Internationalization Core Working Group Teleconference

20 Jul 2011


See also: IRC log


matial, Addison_Phillips, Richard, Aharon, kojiishi
Addison Phillips
Addison Phillips



Action Items

Info Share

<r12a> http://www.w3.org/International/

<matial> I too have a little bit of info

richard: home page rejigged a little
... MLweb project mentioned
... fixed up quick links
... direct link to tracker
... and to reviews

<r12a> http://www.w3.org/International/html5/qa-personal-names

richard: working on personal names qa
... started to use HTML5 in article template
... and the link above shows resulting work
... propose to publish in HTML5 in the future

<r12a> http://validator.w3.org/i18n-checker/

richard: just published i18n checker
... more than just new UI
... file upload option
... html5 rules also checked
... not yet xhtml5 and polyglot

<r12a> http://www.w3.org/International/quicktips/checker

richard: lists all checks/messages
... and details about them
... also points to authoritative sources
... should review this as a WG

addison: will add this to agenda for next time

matial: Israeli Internet Society hosting group to design test cases for bidi support in HTML5
... first meeting is next week
... trying to make automated test cases against browsers
... will be interesting to have that

<r12a> http://www.w3.org/International/tests/tests-html-css/list-text-direction

richard: new framework, so tests should fit that

<matial> Here is the URL for this initiative: http://www.w3c.org.il/event/261

<scribe> ACTION: richard: point out the test framework to Aharon for potential use in bidi tests [recorded in http://www.w3.org/2011/07/20-i18n-minutes.html#action01]

<trackbot> Created ACTION-58 - Point out the test framework to Aharon for potential use in bidi tests [on Richard Ishida - due 2011-07-27].



1. Section 2.5.1. Where it says: -- When a user agent is to strip line breaks from a string, the user agent must remove any U+000A LINE FEED (LF) and U+000D CARRIAGE RETURN (CR) characters from that string. -- Should other line-terminators also be included here?


scribe: such as line separator


ISSUE-35 (New attribute: ubi)<http://www.w3.org/International/track/issues/35> I had left http://www.w3.org/Bugs/Public/show_bug.cgi?id=10807 open only because I was not 100% sure that it was a good idea that the <output> element was also defined to have isolation by default. But I never did make up my mind that it was a bad idea, so I have no intention to make a Last Call about it. Just forget about this.

aharon: need for bidi isolatoin. fixed by <bdi>
... did not close this bug because html5 made one of the new elements (output) isolated by default
... not sure if that's the right choice

richard: looked at a couple of days ago and not sure if it still says that

aharon: it's the default stylesheet
... that would say that
... not the spec itself so much

<r12a> 10.3.5 Bidirectional text

<r12a> bdi, output, [dir=auto] { unicode-bidi: isolate; } /* case-insensitive */

okay to close

ISSUE-36 (New dir attribute value: auto, and a new attribute: autodirmethod)<http://www.w3.org/International/track/issues/36> Both http://www.w3.org/Bugs/Public/show_bug.cgi?id=10808 and http://www.w3.org/Bugs/Public/show_bug.cgi?id=11734 were fixed and are closed. The issue should be closed.

ISSUE-43 (Block-display elements should act as UBA paragraph breaks)<http://www.w3.org/International/track/issues/43> http://www.w3.org/Bugs/Public/show_bug.cgi?id=11829* is important and is still awaiting action by the editor.*

aharon: issue is now wider

bug is 11829

aharon: default stylesheet is non-normative and not very descriptive
... and doesn't tell you what is supposed to be happening
... in words
... and HTML5 doesn't describe Unicode Bidirectional Algorithm
... how to apply it
... etc.
... CSS spec (writing-mode)
... does describe this
... and default stylesheets invoke this
... so default stylesheet does define how UBA works in a non-normative and non-declarative way
... so can't tell what the behaviors should be
... and this was accepted by editor
... but not present in document
... I did provide what the text might be
... and pinged

addison: html5 is kind of hand-wavy about bidi
... also, there is a reason not to have stylesheet be normative
... should have a hard reference in html5

richard: could have two stylesheets: one normative and one informative

aharon: add section that says how bidi works in absense of css
... definition in one place
... comment 8 in:

<aharon> http://www.w3.org/Bugs/Public/show_bug.cgi?id=11829#c8

<r12a> see the 4 paras after "By this role, I mean:"

richard: perhaps have a new comment

<r12a> Issue-49

<r12a> http://www.w3.org/International/track/issues/49

<r12a> shouldn't have been filed, and works out ok

<r12a> let's just closed

<r12a> close it

<r12a> Issue-53

<r12a> http://www.w3.org/International/track/issues/52

<r12a> situation is changing right now

<r12a> since email sent

<r12a> by default lists in html5 behave without interop between browsers

<r12a> some list items are in one direction, others in another direction

<r12a> at heart its a css issue

<r12a> this year the css3 lists module was worked on in earnest - the way it is being addressed is far from settled - keeps changing names and approaches

<r12a> in the end it should be fine on the css side, but css3 tries to be very backwards compatible

<r12a> it's not clear yet whether they will decide to make the css default to be the current browser behaviour or the new behaviour

<r12a> not that they like the old behaviour, it's a question of interoperability

<r12a> what they want to happen if for the html5 spec to say the real desireable default in hte default stylesheet

<r12a> unfortunately they still haven't decided on the syntax or whether it will be necessary to say it in the stylesheet

<r12a> so it's not clear how to proceed

<r12a> the editor is constantly modifying the proposal, never mind finish review

<r12a> hope is to convince them to commit to new behaviour for default

<r12a> aharon will be busy lobyying for that over coming days

<r12a> also

<r12a> default text-align is start

<r12a> aharon suggested that the html5 default stylesheet should say that the default text align for li should be match-parent (already defined) - leads to uniform alignment

<r12a> dependent othe the soutions proposed this is either something absolutely necessary or just ok

<r12a> hopefully it will work out that its not important

<r12a> so not clear whether to ask that of the html5 editor yet

<r12a> so, depending on particulars of work on css lists may need some defaults set, but its not yet clear what

<r12a> richard: so if all becomes clear before end of LC, we should push for that

<r12a> otherwise we should lodge a flag that this is an issue still being worked on and that the spec should take into account the results of the discussion when they become available

<r12a> issue-52

<r12a> http://www.w3.org/International/track/issues/52

<r12a> mostly fixed - need to just expand on a small comment

<r12a> if it doesn't happen it's not the end of the world 0- its non-normative

<r12a> http://www.w3.org/Bugs/Public/show_bug.cgi?id=10823

<r12a> ACTION: richard to figure out how to keep alive open bugs for the html5 LC [recorded in http://www.w3.org/2011/07/20-i18n-minutes.html#action02]

<trackbot> Created ACTION-59 - Figure out how to keep alive open bugs for the html5 LC [on Richard Ishida - due 2011-07-27].

Summary of Action Items

[NEW] ACTION: richard to figure out how to keep alive open bugs for the html5 LC [recorded in http://www.w3.org/2011/07/20-i18n-minutes.html#action02]
[NEW] ACTION: richard: point out the test framework to Aharon for potential use in bidi tests [recorded in http://www.w3.org/2011/07/20-i18n-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2011/07/22 17:12:45 $