 - Write back to timed text and
    note that we've reviewed their charter [on Addison Phillips -
    due 2015-11-19].

Encoding Tests for Encoding Spec





    richard: encoding is into the legacy encoding
    ... uses a form that sends a URL to an iframe in the test page
    ... and the query part of the URL hopefully has the correct
    percent encoding for the character
    ... in this case EUC-JP
    ... set an accept-charset to the form
    ... Edge/IE don't use accept-charset of the form
    ... use the tables from the encoding spec to generate expected
    ... changed actual test file
    ... it was in utf-8
    ... changed the test file to euc-jp and the tests worked
    correctly on edge
    ... but using javascript to generate the code points
    ... before I spend a lot of time on this, want to share this

    addison: approach looks okay to me
    ... maybe consider the negative cases (unmapped code points)

Possibility for a character to be interpreted differently depending
on locale


    <trackbot> issue-506 -- Possibility for a character to be
    interpreted differently depending on locale ⓣ -- open

    <trackbot> [28]http://www.w3.org/International/track/issues/506

      [28] http://www.w3.org/International/track/issues/506



    richard: think backslash is a red herring
    ... actual quesstion is about kanji vs. non-kanji style
    ... need a way to send the information about the content
    ... one suggestion was accept-language

    addison: accept-language not helpful: it's from the user agent

    jck: locale-specific font preference issue

    Because the presentational variations (among other things) do
    matter, the ability to mark up documents and runs of text
    within documents with languages tags is something that the I18N
    WG recommends strongly. But this thread seems to be confusing
    the shape of the glyph with the processing that might be
    applied to the text.




      [30] https://code.google.com/p/android/issues/detail?id=19660

    <JcK> The web interactions with ISO 646 national characters
    also shows up in another place although every instance of it is
    arguably a bug. The problem arose when the default web (mostly
    HTML) rule was ISO 8859-1 and folks in some countries said
    something equivalent to "8859-1 doesn't work for me so I will
    interpret it as meaning 8859-whatever-we-use-locally. There
    too, the problem is one coded...

    <JcK> ...octet represents different abstract characters. The
    solution, of course, is "use Unicode; don't do that"

    addison: this issue is that second screen names or maybe
    descriptions are displayed in your device

    (a) unicode code points don't change and presentation of them
    does NOT change the meaning

    (b) end user's expected shapes of the characters depends on the
    font which depends on the language

    jck: see also arabic styles

    (c) needs to be a way to pass the language of the content with
    the content

    (d) accept-language is looking at it from the wrong side

    (e) language negotiation is a very different thing: offer a way
    to request a specific language

    richard: offer a way to change the presentation
    ... "those few characters that have a different look and feel"

    (f) receiver has to be able to sub different font according to
    the language information

    backslash is a red herring. avoid if possible.

    <scribe> ACTION: richard: respond to presentation-api issue 506
    with discussed set of comments [recorded in

    <trackbot> Created ACTION-481 - Respond to presentation-api
    issue 506 with discussed set of comments [on Richard Ishida -
    due 2015-11-19].

    richard: ... and about that backslash...


    richard: still workingon webvtt roundup

Summary of Action Items

    [NEW] ACTION: addison: write back to timed text and note that
    we've reviewed their charter [recorded in
    [NEW] ACTION: richard: respond to presentation-api issue 506
    with discussed set of comments [recorded in

    [End of minutes]

Received on Thursday, 12 November 2015 17:58:54 UTC