13 December 2001 WCAG WG Minutes


Loretta, Paul, Gregg, Jason, Wendy, Andi, Katie, Eric Hansen


Cynthia, Charles, Kynn, Gian, Annuska, Jo


SC = Success Criteria

CP = Checkpoint

Summary of action items and resolutions

SC for CP 2.2

"similar layout for user interface components is used throughout your site"

JW Seems testable

GV Want things to be similar but not the same. If same, difficult to figure out where you are. Could you have a site that is consistent and predictable that does not satisfy all 5 of the SC? Don't know what consistent means in the context and how predictable.

ASW Difficult to come up with sufficient SC for this CP?

GV Yes.

ASW I agree with that.

WC Reads from Designing the User Interface by Ben Schneiderman, pg 74

Strive for consistency. This rule is the most frequently violated one, but following it can be tricky because there are many forms of consistency. Consistent sequences of actions should be required in similar situations; identical terminology should be used in prompts, menus, and help screens; and consistent color, layout, capitalization, fonts, and so on should be employed throughout. Exceptions, such as no echoing of passwords or confirmation of the delete command, should be comprehensible and limited in number.

GV Would like to see that added. When have lots of things you could do, that gets into techniques rather than success criteria. SC a subset of objective things that give the designer flexibility. Beyond SC, "it is also recommended you look at the following,"

LGR and ASW agree.

JW ATAG and UAAG, required and "strongly suggested" are both included. Similar to GV's suggestion.

GV Solves one issue, path to second. Make SC objective. If lose something, goes to recommendation category. Reading all 5 SC. "your site" what is it defined as? is it a URL? 5 companies might have diff interfaces (if all appearing on one "site"). Ever define "site?"

KHS No, we need to.

GV Also have parts of our site that are like different buildings - different functions, different look and feel.

JW Another problem w/first SC, if it is not have layout, e.g. a voiceXML site.

WC Consistent sequences of actions should be required in similar situations - could be applied across languages/apps.

JW Could be added, but doesn't solve "layout" issue.

GV throughout site is problem. replace "layout" with "layout or flow."

JW Best to separate those two.

GV Let's talk about the concerns that people have w/the current SC and figure out "what is the essence."

PB SC 1 and 2, are UI and not directly re-lated to the CP. Interface could be result of user action. Result of user action, go to another view. What the view looks like is related to but not direct result of user action. 3 and 4 fit, but 1 and 2 do not really fit.

ASW Agree w/PB. Don't agree with SC 2.

WC Same button throughout site, labeled the same

GV e.g. search button shouldn't be called search on one page, find on another.

ASW Covered by SC3?

GV Consistent and predictable...consistent is means to an end. predictable is what you want. Provide consistent to make it predictable.

ASW If always predictable, do we need 2.3?

WC Informative feedback, dialogs that yield closure, error prevention.

ASW Standard usability practice. Used to have in our checklist, use standard actions.

GV #4 SC in this CP.

ASW move bar up to grey area (usability), where will we draw the line?

GV If talk about cognitive, can not make to all people w/cognitive disability. At the extreme you have people who are comatose. Trying to draw line lower than people w/130 IQ. What do we want to mandate?

  1. Make it predictable by following convention.
  2. Make it predictable by telling people in advance (if not conventional). Needs to be more than documenting, needs to help user when they encounter.

GV If you do that, say "do one or the other."

PB A specific scenario: If I am applying online for a car loan. Fields are not explanatory, but link next to that gives more info. Sometimes, I fill out form, but don't want to lose what already filled out. Open new window, copy&paste link into there. Sometimes get the window. Sometimes they use javascript and i get a blank window. Probably the best way to know, is a note that says, "open new window." Would that fit under SC#5 (document). Document means to me, "separate document." I'm thinking of 2 word warning.

WC CNN and Optavia are examples.

GV Yes, that would be warning or explanation.

JW Yes, SC#5.

ASW Sounds like 2.3.

PB True.

WC Goes back to ASW's earlier point: if you satisfy 2.2 do we need 2.3?

GV 2.3 is all about predictability. Extreme case of 2.2.

ASW Familiar things should act in predictable way. Something new could also be predictable, if based on real-world model.

GV How would you know if predictable unless sniffing? Mac or Win or Unix?

ASW I read that, to be ubiquitous, checkboxes, radio buttons, etc.

GV Language the app is written in, since it doesn't matter to the user....We had consistency, conventions, and warnings/instructions.

WC reads from list of questions from "Testing Applications on the Web." pg 158 - questions about user interface. e.g. metaphors.

GV Read recently about something that says don't use metaphors. Can cause misunderstandings b/c people carry the metaphor too far. Good point in what you read, metaphor useful to user or designer. Needs to be useful from user.

GV Can we agree that 2.3 is a subset of 2.2? Should 2.3 be a SC of 2.2? Give them a control or give them a warning. What do we mean by control?

WC Instead of automatically updating, give them a link.

JW Or let them set it up. But, going through that set up could be a higher cognitive load.

PB switching from HTML to PDF. Give them choice to click on HTML or PDF is simple example. By providing 2 links, you give them control.

GV True.

JW Trying to separate 2.2 into SC that more adequately accomplish what we're trying to do. Also, list additional items under it, not as SC but as suggestions. More commentary in Techniques.

GV Good summary.

GV If you have consistent and predictable, you can't have any adventure games.

WC Don't see that. Think all apps are predictable.

GV e.g. game. sometimes the phone does something, othertimes it does not. a book may click on a wall, but same book 10 minutes later will not. How do we include that caveat?

WC General caveat on scope? In content, made caveat for poetry, jokes, etc. and language usage. As with built environments, more interested in public services? However, if people want to make "art" or private site accessible, here are the things you can do.

GV When non-predictability is intended effect...

PB Would not limit to services. I think value of guidelines that can be used by anyone is good to let them apply to anyone. However, correct that exceptions do exist.

GV Too categorical. Captioning a spelling test, part of education. Perhaps, should be predictable, except where that is the effect. Poetry and humor, but not all literature. Therefore, need specific reason with a couple examples. It would allow us to keep stuff on the table.

JW Some time ago we discussed allowing people when claiming conformance that they had deliberately not implemented a CP since it was contrary to their content or purpose. This is similar to what we are discussing here. Although, trying to define that discretion more clearly. This raises the issue of whether our guidelines prescribe anything or entirely up to policy makers to determine what needs to be done.

GV Yes, sometimes we worry about putting abuse of caveats. However, some abuse should be expected and not sure we can completely prevent. It is better to have 10% abuse and server the rest of the 90%. Letting "air in" is probably not as bad as making something so tight people suffocate.

WC "predictability is intended effect" is good. Examples make clear.

ASW re: 2.3 - advertising windows? extreme change in context?

JW discussed b4, CS raised.

WC My opinion, if ad is pushed to the back, not an extreme change since doesn't change your flow of info. Just end up w/lots of windows at the end of the day.

JW The point of regard, or focus, exists in every medium I can think of. Could define it as something that moves focus.

WC UAWG spent a lot of time defining. EH still on the call? We should likely inherit those definitions.

EH A lot of work has gone into. Reuse it would be helpful.

Action WC: talk with KHS about definitions for extreme context changes from UAAG 1.0.

GV "Provide consistent and predictable responses to user actions" should be or rather than and. Either warn or be consistent. GV reads proposed SC. Collapses 2.3 into 2.2.

JW WC's from earlier, re: similar tasks and actions could be added.

GV Every time we make a rules we wipe out standard practice. Could we say something about when you don't use standard conventions.

GV similar layout (lose in new wording?), similar labels (have that), those that look the same act the same - really need to put in?

WC Overloading "more" as a link.

GV Could be appropriate. If a whole bunch of titles, say "war of the worlds - pdf, html, etc."

WC Right, although part of it is the context.

GV That's predictability.

GV (missed some) can always hit "undo" i.e., the back button.

JW We haven't explored cognitive ability that is necessary to use the Web. To what extent does that level of or kind of ability is required?

WC Can we ask Lisa to do more research on that?

JW gets into question of how you classify intellectual functioning and what's involved in using the web.

Action WC: talk with people about what assumptions we can make in regards to level of cognitive ability required to use the Web.

GV reads text aloud.

PB Different actions are reversible in different ways on different UAs. Even between diff versions of IE.

ASW Point to references.

GV's documentation of proposed text and issues:

Some working text we came up with in the telecon -- but which is NOT final. (some problems listed immediately below it. )

Checkpoint 2.2 Provide consistent or predictable responses to user actions.

  1. Where the response to user action will be different than general convention and is not always completely reversible using a conventional "undo" mechanism (e.g. back key) then provide the user with a warning.
  2. If user action will cause an extreme change of context (e.g. opening a new window) then either provide a mechanism to avoid the extreme change in context or provide a warning so that the user can defer from making the action.

Exception: Those instances where non-predictability is the intended effect such as gambling machines, mystery games etc.

UNSOLVED PROBLEM - what would the def of General Convention? - what would conventional undo - will these undo conventions work on all browsers - need to define "extreme change of context"

Next meeting

Action JW take next F2F meeting to list.

Resolved: Meet on December 20th not on the 27th.

$Date: 2001/12/14 00:55:39 $ Wendy Chisholm