Internationalization Teleconference -- 21 Oct 2009

<scribe> Scribe: Addison Phillips

<scribe> ScribeNick: aphillip

<scribe> Meeting: W3C Internationalization Core WG Teleconference

<David> Hello Addison

hello david

<David> How did the unicode conf go?

Agenda

note from michael kay

Action Items

addison: working on the timing table
... ping martin about iri wg activity
... set up an introductory call or meeting with DOM Events folks to discuss

probably our next call

richard: ask Ian about errata correction of UTR#20

<scribe> done...

<David> http://www.w3.org/TR/unicode-xml/#Charlist

<scribe> ACTION: richard: ask Ian about errata correction of UTR#20 http://www.w3.org/TR/unicode-xml/#Charlist [recorded in http://www.w3.org/2009/10/21-core-minutes.html#action01]

all: look at XForms 1.1 PR for internationalization issues, especially "input mode" at the back

now moot

since it is now a REC

addison: reply to Hixie in support of text at end of thread

all: review UAX#46 draft for review in two weeks

richard: publish language tag articles for wide review

<scribe> done

Info Share

richard: new web fonts charter
... is this MO?
... xforms 1.1 is a REC

DOM KeyEvents

http://lists.w3.org/Archives/Public/public-i18n-core/2009OctDec/0067.html

http://dev.w3.org/2006/webapi/DOM-Level-3-Events/html/DOM3-Events.html

<shepazu> http://dev.w3.org/2006/webapi/DOM-Level-3-Events/html/DOM3-Events.html#keyset

doug: how do i get the character I want when I press a (sequence) of buttons?
... such as dead keys (Option-U + o == ö
... or perhaps Shift-Q => Q
... or third case, IME
... (transformation, choose from lists and such)
... problem #2
... you write a game, for example, and want 'a', 's','d', 'w' be arrow keys
... or you want Ctrl-C to be your copy function
... even if 'c' is remapped to 'h'
... the physical layout is the same
... this is not in the scope of dom3 events

addison: so problem 2 is detecting a physical key press on physical keyboard

richard: so access keys are not included?

doug: never been part of it
... but for example XHTML2 Access... for accesskey
... handwaves over the problem of keys
... rewrote that document and hoping to get published in terms of key identifiers
... solved problem #1
... broke it down
... want single token to id each key
... don't believe that a single id can be had for each key
... key id mechanism can produce similar results across operating systems
... if i have a US-QWERTY keyboard...
... and press 'q'
... when I get 'Q'
... as keycode
... further level of abstraction from java/.net
... if the key produces a character
... then that's key id
... some keys can have more than one identifier
... text input event might give you same value as key id
... if you paste in text or are using IME
... those come as text input events
... also something new...
... composition events
... start, update, end
... implemented in FF
... to enable on-the-spot editing

richard: once you get key id then you do someting with it
... so that's where an access key or accelerator might come in
... so when typing in japanese
... the ime captures qwerty keys and converts to hiragana for example
... but when you have an accelerator key, like, say 'k'
... you don't want to interpret as hiragana first

<r12a> http://rishida.net/utils/keyevents/

doug: some things to watch out for
... list of key ids
... will change in next draft
... some have one value, some have 2, or 3
... some different colors
... for whatever reason, syntax is U+####

(U+20AC"

<r12a> 'j'

<r12a> 'U+006A'

<r12a> The Latin Miniscule Letter j key.

doug: problem I wish to solve
... see if character typed was in a certain range
... meaning block, script, etc.
... such as \p{Greek}
... want to say regex "is this in a range I want"
... rather than U+

addison: problem supplementary

<shepazu> \u006A

U+10000

\ud800\udc00

<r12a> http://rishida.net/tools/conversion/

<r12a> \U00010398

doug: added util method to DOM3 events
... converts to some format
... entity, name, or U+ representation, or a character
... for some characters, for example 'j'
... or its Unicode code point string

<r12a> j U+006A: LATIN SMALL LETTER J

addison: use Unicode names for characters

doug: so some people "hey, do we need conversion"?
... propose instead of U+
... use \u string

supplementary planes

addison: the characters on those planes are real and keyable
... example: emoji
... or Chinese characters on plane 2

doug: we need to hear this feedback
... need even one specific example
... please send in a comment saying "this is a real world use case"
... will point to thread
... is there utility in tranformational utility

addison: could see use for normalizer of char names
... and key ids

doug: so now: key ids can be 'j' or 'U+006A'
... but plan to collapse to '\u006a'

<r12a> 𐎄

<r12a> U+10384

(see cuneiform character)

<r12a> \uD800\uDF84

<r12a> \U00010384

C version has fixed length

<r12a> 𐎄

vs. j

<r12a> shepazu requests:

<r12a> explain why it would be better not to use the javascript escapes for supplementary characters

\u{10384}

\u{10384 6a}

doug: make blog post or summary or something to point to

<r12a> http://www.w3.org/TR/charmod/#sec-Escaping

doug: another spec, WebIDL
... TC39 at ECMA
... lot of Ecmascript folks looking at deliverables
... so maybe fix JS
... possible that if we prosecute our case correctly, might get JS unified with other escape syntaxes
... so... please provide an explicit comment about supplementary support
... and also discuss escapes
... also: please put energy into reviewing this spec please

- DRAFT -

Internationalization Teleconference

21 Oct 2009

Attendees

Contents

Agenda

Action Items

Info Share

DOM KeyEvents

AOB?

Summary of Action Items

Scribe.perl diagnostic output