[minutes] Internationalization WG telecon 2013-02-14


Text version follows:

Internationalization Working Group Teleconference

14 Feb 2013



    See also: [3]IRC log

       [3] http://www.w3.org/2013/02/14-i18n-irc


           aphillip, davidClarke(IRC), johnKlensin, matial, r12a,
           fsasaki, aharon, johnOConner, kojii, najib(IRC)

           Norbert, ning




      * [4]Topics
          1. [5]Agenda
          2. [6]action items
          3. [7]Info Share
          4. [8]Isolating Direction in HTML5
          5. [9]Ruby Annotation Extension Spec
          6. [10]Moving Forward Bidi Change Proposal
          7. [11]AOB?
      * [12]Summary of Action Items


action items


    <trackbot> ACTION-165 -- Richard Ishida to contact Tab about
    plans to make the ruby spec extension -- due 2012-12-12 -- OPEN


      [13] http://www.w3.org/International/track/actions/165

    richard: can be closed

    close action-165

    <trackbot> Closed ACTION-165 Contact Tab about plans to make
    the ruby spec extension.

    richard: robin berjon has produced an HTML extension draft


    <trackbot> ACTION-166 -- Richard Ishida to inform wg when bidi
    documents are ready for review -- due 2012-12-12 -- OPEN


      [14] http://www.w3.org/International/track/actions/166

    close action-166

    <trackbot> Closed ACTION-166 Inform wg when bidi documents are
    ready for review.

    richard: ready for review


    <trackbot> ACTION-170 -- Addison Phillips to coordinate with
    richard regarding KLReq -- due 2013-01-16 -- OPEN


      [15] http://www.w3.org/International/track/actions/170

    aphillip: placeholder


    <trackbot> ACTION-174 -- Addison Phillips to write article
    about floating times in HTML -- due 2013-01-30 -- OPEN


      [16] http://www.w3.org/International/track/actions/174

    aphillip: have started this


      [17] http://www.w3.org/International/wiki/FloatingTime

    aphillip: feel free to comment. Have xxx third of the text in


    <trackbot> ACTION-175 -- Richard Ishida to contact Robin Berjon
    about what he's doing wrt ruby extension work -- due 2013-01-30
    -- OPEN


      [18] http://www.w3.org/International/track/actions/175

    richard: done

    close action-175

    <trackbot> Closed ACTION-175 Contact Robin Berjon about what
    he's doing wrt ruby extension work.


    <trackbot> ACTION-176 -- Addison Phillips to ping anne about
    encoding spec -- due 2013-02-07 -- OPEN


      [19] http://www.w3.org/International/track/actions/176

    aphillip: will keep open


    <trackbot> ACTION-177 -- Addison Phillips to establish scribing
    rotation -- due 2013-02-07 -- CLOSED


      [20] http://www.w3.org/International/track/actions/177


    <trackbot> ACTION-178 -- Norbert Lindenberg to coordinate
    proposing text for ITS directionality with richard -- due
    2013-02-07 -- CLOSED


      [21] http://www.w3.org/International/track/actions/178


    <trackbot> ACTION-182 -- Richard Ishida to send note with
    updated text for directionality to MLW-LT -- due 2013-02-14 --


      [22] http://www.w3.org/International/track/actions/182



    <trackbot> ACTION-183 -- Addison Phillips to write draft of new
    charmod-norm -- due 2013-02-14 -- OPEN


      [23] http://www.w3.org/International/track/actions/183

    aphillip: not done yet

Info Share

    aphillip: john klensin joined as invited expert, welcome!

    johnC: have worked on i18n issues on IETF and ICANN for some
    years, IDN and other topics
    ... felt to participate in this work no an ongoing basis rather
    than complaining later

    aphillip: great, welcome!

    johnC: recently adobe received funding to work on ITS2
    ... we are attempting to add tags to jack rabbit

    aphillip: cool


    very cool, joconner :)

    richard: what is jack rabbit?

    joconner: an open source repository
    ... would like to add the ITS tags into the schema for content

    richard: this week published program for MLW workshop in rome
    ... if you want to go register asap. many spaces are taken
    ... have a good program
    ... Mark Davis is doing the keynote
    ... have the usual format with many different people in the
    room who haven't met before
    ... other topic - bidi in HTML work:
    ... some of the things work, thought it is good to put this out
    so that people can look at it
    ... then korean layout requirements doc
    ... lot of english + format editing


    aphillip: next will speak at imug conference



Ruby Annotation Extension Spec

    richard: robin berjon working on HTML5 ruby extension spec
    ... he was unaware of stuff we are doing, need closer
    ... robin sent the draft to public i18n cjk list for review
    ... we need to look into this in the next couple of weeks of

    <aphillip> ACTION: addison: remind group to review berjon's
    ruby spec extension proposal [recorded in

    <trackbot> Created ACTION-184 - Remind group to review berjon's
    ruby spec extension proposal [on Addison Phillips - due

    richard: what is proposed: availibility of an rb tag
    ... fantasai proposes an rb.rb.rt.rt model for jukugo ruby
    ... that is with multiple bases at the start of the ruby tag,
    and multiple rts at the end
    ... I think that this is not necessary
    ... fantasai is concered that if you have an inline set in your
    ... annotations at the end of the base characters should be
    dropped after the ruby tag
    ... I think that can be done easily done by the browser
    re-organizing information
    ... doing that with markup is a very small tail for waging a
    large dog
    ... now double sided ruby:
    ... fantasai is against nesting of ruby elements
    ... she wants to go closer to what the old Ruby Annotation spec
    ... and use an rtc element for the secondary level of
    annotation, eg. rb.rb.rt.rt.rtc(rt.rt)
    ... having done nested stuff I think it's not so difficult and
    may be ok given the frequency
    ... but fantasai's proposal for double-sided using rtc may be
    easier - i need to review the details carefully

Moving Forward Bidi Change Proposal

    richard: in the past we wanted to get isolated spans used in
    inline markup
    ... and we planned to write an extension spec for HTML5
    ... with two new attribute values, corresponding to new unicode
    control characters
    ... however, thinking it over, I thought this might not be the
    best way forward
    ... see the email at
    ... the best way might be: to create a new attribute
    ... that attribute would have ltr, rtl, auto as values
    ... and it would produce isolation by default
    ... you cannot, like in CSS, put in two things and have the 2nd
    one taken up
    ... e.g. two instances of "dir"
    ... we already have problems confusing rtl and ltr when they
    ... adding lri rli has the potential to increase typos
    ... people have to take "intelligent decisions" what to choose
    ... most people have no clue about the importance of isolation
    ... and continue what they have always done
    ... there is a downside of using new values

    aharon: another downside:
    ... lri and rli more or less explicitly say "this should be
    ... that is fine for inline markup, but what about a block
    ... it doesn't makes sense to write <p dir="lri">
    ... so ltr and rtl would not get completely deprecated really

    richard: an alternative would be the "direction" attribute
    ... with the three values ltr, rtl, auto

    <aphillip> <p dir="rtl" direction="ltr"> :-(

    richard: it has the advantage to use it alongside dir in the
    short term to ensure transition

    felix: what guidance would we give people for non-HTML

    richard: good question - will come back to that later?
    ... if we deprecate dir we can move into the direction of

    aphillip: that would make many pages invalid

    richard: correct
    ... additional advantages of direction attribute: people
    continue to use the same values
    ... they just need to know "direction" rather than "dir"
    ... and it doesn't rely on CSS available
    ... the drawback is that it is a pain to use two attributes at
    the same time
    ... one new idea related to this that only occurred to me
    ... we almost never want to do non-isloated text
    ... in unicode report that describes lri and rli it says you
    can still use the old values
    ... if we go away from the old values we will loose the ability
    to embed
    ... which may be useful in some circumstances
    ... we could have a new attribute that produces isolation
    ... and we could keep dir
    ... the new attribute could be bdi
    ... or bd for "base direction"
    ... aharon already said you can use direction aspect of block
    ... so we can give clearer guidance

    <matial> I don't see the difference between the direction attr
    and the bd(i) attribute

    <aphillip> <p dir="rtl" embed>

    richard: some people have used directionality in a way they

    <aphillip> <p dir="ltr" isoloate>

    richard: so generally we always want isolation

    aharon: yes
    ... I support Richard's second proposal (to use a new
    ... use of dir for intentional embedding is not theoretical:
    ... I know people who use dir in a way that does embedding, not
    ... changing the semantics of dir is not an option, and taking
    it away is not an option either
    ... it is incorrect to say that block elements are isolated
    ... what to do about dir name and CSS - that is not an issue
    ... if you have <span direction="ltr">
    ... meaning it is isolated
    ... directionality is ltr, i.e. <span dir="ltr>
    ... that is the same, since direcitonality influecnes what you
    want to do inside the element
    ... islolation talkes about what is outside the element

    <aphillip> <span dir="rtl" isolate>

    aharon: dir is then ignored by new browser if they use

    aphillip: is auto equivalent to FSI?

    aharon: dir=auto is isolating

    richard: in implementations currently it isolates but doesn't
    do the "auto" part

    aphillip: what actions follow from this?



    richard: need to continue discussion, basic proposal is in mail
    I sent
    ... need to decide whether we should choose option two
    ... and need to think about getting things moving
    ... and then we need to find something to author extension spec
    ... before we do all that I think



    felix: should we look into this proposal?

    richard: that proposal just said that there are two types of
    ... there is a note about isolation, and I think the proposal
    is still valid
    ... the point is then you define the attribute values

    aphillip: we were pointing to HTML4

    richard: rlo and lro also don't exist in HTML5, but as concepts
    they do
    ... like to hear opinions

    aphillip: need to look into more material before making

    matial: we cannot change dir behavior because we don't want to
    break existing pages
    ... we cannot unify one attribute because often documents are
    composed from various pieces
    ... telling to use two different attributes for direction, one
    for blocks and one for inline, will be hard to educate people
    ... creating a new attributeis the best thing
    ... if we want a name different from direction it could be
    called "bdir"

    <aphillip> richard: enjoin folks to read email and send

    richard: publish blog post?
    ... could publish on i18n home page

    <Najib-ma> I favors option 2 [create a new attribute]

    <Najib-ma> Thinks it is reasonable.

    <Najib-ma> But I fear people think of "direction" attribute as
    a new name of "dir" attribute.

    <Najib-ma> and then will be surprised by the effect of isolate,
    if any.

    <Najib-ma> That's all

    <scribe> ACTION: richard: publish blog post as wiki, put on
    front door, and announce to winter list [recorded in

    <trackbot> Created ACTION-185 - Publish blog post as wiki, put
    on front door, and announce to winter list [on Richard Ishida -
    due 2013-02-21].

    richard: one more question for aharon
    ... people have used dir embedding wrongly
    ... but are there other cases where you want to embed rather
    than isolate on purpose?

    aharon: no, because then you get with LRM,etc.

    richard: if you unnecessarily/accidentally put dir on span in
    same direction, no problem?

    aharon: other way around

    john: classic problem with domain names


    aharon: in LTR context, span with LTR starting with RTL

    <matial> Be

Summary of Action Items

    [NEW] ACTION: addison: remind group to review berjon's ruby
    spec extension proposal [recorded in
    [NEW] ACTION: richard: publish blog post as wiki, put on front
    door, and announce to winter list [recorded in

    [End of minutes]

Richard Ishida

Received on Thursday, 14 February 2013 18:08:18 UTC