[minutes] Internationalization telecon 2019-06-06


text version follows:


            Internationalization Working Group Teleconference

06 Jun 2019


           Katy, David, atsushi, Bert, xfq_, JcK, r12a





      * [2]Topics
          1. [3]Action Items
          2. [4]Info Share
          3. [5]RADAR and Active Work Review
          4. [6]String-Meta and RDF direction
      * [7]Summary of Action Items
      * [8]Summary of Resolutions

    <atsushi> I need to get out by 30min to airport, sorry for that

    <r12a> trackbot, start meeting

Action Items

    <r12a> [9]http://www.w3.org/International/track/actions/open

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

    <scribe> scribe: Bert

    r12a: Going through actions.


    <trackbot> action-799 -- Katy Mawhood to Check whatwg for issue
    518 (multiple meta description allowed) -- due 2019-05-09 --


      [10] https://www.w3.org/International/track/actions/799

    Katy: I was wondering if somebody could take a look.

    r12a: Not done 803 and 804 yet.


    <trackbot> action-808 -- Atsushi Shimono to Research the 6
    close suggested issues vs. the whatwg specification -- due
    2019-06-06 -- OPEN


      [11] https://www.w3.org/International/track/actions/808

    atsushi: Haven't had time yet, sorry.

    <r12a> opic: Info Share

    atsushi: May be able to do it this week. Willbe traveling part
    of next week.

Info Share

    <r12a> updated test guidelines wiki page to reflect the latest
    practise for creating pages for the i18n test suite


      [12] https://github.com/w3c/i18n-activity/wiki/Writing-i18n-tests

    r12a: For those writing tests, I updated a page of guidelines.

    <r12a> updated cursive behaviour tests to latest format

    <r12a> [13]https://w3c.github.io/i18n-tests/results/cursive

      [13] https://w3c.github.io/i18n-tests/results/cursive

    r12a: Goal is that people coming from gap-analysis docs
    understand what the tests do.

    <r12a> short review, reduce all alternative pages to one


      [14] https://www.w3.org/International/techniques/shortchecklist.html

    r12a: I also updated how to get into touch with us, see ^^

    <r12a> review request: reorganise so short review link is in
    FPWD, etc

    <r12a> [15]https://www.w3.org/International/review-request

      [15] https://www.w3.org/International/review-request

    r12a: There is now a single location. Makes it easy to put
    issue on GitHub.
    ... Emphasizes self-review. Page should now be clearer.

RADAR and Active Work Review

    <r12a> [16]https://github.com/w3c/i18n-activity/projects/1

      [16] https://github.com/w3c/i18n-activity/projects/1

    r12a: CAPTCHA is in review, who is on that? And we had a
    request for Audio last year. Anybody?
    ... And DCAT, also a request for review.

    Bert: I could do DCAT, but only in July.

    r12a: A bit late. I'll talk to Addison to check.

String-Meta and RDF direction



    r12a: We had a meeting yesterday, some people from JSON-LD and
    some from us.
    ... Addison and other argued for re-engineering JSON-LD to
    support explicit direction.
    ... Others preferred extending the existing language info.
    ... Conclusion is the re-engineer JSON-LD, but that will take
    ... There is still the question of what to do until then.
    ... Interimg solution could be to use private-use area.
    ... Discounted the idea to detect the direction from the
    language. That would only work if script sub-tag is always
    present, and is not the purpose of that tag.
    ... The alternative is better: rely on first-strong detection,
    with fallback.
    ... And in the future rely on extra metadata.
    ... The link ^^ has more details.

    Bert: Is interim solution forward-compatible with future

    r12a: Private-use is inherently a hack, clear to everybody.

    JcK: Long-term use is terrible. The feature "x-" language
    should be deprecated...
    ... I'll see if I can add a note or a reference to the issue
    about that.

    r12a: There was excitement about the fact that you could
    normally detect the direction form the language.
    ... So suggestion was to make a list of languages and
    ... We looked at it in more detail and there is quite a long,
    and open-ended list of language that can be rtl.
    ... And a big gray area.
    ... As I noticed when working on cldr this morning, Azeri
    ("az") language does usually not have a sub-tag
    ... So not feasible to make a finite list of languages.

    JcK: There are a bunch of ISO standards about transliteration.
    ... Cannot determine automatically which transliteration is

    r12a: There's a processing issue as well.
    ... The first-strong heuristic is quick, but sometimes wrong.
    ... The language can be used next.
    ... But if there is explicit metadata, that always wins.
    ... We want there to be metadata always.
    ... Another case is if there is direction info, but not
    language info.
    ... Assume you type in a Hebrew book title in order to search
    for it.
    ... May end up in a situation that inferred language is
    incorrect or missing.

    JcK: Does a direction tag apply to rendering? or to storage?

    r12a: The specs we're talking about are concerned with storage.
    ... It helps to think of 3 stages: production, storage,
    ... The producer needs to provide the info, but storage needs
    to provide enough capabilities to store it.

    JcK: I imagine there are still texts around that store the text
    in the order it was scanned.

    r12a: visual order...

    JcK: Sort of.

    r12a: For our purposes we assume only storage in logical order.
    ... Have to be able to insert text in the middle and rewrap the
    lines. Only possible with logical order.
    ... We also don't think there is much storage in visual order
    any more.
    ... So it seems we are making progress on RDF and JSON-LD, but
    it will take time.


Summary of Action Items

Summary of Resolutions

    [End of minutes]

Received on Thursday, 6 June 2019 14:53:26 UTC