20 Sep 2016


kirkwoood, JohnRochford, kirkwood, Mike_Pluke


<Lisa_Seeman> scribe: kirkwoood

<Lisa_Seeman> start with intorductions

<stevelee> Lisa_Seeman: I need the webex pwd pleas

<Lisa_Seeman> i put in in a 1 on 1 chat steve

<Lisa_Seeman> let me know if you have problem

<Lisa_Seeman> steve did u manage with webex

<Lisa_Seeman> use the link on the tpac page

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/gap-analysis/

<kirkwood> LS: Gap analysis based on two different user resarg works based on materials from University documents and other research

<kirkwood> LS: issue papers based on horizontal problems

<kirkwood> LS: looked at security

<kirkwood> LS: short term memory etc

<kirkwood> LS: made proposals for WCAG and alos a reoadmap. The gap analysi

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/gap-analysis/table.html

<kirkwood> LS: interesting thing is table of user neeeds . URL in irc

<kirkwood> LS: we had gap analysis and just approved second revision

<kirkwood> LS: here's sehat they are sections: authentifiction safety, etc and p=split into tables to make manageable

<kirkwood> LS: first thing is a proposal to WCAG,

<kirkwood> next thing is any authoring techniques

<kirkwood> LS: next thing is proposed new semantics. which you need user settings

<kirkwood> LS: next thing is new symqntics

<kirkwood> LS: semantics

<Lisa_Seeman> test

<Lisa_Seeman> technical issue

<stevelee> i'm in and can hear you

<stevelee> but you seem not to hear me

<stevelee> yes - and then I could hear you

<stevelee> strange - works normally

<stevelee> I can hear you

<stevelee> trying all the options

<JohnRochford> I will scribe for awhile if needed.

<JohnRochford> Lisa: Let's take a look at Table 3. Does it contain all the user needs we can think of?

<JohnRochford> Michael: Suggests people try logging in again to IRC.

<kirkwood> test

<kirkwood> LS: if i do make a mistake I need to know how to correct the mistake and it shouldn't be too scary to make it

<kirkwood> LS: need enough time, dif does timeout need to not lose work

<kirkwood> LS: want to use applications or API to help, including spell checker that we could add to user need

<kirkwood> JB: UX issues

<Lisa_Seeman> scribe: JohnRochford

Lisa: Always overlap between accessibility and usability.

Jim: Isn't avoiding mistakes something everyone wants?

JohnK: We could define disability as impeding life tasks.

Lisa: There are situations in which people without disabilities would be in situations that mimic (or similar to) those for people with disabilities.

John: (That last was awkward, but I don't know how to make it better.)

Lisa: We have a few success criteria that are being reworded.
... Let's start with context-sensitive help.

<Lisa_Seeman> rewording https://rawgit.com/w3c/coga/master/extension/rewroded%20sc%203.html

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/gap-analysis/table.html

Lisa: Do we want to talk through it, or review it and say what you think?
... If an input error is detected, an automatic correction is made where reilable and possible.

reilable = reliable

Lisa: Error messages should be clear.
... Do instructions include error messages? (A or AA)

Jim: Errors must be understandable else severe frustration.

Michael: Error messages can be acknowledged by the user.

Jim: Error messages (at level A) should stop you in your tracks.
... Error messages at AA should be informational.

Lisa: If WCAG wants to shorten it, they can, but at least they will know what we are talking about.

<kirkwood_2> Lisa: "error messges that require response to continue" should be A

Lisa: Now talking about rewording 3.3.2 as in Table 4.
... Working on language. A user study is needed with people who have attention and memory issues.

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/extension/rewroded%20sc%203.html

Lisa: Now talking about 3.3.2 - "TODO - add Identify Charges"

JohnK: Someone with a cognitive disability would not attempt a process if extra charges are applied on the back end.

Lisa: They might if they do not want to go through the whole process again.

Clarification: might pay an extra charge

Jim: There are some organizations that will not care if they are AA or not; they just want the money.

Lisa: Explaining example of wheelchair purchase that automatically added another wheelchair.

JohnK: "Fine print" page is usually the last.

Jim: Some conditions could be forbidden.

Lisa: Perhaps semantics are added to fine print such that the tagged portions appear at the top of a page (via a browser extension).

Jim: Who decides what is important information?

Lisa: We do by defining it.

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/extension/index.html

Lisa: important information definition: 1) information the user may need to complete any action or task including an offline task. 2) information the user may need to know related to safety, risks, privacy, health or opportunities.

JohnR: Low vision task force is using our definition of important information.

Lisa: Info garnered from personalization must be kept absolutely secure, else people with cognitive disabilities could be victimized.

Must step out for a minute. BRB

Lisa: Not everything in the user research has made it into the personalization table.
... May ask everyone to check their user research to make sure everything made it into the personalization table.
... Let's have a look at Undo.
... Undo is in error prevention.

Jim and JohnK: Lisa's undo example may not actually be an undo, but a modify,

Lisa: Undo is also for mistakes.

Jim and JohnK: Users should have to be taken back as few steps as possible in a transaction.

Lisa: 15 minute break

<Lisa_Seeman> starting agian

<Lisa_Seeman> anyone on the phone who can scribe

<Lisa_Seeman> we are joining apa

<Lisa_Seeman> in the apa room. we will leave our other phone line open so that we can get back afer lunch m but now we are on the apa webex

<Lisa_Seeman> (see the scedule

<Lisa_Seeman> steve do you need the apa password

<Lisa_Seeman> we are calling into coga

<kirkwood> ls: writing SC will be good for presenting the case for the SC. To get ahead on the usercase. We can put in propsed wording and state the issue

<kirkwood> more importsnt to make sure everying in

<Lisa_Seeman> scribe: kirkwood

Michael: rewording SC, want you want to do SC naled down from WG by January or so should do a walk through of SC and roadmap

MC: the walk through doesnt need to be a big project. Once in the beginning and once at the end
... i am certainly not guarantee to be in 2.1

LS: it's all pretty important

MC: each TF is proposing a lot of success crtieria not sure if we'll be able to incorporate it all

LS: we put things in A and AA quite meticulously

MC: you might want to document the rationale for conformance levels

LS: when we had some things A rather than AA to put into content. Andrew says there could be procedure.

MC: we care about objective testability
... finding navigation elements, as a critical feature you need to define what it means
... trying to get into examples for it being testabel. What would a tester be looking for wether its a human or bot.

SC: if you saying level A and you need a rebuttal in hand if want to move to AA. If you have documentation you will be able to have that conversation

MC not SC

<Lisa_Seeman> scribe: Mike_Pluke

Testability for timing adjustable is concrete (good) but maybe awkward to stop and restart

MC Personalization test has some clear testable parts (at the beginning) but fuzzy parts (at the end)

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/extension/timed-events.html

<Lisa_Seeman> https://rawgit.com/w3c/coga/master/extension/minimize-errors.html

<MichaelC> If an input error is detected suggestions for correction are known, then the suggestions are provided to the user or automatically made where reliable, unless it would jeopardize the security or purpose of the content.

<Mike_Pluke_> MC: Personalization testabiliity. Initial part very concrete the last part too fuzzy.

<Mike_Pluke_> Key element: Key part of SC is that if errors are automatically they are automatically corrected when the correction is reliable

<Mike_Pluke_> "detected" should follow "automatically" in above

<Mike_Pluke_> Debbie has joined

<Mike_Pluke_> Dan Burnett editor and member of multimoda

<Mike_Pluke_> Debbie: Interest in communication between people who have different levels of ability with language

<Mike_Pluke_> Debbie: Authoring with ability to code different levels of complexity

<Mike_Pluke_> Debbie: Demonstrating "Emma"?

<Mike_Pluke_> Emotion ML is a markup to the content. Emma includes the input and output pof content in many ways.

<Mike_Pluke_> Emma a broader way of describing human input and output than EmotionML

<Mike_Pluke_> Describing EMMA: Extensible MultiModal Annotation markup language Version 2.0

<Mike_Pluke_> Instead of "probability of precipitation" can say "it is likely to rain"

<Mike_Pluke_> Measuring text complexity with a number is really a major challenge

<Mike_Pluke_> Debbie describing xxxxxxx Kincaid (a measure of text complexity) - not bad

<Mike_Pluke_> "xxxxxx Kincaid" in above should be "Flesch–Kincaid"

<Mike_Pluke_> LS: Explaining how an SC that addresses blocking errors would be A, others AA

<Mike_Pluke_> LS: Explaining how words like "simple text" used in our SC have been defined

<Mike_Pluke_> LS: Frequently used words in a specific context can be used even if not in the basic set of words

<Mike_Pluke_> Debbie: Possibly refer to "contemporary" words.

<Mike_Pluke_> LS: Not a fan of the use of "contemporary"

<Mike_Pluke_> Have to know what people are likely to say for voice recognition - very time and subject specific - one of the most difficult things to discover

<Mike_Pluke_> Debbie: Misuse of pronouns will have a big effect on intelligibility.

<Mike_Pluke_> In China they have a common pronoun to mean he or she.

<Mike_Pluke_> When you want to alert somebody you have to choose a very distinct word.

<Mike_Pluke_> The frequent words are not always the best to use for alerts/instructions

<Mike_Pluke_> Debbie: Emma has image and audio output as well as written language

<Mike_Pluke_> LS: Could use EMMA in a technique

<Mike_Pluke_> Debbie: Continuum in language complexity

<Mike_Pluke_> LS: We need to say conform/not conform

<Mike_Pluke_> Dan: Require or encourage, that the specification writer consider or develop a complexity measure for their expected audience

<Mike_Pluke_> Mike: Self defined complexity level for expected audience is a get out clause!

<Mike_Pluke_> Joined ? (speech clipped) from Microsoft

<Mike_Pluke_> Jason: Saying its about how to get to content and then use of it

<Mike_Pluke_> Jason: What can be done at the browser level irrespective of the authored level (not place burden on author)

<Mike_Pluke_> Jason: ARIA only used on a 6% of websites - so placing burdens on authors is unlikely to work

<Mike_Pluke_> Lisa: Personalized placing of buttons on websites might be achievable.

<Mike_Pluke_> Jason: The large predominance of JSscript sites makes the problem greater

<Mike_Pluke_> Jason: If can get the text out of the content it would be possible to process it in ways that suit the user's needs.

<Mike_Pluke_> Jason: Good to push more of the solution onto the browser because there is a lack of suitably knowledgeable authors to guarantee good practice.

<Mike_Pluke_> Jason: Browser does track where the user has been and navigation and re-tracing can be helped at the browser level

<Mike_Pluke_> Jason: Chrome extension model helpful when it comes to providing solutions that fit the needs of small segments of users

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2016/09/20 16:04:28 $