MINUTES: W3C WAI User Agent Telecon 25 May 2000

Attendance

Chair: Jon Gunderson

Scribe: Jim Allan

Present:
Kitch Barnicle
David Poehlman
Harvey Bingham
Gregory J. Rosmaita
Madeleine Rothberg
Tim Lacy
Denis Anson
Eric Hansen
Dick Brown

Regrets:
Ian Jacobs
Mark Novak

Absent:
Mickey Quenzer Charles
McCathieNevile
Rich Schwerdtfeger
Al Gilman
Hans Riesebos


Action Items

Open Action Items

    1.IJ: Draft a preliminary executive summary/mini-FAQ for developers. 
(No deadline.)
    2.CMN: Propose a technique that explains how serialization plus 
navigation would suffice for Checkpoint 8.1.
    3.GR: Look into which checkpoints would benefit from audio examples in 
the techniques document.

New Action Items

    1.Editors: Update document based on MR proposal for control and 
configure and the resolutions made during this telecon
    2.Editors: Cross reference 4.8 and 4.10 and make clear that checkpoint 
4.8 for non-syntheisized speech audio
    3.GR: Research history of the priority of checkpoint 4.8 on audio volume
    4.EH: Propose new definitions for control and configure

Completed Action Items

    1.WG: Read Madeleine Rothberg review of the UAAG related to control and 
configure and re-read the guidelines document subsituting control for
      configure (except in Guideline 10).
      http://lists.w3.org/Archives/Public/w3c-wai-ua/2000AprJun/0354.html


Minutes

PR#284: Consistent use of the words "control" and "configure" in checkpoints

http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#284

JG: mr reviewed control & configure, control is set dynamically, configure 
is set up before, control is short term configure is persistent some want 
both when
a control is set it can become permanent. set parameters before something 
starts (playback rate in video) mr most check points can be control jg says use
control for all, then have configure option in gl10-default user configuration

EH: what is wrong with mr definitions?

DB: mr seems reasonable

JG: people like distinction

DB: yes

DA: important, some things need to be configured. mr analysis was good tl:

DP: if you want both then say both

EH: key thing, if priorities are same, then need to be accurate with 
priorities. if difference between control and configure priority then match 
checkpoint
priority. would be nice if mr would look at need for control and configure 
in checkpoints. unease with locking in developers to tightly

JG: control and configure size of text. make it 2 checkpoints? need minimum 
requirement for priority 1 items 10.7 save in profile (p2) control is p1
configuration is p2 but can be seen as part of 10.7 --last control setting 
will be saved as configuration

HB: be careful, that last font was not for header but for normal text size.

JG: if setting only header then it should be persistent for headers. don't 
have a range for control of font sizes. group has tried for other checkpoints.

DA: don't always want last setting

GR: set UA to reuse last setting, or retain old settings,

JG: may be techniques, different OS may need different techniques - ij 
example of editing initialization file.

DA: what you want with configuration-this is the way I want this to behave, 
need reset to default, last use may be special case, don't want it to be 
default
automatically.

EH: need many ways to set configuration

HB: in techniques have pointers to style sheets

EH: don't over proscribe what configuration means, unless all agree

DA: looking at priorities, control vs. configure, need to look at 
individual checkpoints to evaluate. mainly for control and configure are 
both important

JG: mr identified 3 cp for control and config. size of text, volume 
control/config size of text both p1?

MR: yes, checkpoint delineates

JG: is it impossible for some users to access content.

MR: p1 for one or the other

JG: p1 for control, need it dynamically config is for things that are 
difficult

MR: depends on what web authors are doing

GR: cognitive load for control is less than for config

DA: I think its opposite. config is set once and not worry

JG: when set control should be persistent for all pages in session

DA: look at things that require control of, may take 30 minutes to set up

DP: but may never use synthesized speech

JG: 10.7 not impossible

GR: opera persistent control of images don't champion one over the other,

EH: back to 4.1 control and config size of text. both p1

GR: yes, application in real world, config assumes UA source is properly 
marked up, client side CSS. if marked up differently then personal CSS doesn't
affect page, then I need local control.

JG: 4.1 in IE need to use two controls,

JA: IE control font dynamically, but must have "ignore font size" configured

Resolved: Checkpoint 4.1 is p1 for both control and configure.

JG: control configure for audio rate- p1

GR: yes

DP: yes

MR: question for screen reader- have configure for rate, but not dynamic 
(control) rate

DP: starting to have control for rate.

GR: also starting to have control for punctuation - change dynamically

HB: change in middle of document

GR: yes, if working in IE and note pad, and change

DA: isn't this p2, don't have to have control

DP: talking about rate not punctionation, need quick way to fix rate 
problems, and a way to make it persistent, need both, so it is p1, config 
is a global
setting, there are times that rate needs to be adjustable quickly depending 
on the task

JG: webspeak has control of rate, hpr may have control also, both have 
configure.

GR: not just a blind issue, low vision users, second language people

DA: no problem with control and config

JG: p1 for both control and config audio rate

DP: public ally available speech synthesis, have little or no control. such 
as Microsoft agent give audio but no control. important that we show developers
that control is necessary .

EH: over time we get an understanding about what is a p1, etc. key point is 
apply same standard across all guidelines, don't front load everything just
because we like it. implementation becomes difficult, if can only implement 
50% then which are most important

GR: I get lots of requests for setting up user work stations in public 
settings. lots of people use public terminals, they are looking into ADA 
compliance, must
have control

JG: first level is what is needed by pwd, spent much time working on 
priorities. few statements about priorities during last call, don't see 
much trouble with
priorities in document. have pretty good consensus.

EH: no problem with p1

HB: reset to default, should that be a technique

GR: yes

JG: user may not have access to that. has been dealt with before. good for 
technique for public access terminals. Have checkpoints to restore to default
10.7 p2

Resolved:Checkpoint 4.9 is p1 for control and configure

JG: 4.7 config and control audio volume p1

DP: control is p1, similar to reasons for rate

JG: what about hearing impaired

MR: hadn't thought about it. if no text with only speech then it is relevant.

DA: with synthesized speech isn't synthesizer with client.

JG: if using telephone, synthetic speech is in host

EH: control and configure for 4.10 is p1

JG: yes

DA: would set home terminal for something I can hear, public terminal may 
have default median volume

DP: synthesizers without front end from the author have not control. synth 
interperts the markup and outputs, if no markup for control of volume then no
control.

DA: is an imbedded object

GR: acss have aural properties (lists properties), many things built into 
w3 spec

JG: control is p1, what about config?

GR: config, most screen readers use sound card cannot set volume through 
screen reader, must use OS

JG: can control volume through sapi, actual volume is sapi, master control, 
and manual speaker volume

GR: is this a case where OS control covers all audio control

JG: this is a technique, dependent on OS, control must be accessible

GR: ok

JG: gr is config as important and control

GR: I lean that way

JG: note that for regular audio, 4.8 talks about audio volume (not speech) 
this is a p2, mr recommended p1 for control and configure. other audio events
must have text equivalents

EH: language used in regard to audio, depends on how terms are used. 4.8 
audio is a generic term, includes synth. speech

JG: long discussion on this

EH: need to define better, prerecorded vs. synthesized, discuss in exactly

DP: syn. speech is a special case of audio

DA: syn. speech may need special control, different from general audio

GR: control is situational.

JG: is config a p1?

DP: should be same as rate. same situations, need same type of control and 
config--should be p1

JG: objections?

DB: situations where you can configure but not control makes it impossible 
to use.

JG: if volume is so low that a person with hearing problem cant hear it 
then can

JG: most systems provide config but not control

DB: talking about UA not AT,

GR: if they choose to do this then these are the parameters

DB: I know it is not hard to do, little uneasy about it being impossible to 
use thus requiring a p1

JG: it is a present value, was reviewed as a p1, no comments during review 
process.

DB: I can live with it. not positive that it meets the impossible rule. can 
imagine many scenarios for all GL that can make all be a p1

JG: many public access computers, allow users to configure. set up a 
system, so individual profiles follow person around. then can config.

DB: if you log in than your last used profile

DP: if you cant log in then must be able to config/reset values

DB: cant control because you cant hear it

JG: system volume, get back to default value. argue about different 
situations where config or control is better or more needed. similar to speech
playback--volume and rate are important for hearing impairment.

DB: can live with it, move on

Resolved :4.10 control and config p1

rest are p3. also add control and configure to 4.8 as a p2

MR: already a p1 need for text equivalent. must have p2 to control and 
config audio volume

GR: if have page playing midi file, plays at a volume that cant hear 
synthesizer. OS control globally controls all volume. ability to control 
volume should be p1

DP: should be p1

JG: address both now

DA: if config/control speech is p1 then this can be p2

GR: cross reference 4.10 and 4.8 add note "if user agent supports sny 
speech natively then changes in audio volume should be independent of syn 
speech
volume"

JG: minimum requirement that volume controls are independent.

EH: sounds fine to me

JG: how do I know when I comply. group working on this. 4.8, 4.11 
independent control, two type of minimum requirements--provide structured
navigation-minimum things are defined, another type is provision of a range 
- such as change rate

KB joins don't have range for audio (hardware limitations). syn speech 
playback rate was specified - use default of synthesizer else offer a range

EH: on 4.8 divergence on priorities. p1 control, p2 config.

JG: need action, why 4.8 is p2, check archives

Action GR: why 4.8 is p2, will post to list

GR: recollection, speech is p1 and audio is p2 is for reasons stated, audio 
overwhelms synth. speech.

EH: need to define that 4.8 is non synth. speech. or audio other than 
synthesized speech with pointer to 4.10

Action Editors: cross reference 4.8 and 4.10.

Resolved: independent control of synthetic speech and audio volume.

Action Editors: 4.8 indicate non-synthesized speech audio.

JG: combine volume control for audio and synth speech. must have 
independent control as a minimum

GR: originally 4.8 was turn on and off background audio.

JG: have separate check point for that adopt mr proposal

Resolved: Use the rest of MR proposal for changing the wording of control 
and configuration checkpoints

Action EH: Propose new definitions of control and configure.

2.PR#283: Delete checkpoint 10.4 Allow the user to change the input 
configuration.

http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#283

JG: Checkpoint 10.4 currently does not have any clearly identified minimum 
requirements. Its minimum requires seem to satisified by other checkpoints:

    1.We already have a requirement that all functions be accessible 
through the keyboard (1.3)
    2.We say to use system conventions (5.8)
    3.Allow single key or command configuration (10.5)
    4.Configuration of graphical controls (10.9)
    5.Saving custom and restoring default configuration information 
(including keyboard bindings) are in checkpoint 10.7

What additional functionality does checkpoint 10.4 add that can be put in a 
minimal requirement for satisfying the checkpoint? We need just one requirement
of a minimal specification to satisfy the checkpoint that is not covered 
somewhere else.

NOTE: I would modify the minimum requirement of 10.5 to include either 
single key or modifier + single key for configuration of systems with 
keyboards.
This would generalize 10.5, but still provide the single key functionality 
for those that want just a single key.

EH: for 10.5 need to describe minimum requirements. so many steps. define 
functionalities.

JG: all things that need control need keystrokes. part of minimum spec for 
10.5

EH: is there language for 10.5 minimum requirements. they are posted for 
Netscape.

JG: I will look and post something. 10.5 is next issue.

KB: jg turning in 508 comments

JG: yes

EH: are you KB

KB: yes have lots of work
Jon Gunderson, Ph.D., ATP
Coordinator of Assistive Communication and Information Technology
Chair, W3C WAI User Agent Working Group
Division of Rehabilitation - Education Services
MC-574
College of Applied Life Studies
University of Illinois at Urbana/Champaign
1207 S. Oak Street, Champaign, IL  61820

Voice: (217) 244-5870
Fax: (217) 333-0248

E-mail: jongund@uiuc.edu

WWW: http://www.staff.uiuc.edu/~jongund
WWW: http://www.w3.org/wai/ua

Received on Thursday, 25 May 2000 16:24:21 UTC