Techniques Task Force of the WCAG WG Telecon

Plan for addressing techniques and test suites for 1.1, 1.3, 2.4, 4.2

mc: techniques for the guidelines above

wc: techniques taskforce michael john mike david becky wendy, chris on call
... should take up test cases and techniques at same time to harmonize
... how to divy up the work???
... who's willing to do what, how much time do people have
... we have lots of people, if everyone took 10 techniques we could do it

bg: seem to be loosing people from wed

wc: resourses is an issue

mc: that may be good because we can focus

w: if you are doing a lot of action items on this call don't do much on Thurs

js: and vice versa

wc: next wed we need to talk about techniques....divy up work today. by monday issue summaries, harvest more on Wed. then update propsoals following week then following wed close
... i think some tech for 1.1 could be combined
... can't get into bugzilla

js: I got in the day before yesterday

wc: I'll ask gregg to restart that server

mc: I'll write up search tips
... take on tech for a guideline, in context of current, review each tech related to a guideline keeping in mind the discussions in main group thurs, discuss enndotes, bugzilla, then propose action, to close,change, create new combine etc

setp2: go through mapping doc, think of techniques that appear to be missing

css html and script

mc: test cases: go through the test cases for the techniques.. identify issues with existing test cases and propose additional test casses

wc: I have ideas for test cases in CSS
... perhaps test cases with tim

<wendy> ACTION: wac ask tim about creating css test cases [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action01]

js: jim allen, james craig might help

bg: we have no JS scripts techniques now

mc: there open issues about how to write techniques in relation to baseline
... havestin WAIG list ok to do also but not necessary

bg: but guidelines in flux

mc: make conditional techniques

js: we may close 1.1, 4.2, 2.4, 1.3 soon

wc: I'm tempted to take 1.1 cause I'm doing the guideline summary.

<Michael> ACTION: Wendy take 1.1 [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action02]

wc: labeling functional apps will work with Becky

bg: label is ot an alternative

wc: it is test that serves the same purpose

mc: label is label not an alternative

mc: will be alot of work - review list for info about GL plus techs and tests

wc: so makes sense to work with person reviewing the GL

dm: have been working on 4.2

js: but there is a proposal on the list - can start with that
... hope is that WG comes to consensus on proposal - then techs group can start work based on newly adopted proposal

<Michael> ACTION: David take 4.2 [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action03]

js: so hopefully don't have to sift through all the prior proposals

mc: so David is taking 4.2

<Michael> ACTION: Michael take 2.4 [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action04]

mc: MC will take 2.4

<Michael> ACTION: Becky take 1.3 [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action05]

bg: will take 1.3

<Michael> action 3 = David take 4.2, work with Loretta

<Michael> action 4 = Michael take 2.4, work with Yvette

bg to work with jc; dm to work with LGR

<Michael> action 5 = Becky take 1.3, work with Joe

mc to work with yh

wc: by Monday, may 2 we will all have sent an issue summary related to the techs; proposals for changes would be great
... on May 9 will submit proposals for changes (based on dicsussion on May 4)
... goal will be to discuss and close on May 11
... concerned that am travelling week of May 9

js: perhaps can coordinate with Ben?

wc: will be in Japan to will be ahead of all of you :-)
... will be at WWW conf and also speaking to JIS folks while in Japan

js: have sent draft of 3.1 proposal to JIS group since they are concerned with English language centric issues

mc: work is assigned; will send a summary of what we are doing; May 2 is this Monday - so not lots of time, plan accordingly

dm: this includes bugzilla stuff - but we can't access bugzilla right now

mc: will work on access

dm: only dealing with techs that are still related to gL

mc: sent emails with mapping of techs to guidelines - use those emails as starting point

bg: all CSS techs showed as not being mapped

mc: then probably not mapped properly
... so consider techs that aren't currently assoc. and perhaps propose associating them with a particula Gl

dm: have end to ends on my site that may help

mc: will try to include pointers to all sources in set of instructions

Requirements for Checklists and Techniques

<David_> http://www.eramp.com/david/end_to_end/index.htm

<wendy> http://www.w3.org/WAI/GL/WCAG20/WD-wcag2-tech-req-20050426.html

mc: sent a change log to list

<wendy> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0263.html

<David_> end to end links above have associations that I eyeballed from Guidelines to Techniues docs, (6 months ago)

mc: added info to intro about techs not being comprehensize - no garuntee we cover every possible tech and technology
... following techs is not req. to conform to WCAG - is a set of well thought out suggestions
... for most users is best way but not only way
... def. of reliably human testable - understood that want to remove "80% of human testers would agree" so rewrote
... added def. of postitive and negative tests
... pos test demonstrates proper applicaton of tech; neg test demonstrates improper application
... remove ref. to additional ideas
... added bullets about sufficeint and optional
... provided clarification of AND and OR
... removed req. for showing support by all AT (provide where approp)
... removed req. about untestable techniques
... added placeholder about baseline
... this is probably biggest thing I want to deal with
... removed appendix fields
... removed req. about general techs that are req. to conform

js: issues about req. for conformance
... positive test means test has been correctly implemented

mc: reads from doc about pos and neg tests

js: this should only be for techs that are sufficient
... may not need second clause, where tech is sufficient for conformance a positive test would indicate conformance - something like that

mc: wanted to say techs are a way to meet SC - a positive test is a good thing for access.

js: but pos. test isn't always a garuntee of conformance

mc: action to delete clause about positive tests and conformance

js: sufficient and optional bullets added?

<wendy> ACTION: michael delete clause about positive tests and conformance [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action06]

mc: reads info from doc

js: does that entail that we have to spell out benefit?

mc: why would we document an optional tech unless we thought it was worth it?

js: so are we requiring ourselves to say what the identifiable benefit is?

<Zakim> wendy, you wanted to ask in tech summaries, start to categorize the techs as sufficient, future, etc?

wc: if we can describe them well - we can also use the categories for a tech. summary categories
... if we want to categorize techs as optional, sufficient, future, etc

mc: yes want to categorize; need that for when the tech spans baselines
... at lowest baseline most techs are not optional but may become optional in higher baselines
... or may be optional in one baseline and sufficient in another or vice versa

wc: might be helpful if in the list of defs we list ooptional, sufficient,

mc: so action to create definitions

wc: other categories?

<wendy> ACTION: michael add definitions for sufficient, optional to section on definitions [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action07]

mc: analysis includes a not recommended category - where following tech was destructive to access.

js: so there are places were we tell people what not to do

mc: no - it is a tech for one baseline that is not rec. for another baseline

dm: do we want to say that a tech becomes optional in another baseline?

mc: req. doc does say tech must specify the baseline; probably should expand

dm: need to carefully define optional

mc: do we want to include the three baselines we have been discussing so far?

bg: do need to define them

js: instinct is to not put in req. since they may be somewhat fluid
... techs need to def. baseline and which baseline a given techs works in

<Michael> ACTION: Michael put information about relation of sufficient and optional in baseline section [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action08]

js: but don't think req. doc should contain the definitions of baseline

wc: req doc should be like a checklist to make sure we have done the all necessary work
... saying what we are going to implement without saying how

js: say techs will ref. more than one baseline but not what they are

mc: so current wording about techs referencing the baseline it applies to is ok

js: yes

mc: but still need to define baselines and where to include those defs

js: need to figure out grouping etc.

mc: so should add an agenda item for next wed. to discuss baseline

wc: there will probably be issues related to baseline in our techs issues summary
... still see the 80% figure in the req. doc

mc: didn't remove it - wasn't sure if that was the action -
... at first deleted the sentence - then just updated the wording

dm: we are giving an exact % for an arbitrary number

js: actually 80% is a common target

dm: what is the exact term

js: inter-rater reliability
... one example is if hand 5 people the same set of content they will agree on the rankings from best to worst but might give diff. numerical scores

mc: we are tangenting a bit - do we want to zap the 80% testable by humans sentence or not?

js: will take action item to see if there is agreement about % cut off

<wendy> ACTION: john research "inter-rater reliability" - is there a % cutoff point? [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action09]

dm: can get a book about inter-rater reliability for $39.99 :-)

mc: is there more on req? do we need a break?

wc: suggest putting it on agenda again so people have more time to review

mc: will update based on actions from today's call and repost req. doc

<David_> book called "handbook of inter-rater reliability"

<David_> http://www.stataxis.com/interrater.htm

Mapping of script techniques to guidelines

http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0249.html

bg: didn't map too many to 1.1. didn't go through existing script techniques.
... went through a list apart, etc. tried to propose techs for 3 categories: 1. don't have scirpt, 2. can help w/accessibility 3. scripts not helping accessibility, but common scripts and how to do accessibly.
... only additional in noscripting e.g., about making accessible pop-ups (using real links instead of javacript uris). mapped to 2.4
... felt a bit of a stretch, but using a tags appropriately.
... could probably go under 4.1 as well.

js: re: 2.4, is it supporting navigation?
... under 3.2 have extreme changes of context.

bg: tech is not to show how to mark extreme change of context.

js: if it's about using a tag might be about 1.3

wac: 3.2 level 1 #1 - "Any extreme change of context is implemented in a manner that can be programmatically identified. " is using a in a way that the ua knows what's going on. thus "programmatically identified"

bg: the next group were to enhance accessibility
... focus to form element. 2.4, level 2, #4
... could also be #2 of that GL and level
... someone on the list said hadn't seen onload reliably.

mc: concept of setting focus is not under debate?

bg: no. someone else suggested "here's how i do it"

mc: like it, however, nothing before the focus is read.

bg: for filling out a form, is good.

js: google starting doing that/jaws supported, but field not labeled.

bg: leave it with one

mc: prefer to map to only one when possible. may have weak mappints to others. technique should only map to one.

bg: next is how to catch onchange of input, validate, and set focus on next element. sort of help ppl correct errors, but helps navigate sequentially.

mc: also helps when do validation to do not mess it up.

js: what if script validates and input wrong?

bg: leaves you in the same field. however, this hard b/c depends on the script.
... may not be able to come up with techs that are acceptable to the group. need code and testing...

js: what is the main thrust of the technique? helping users avoid errors and make ease to correct? or navigating sequentially?
... then 2.4 l2 sc 2

bg: navigating sequentially
... it assumes you are doing validation. vs the other technique that says "move to the next field when there are 3 characters"

js: the 1st one goes under 2.5 (making easy to correct errors)

bg: say that it's about validation

input assist to change the background color or border of the element with focus [GL 2.5 L2 SC1 (If a user error is detected, the error is identified and provided to the user in text ) although color and border isn't really text and this technique isn't specific to errors.]

bg: could be "help navigate" or "error handling"

mc: could also set text (not only color)

bg: adding text effects the layout.

input assist auto advance through fields. Fore example, a US phone number input with 3 fields as you finish typing 3 numbers in the first

field the cursor moves to the next field and after three numbers there jumps to the last field. This one is likely to be controversial as I'm sure there are some people that dislike the auto advance behavior. [GL 2.4 L3 SC2 (When a page or other delivery unit is navigated sequentially, elements receive focus in an order that follows relationships and sequences in the content.)]

bg: some ppl like this one, others don't.

js: in an app, a user preference?

<scribe> ACTION: david send bg techs for 2.5 [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action10]

mapping makes sense

input assist to show a specific format in the field and user input over types the format with the actual value. For example, a field for a US social security number that shows "###-##-####" - as the user types the # gets replaced by the actual input. [GL 2.5 L2 SC2 (If a user error is detected and suggestions for correction are known and can be provided without jeopardizing security or...

scribe: purpose, the error is identified and the suggestions are provided. ) This technique is not specific to errors, though. ]

bg: instead of 3 fields, one field with format.
... put under 2.5, however the user hasn't created an error helps prevent an error.

<jslatin> a mechanism is available to help users enter data correctly

looks like we need to propose a success criterion in guideline 2.5 at level 2 or 3

mc: info needed to fill out the form is before the form

wac: google suggests - need info about how to fill in. not sure that "###-##-####" will be intuitive to users.

question about creating a SC that might not be met by HTML. although, seems that there are ways to do it in HTML.

wac: next steps?

figure out how to create an accessible pop-up (NOT a new window) that can be made accessible. I'm not sure how to do this, yet, but it could be used to provide more detailed information or help without leaving the page. [GL 2.1 L1 SC1; (All of the functionality of the content, where the functionality or its outcome can be described in a sentence, is operable through a keyboard or keyboard...

scribe: interface. ) and GL 2.1 L3 SC 1; (All functionality of the content is designed to be operated through a keyboard or keyboard interface. ) ]

<David_> http://www.eramp.com/david/general/

wac: extreme change of context?

<David_> above is general tech proposals for 2.5

discussion about how this could map to a variety of SC depending on the example. could also map to 3.2 (if avoiding an extreme chnge of context) or 2.5 (if helping user correct an error)

bg: pretty open ended. perhaps too early to map.

convert information provided via <link> elements into a select list on the page. From alistapart.com entry: Dynamically Conjuring Drop-Down Navigation [3] [GL 2.4 L1 SC1 (Structures and relationships within the content can be programmatically determined );GL 2.4 L2 SC 1 ( Documents that have five or more section headings and are presented as a single delivery unit include a table of...

scribe: contents with links to important sections of the document. ); and GL 2.4 L2 SC2 (There is more than one way to locate the content of each delivery unit, including but not limited to link groups, a site map, site search or other navigation mechanism. )]

mc: tech to get around UA deficiency

bg: also navigation

js: level 2 under 2.4 (multiple nav mechanisms)
... l2 sc2 of 2.4

provide alternative text sizing on the page (in combination with CSS). Here is an example for alistapart: Power To The People: Relative Font Sizes [4] Warning - it uses JavaScript uris :-) [This seems like a helpful technology but I can't find a mapping???]

js: take it for granted that text is perceivable (by default). perhaps a criterion that text is actually perceivable.

mc: related to tech on absolute size, have been unable to map them.

dmd: the only reason that we recommend rel is that there is a bug in ie?

mc: css mapping to 1.3

need a criterion about font size and reflowing content when font size increases.

js: fits under 3.2?
... b/c talk about consistency and design and effect on layout
... want to allow for changes in font size w/out blowing up layout

bg: it's hard to do.

js: specify a threshold, beyond which it is not the designers responsibility

dmd: address what % users can increase font size?

wac: what's the general rule of thumb from moving from increasing font size in browser to using magnification software?

<scribe> ACTION: wac new issue for 3.2 related to increasing font size [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action11]

formatting table rows to distinguish one from another. This should help the readability of all users but particularly screen magnifier users if the table is wide. Another alistapart example: Zebra Tables[5] [GL 1.4 L1 SC1 (Any text that is presented over a background image, color, or text can be programmatically determined. ) although that is a bit of a stretch since the technique itself...

scribe: is using color]

discussion of benefits - people who use magnifiers, people with learning/reading disabilities

mc: may fit under understandable

js: perhaps an optional technique under 3.1 or 2.4 (re: making tables more navigable)
... perhaps another mechanism for locating content

discussion about variety of places it could map to.

wac: depends on primary benefit - if learning disability, 3.1. if magnificiation, orientation and 2.4

<scribe> ACTION: dmd do either magnification tool have ability to zebra stripe [recorded in http://www.w3.org/2005/04/27-wai-wcag-minutes.html#action12]

dmd: i've been looking re: inter-rater reliability and 80% seems reasonable number. 70% they say is low, 90% is high. sent john articles have found.

next week: issue summaries for techniques/test cases, baseline, requirements, script mapping