Technical Plenary Notes

Internationalization Core Working Group

Monday, 28 February 2005

Invitations and Scratch

Invitation from MMI to review progress on Pronunciation Lexicon Specification 1.0 at 5PM Monday. Accept? Yes. Grand Ballroom C.

XHTML2/CSS <quote> handling discussion.

WSCG: I owe Stephen R-T some time: lunch Wednesday

WAI: language declarations

SWBPWG: chat about language tags and 3066bis. Tuesday Morning/Afternoon.

Discussion of 3066bis and friends and our relation to it.

Thomas Roessler: ICANN and IDN mess. Also target UA representatives.

Chit-chat about ITS


Felix Sasaki: new Team member. Introduction. Felix will be the new Team contact for I18N Core and will be starting full time in late April.

Reviews. Discussion of different issues that reviews have faced. It would be nice to have a better record of decisions and the like.

CharModFun is in French ici

Language Tags and Locale Identifiers

Very long and detailed discussion here. Richard learned a lot. Evil plots were hatched.

Discuss when and how to do this work. Addison prefers it not be the Tuesday telecon. Need to recruit an editor.

Body Acquisition:

Limited work strategies, such as only doing a sporadic review. Ability to use resources without full sign-up or to have more than one person from a company.

IDN: discussion of the problem and the W3C response.

XML Query Meeting

XML Query WG. Paul Cotton. New Last Call probably April (4th?) for six weeks. Can use WD's that can be looked at now. Review comment responses now especially. Will receive formal request.

Let's see if we can find the links.

Addison: forward Paul links to 3066bis and friends.

Multi-Modal Interaction

Participated in a presentation/review of PLS spec work. This work is Member, so not discussed here.

Tuesday, 1 March 2005

Visitors: Jyri Hagman (Nokia)


Reviews Redux

ATAG 2.0: Deborah signed up (kudos, thanks)

DI-DPF. Addison can look at it some. Others??

QT: excavating our old comments and the current state. Felix and Martin to process our old comments versus the issues lists in each spec.

Action: XSLT 2.0: richard to look at chapter 16 (numbering stuff)

Character Model Resource Identifiers and Normalization

CharModResId: Martin: writing to Bjoern, do some fixing.

ResId testing: need to build the table of tests; RI and MD to coordinate as students take up some of the test development work.

Normalization. FY Skype call.

FY has completed one pass, but the document is not complete and questions remain.

Discussion of C302 and the issues surrounding EUN vs. LUN. Security problem. MUST vs. MUST NOT in handling of suspect text.

(Jeremy joined us here.)

Semantic web RDF Core experience: weakening charmod (for xml) resulted in a knock-on effect of weakening the position of EUN in RDF to should as well.

Add C301a: Specifications that do not require normalization MUST specify how the specification addressed normalization issues that arise. (Proposed text by AP)

Jeremy: Do early or do late or document security limitiations including potential attacks.

Modification of 302. Convert to MUST from MUST NOT?? Perhaps we need to refashion the items to deal with EUN/LUN and reorganize. Split 302 into [S] and [I] components. Also split based on EUN and LUN and no-UN.

Martin to provide suggested version of a split 302.

Addison: fix the link to CharModNorm on core home page

(adjourned for lunch)

Semantic Web Best Practices WG Outreach

Jeremy gives us a synopsis of their activity. Do we have any i18n work for them to consider? Would we care to send a formal communique to that effect?

SWBP assists implementation of RDF and OWL in the most effective manner. Output is generally WG-Note.

Two suggestions: Multilingual Article and document on RFC 3066bis ontology incorporation (with Addison).

Addison to send a marvelously worded email suggesting work on RDF incorporating RFC 3066bis in a few months. Jeremy to send me the words.

WordNet review suggestion (Felix). Need to review TF work for internationalization issues (mentioned by Jeremy).

Martin acquires tools. Jyri leaves us.


CSS and XSL line-breaking.

   discard | preserve | auto-collapse-breaks | space-collapse-breaks | zero-width-collapse-breaks | discard-breaks | preserve-breaks
    wrap | no-wrap | simple-wrap | long-word-wrap | keep-all-words | break-all-words
   normal | strict | none

Visitors (CSS): Melinda Grant (HP), Bert Bos, Steve Zilles, Ian Hickson, David Baron (Mozilla), Tantek, Markus (Microsoft)

References:Last update on CSS page of the fragment Fantasai's current work

Head scratching to resurrect the issues.

Is it essential to have compatibility with XSL-FO (and SVG??) (and IE6)? SZ: Possibly create shorthands/okay to create. Need to be compatible with FO. New unmapped options would present an interop problem in compound doc model.

RI: some of the names chosen were ill-chosen. It isn't just the names, however. Some of the names were confusing, though. Most items not in CSS2. Only CSS3 really mentions specific space handling. Basic behavior of reflowed text (CSS2.1 contains some very limited text? Richard recalls writing a few tests.)

Specific text of what came before: CSS 2.1 16.6.1 reference

SZ: need to map items to more primative properties.

BB: leave up to UA and script combination to interpret.

Discussion of embedding of one script in another: ability to detect language and script boundaries and also using controls to affect differences.

Discussion of SVG's goals and problem space. Some dependency on CSS, since ideally the controls are the same. Richard: requests that we get a picture of where things are at and make it a higher priority.

Bert: need help to address, since 2.1 and selectors is highest priority. Need help from outside.

RI: talk to Michel?

DB: having a base default that works is more important than having knobs to switch it to some other mode.

Action: purchase Richard a waterproof laptop so he can work in the shower.

[14:53] apphillips:
[14:53] fantasai:
[14:57] Bert_lap: Discussing the goal and direction of that draft.
[14:58] Bert_lap: Recalling old discussions and threads.
[14:58] Bert_lap: OK to reduce features to cover just the needed cases.
[14:58] Bert_lap: But are we sure we know all the needed cases?
[14:59] Bert_lap: What is the story for SVG, XSL and IE6?
[14:59] Bert_lap: I.e., what should they do?
[14:59] fantasai: Bert_lap: are you asking me?
[14:59] Bert_lap: SZ: Shorthands with a limited set of choices is of course OK
[15:00] Bert_lap: (No, I'm telling you what is beeing said :-) )
[15:00] fantasai: ok, that's what I thought :)
[15:00] fantasai: thanks
[15:00] * fantasai just opens her textbook and  panics quietly 
[15:01] Bert_lap: Richard: the old names for some properties were also misleading.
[15:02] Bert_lap: SZ: But that is past: they were chosen.
[15:02] fantasai:
[15:05] Bert_lap: MartinD: History is that HTML first didn't say anything, then did, then left it to CSS.
[15:05] Bert_lap: MartinD: SVG is struggling with it now.
[15:05] Bert_lap: Ian: Not in CSS2.1, but should be in CSS3
[15:05] Bert_lap: MartinD: Timing issue between specs.
[15:07] Hixie:
[15:08] Hixie: nevermind, 2.1 does handle this.
[15:08] Hixie: 16.6.1 bullet 3
[15:10] dbaron: fantasai: when are you arriving in Boston?
[15:11] Bert_lap: Richard likes to talk about white space, but can't during Friday during normal meeting time
[15:13] fantasai: dbaron: I'm arriving 8pm on Thursday
[15:13] fantasai: dbaron: and I will be there until Sunday at the latest
[15:14] fantasai: dbaron:is that good enough?
[15:19] Bert_lap: SZ's question is whether every value in the new CSS3 Text will map to a combination of properties/values in XSL and can thus be seen as a shorthand, possibly a shorthand for properties that won't actually exists in CSS3 itself.
[15:20] Bert_lap: Richard: SVG wants to add word wrapping in SVG
[15:21] Bert_lap: Richard: They want *pixel perfect* interoperability.
[15:21] Bert_lap: Richard: The new CSS3 Text props should be available to SVG authors as well.
[15:21] *** fantasai has signed off IRC (Ping timeout).
[15:23] *** fantasai has joined #i18n.
[15:24] bjoern: [21:20]
<Bert_lap> SZ's question is whether every value in the new CSS3 Text will map to a combination of properties/values in XSL and can thus be seen as a shorthand, possibly a shorthand for properties that won't actually exists in CSS3 itself.
[15:24] bjoern: [21:20]
<Bert_lap> Richard: SVG wants to add word wrapping in SVG
[15:24] bjoern: [21:21]
<Bert_lap> Richard: They want *pixel perfect* interoperability.
[15:24] bjoern: [21:21]
<Bert_lap> Richard: The new CSS3 Text props should be available to SVG authors as well.
[15:24] Bert_lap: Richard: SVG is in a hurry.
[15:25] fantasai: bjoern: I'm not 100% sure about Every value,
[15:25] fantasai: bjoern: but I think that mapping to a combination of properties/values in XSL should in many cases be possible
[15:25] fantasai: bjoern: with the general exception of vertical text properties
[15:25] * bjoern just quoting what you might have missed :-)
[15:26] fantasai: bjoern: thanks much :)
[15:26] Bert_lap: Richard: So we need something published.
[15:26] * fantasai missed dbaron's reply also, if any
[15:26] bjoern: there wasn't one so far.
[15:26] fantasai: k
[15:26] dbaron: fantasai: none yet; were discussing whether to schedule an additional meeting
[15:26] fantasai: ok
[15:27] Bert_lap: Bert: But within CSS WG, Text is low priority, compared to 2.1 and Selectors. Would probably need outside help.
[15:27] fantasai: Bert: true, but for me, it is the only spec I am editing
[15:28] Bert_lap: Richard: I think internationalized text is very important. But I have too much work as well :-(
[15:29] Bert_lap: DBaron: Highest priority is to describe the default correctly, more than provide properties to change the behavior.
[15:30] fantasai: Following on that, I would like to know what the i18n wg thinks of the white space collapsing rules defined in that draft
[15:30] fantasai: and if they can contribute any information on line breaking
[15:31] fantasai: besides "strict means no breaks before small kana"
[15:31] fantasai: and "copy ms word"
[15:31] dbaron: people need to read it
[15:31] Bert_lap: Nobody has read the latest version yet.
[15:32] fantasai: ok
[15:32] fantasai: the main difference is that line breaks will
[15:32] dbaron: r12a is chairing GEO meeting Thu and Fri
[15:32] fantasai: collapse to nothing in ideographic context
[15:32] fantasai: as before
[15:33] fantasai: but they won't collapse to zwsp in Thai etc.
[15:33] fantasai: instead, if the line break is next to a zwsp, /then/ it will collapse to zwsp
[15:33] fantasai: because web sites already rely on line breaks collapsing to spaces in Thai
[15:33] fantasai: so, afaict, we cannot make them collapse to zwsp without breaking things
[15:34] dbaron: ----break----


Continuation of Meeting with CSS WG

<quote> and <l> handling in CSS and XHTML

If you have quote marks in your text as characters, they should go outside the actual <quote> element: (I18N)

The quotes must consider the xml:lang on both sides of the element.

Automatic default supplied by UA, intervention via styling, intervention via placement of characters in the text.

(various implementations have attempted to provide context sensitive quotes for a default render)

We have three things, they work, no need to solve this further?

[15:58] Hixie: </br>
[15:59] r12a:
[16:02] Hixie: <qm>&lquo;</qm><quote>Hello world</quote><qm>&rquo;</qm>
[16:02] Hixie: vs
[16:02] Hixie: &lquo;<q>Hello world</q>&rquo;

Date/Time Formatting

CLDR references


Other Items