See also: IRC log
<scribe> scribe: JR
JA, CL General comments on how quickly things are moving forward...same prob of people devving new techs not taking into account access
JA: my note was just that Unix was missing
PP: Everything back from UTSPI in
UTF-8
... Then based on users locale put in appropriate lang
... IBM speech engine expects unicode in char set it knows
about
<jallan> JR: does this really need to be here. checkpoint just supports the API
<jallan> cl: get informatoin from DOM or accessibility interface. when you send info to speech engine, you must convert to ascii
CL: When getting info out of API and into speech engine etc. a conversion is often required
<jallan> UA is not helping, AT does all the work
PP: If UTF-8 is being sent a locale also must be sent
JA: From HTML lang element?
CL, PP: No
scribe: but not sure from where
PP: Discussing with AL yesterday
about how to deal with docs made of text in several langs
... Looking at defaults...
... defaukts to ISO-8859 but there is also auto-detect
setting
CL: AL they get it from HTTP or lang attribute
JA: Meta tag?
CL: Or doctype?
PP: In XML tag as well
CL: So yes coming from what author says in web page
JA: Checkpoint seems wide open
CL: But may want to say something about lang and locale...we should say who benefits...speech engines
<scribe> ACTION: CL to Write a bit about speech engines and braille devices needing lang and locale info [recorded in http://www.w3.org/2006/09/14-ua-minutes.html#action01]
JA: so 8859 is just latin languages.
PP: And yes there are Chinese and Japanese speech engines etc.
CL: Just saying browser has to be able to pick up info supplied by author and give it to the ATs
PP: Linux ATK tells locale
... UTF-16 is yet another unicode encoding
JA: Are speech engines moving towards additional char set supports?
PP: From actual checkpoint, browser has to be able to encode strings in form API wants
CL: Browesr could help out by providing the locale
JA: So AL said we can't necessarily do thi.
JR: CL's msg: http://lists.w3.org/Archives/Public/w3c-wai-ua/2006JulSep/0045.html
CL: DOM not supposed to worry about presentation info
JA: And prob is that CSS allows
content to be added
... Really tied up in AJAX and DHTML
... Scripting changing style sheets...
CL: No solution yet
<scribe> ACTION: JA to Follow up on 2 issues with 6.9 - events and content insertion [recorded in http://www.w3.org/2006/09/14-ua-minutes.html#action02]
CL: In that conversatrion Rich suggests getting CSS people to look at it
JA: At plenary Rich was
suggesting a joint meeting on this
... ALso concerned this is a p2...given the new stuff
... This stuff is going to be critical
JA: HPR have font or colour ability?
CL: Possible to get this from
native API
... But today the goal is to get everything via access API
JR: My comment is "WHat's timely?" - more of a technique for other checkpoints
JA: Yes...rationale just says "play well with others"
CL: To us this means providing
via an engineered API rather than a text sream.
... Pref event when things happen rather than having to
detect
<jallan> cl: want api to send semantic information to other API rather than a text string
<jallan> cl: semantics=content + event + location
CL: DOM spec is implemented by browser - that's the engineered API
<jallan> cl: using DOM spec to design API for each UA, need to have semantics built in to UA implemented API
Barry F. says in ATAG 2.0 "accessibility platform architecture(s)"
CL: If following DOM spec then
semantics are built in
... eg. many Windows browsers expose both a DOM and to
MSAA
... Firefox on Linux doesn't expose DOM, but instead exposes
via access API
... Maybe will look at ATAG for possible wording
... "Timely" would seem to indicate events - otherwise how do
you know
PP: ex. an instant message system that doesn't notify user of msg right away
JR: Maybe some objective limit based on response time research
CL: IBM testers were ok with delays of 2-3 seconds
JR: Also relates to input
CL: Many factors including internet communication etc.
JA: Maybe timely is the best word for now?
PP: We could provide examples of what not to do
<scribe> ACTION: PP to Write up a few examples of situations where an API exchange is too slow [recorded in http://www.w3.org/2006/09/14-ua-minutes.html#action03]
JA: Next mtg in 2 wks
... Will pick up with 7
<scribe> ACTION: JA to update review chart and send to list [recorded in http://www.w3.org/2006/09/14-ua-minutes.html#action04]
This is scribe.perl Revision: 1.127 of Date: 2005/08/16 15:12:03 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found Scribe: JR Inferring ScribeNick: JR Default Present: Jim_Allan, Pparente, cklaws, Jan_Richards Present: Jim_Allan Pparente cklaws Jan_Richards Regrets: D Poehlman WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 14 Sep 2006 Guessing minutes URL: http://www.w3.org/2006/09/14-ua-minutes.html People with action items: cl ja pp[End of scribe.perl diagnostic output]