User Agent Accessibility Guidelines Working Group Teleconference

30 May 2013

See also: IRC log


Jeanne, Jim_Allan, kford, Kim_Patch, Jan, +1.609.734.aaaa, Eric, Greg_Lowney
JimAllan KellyFord


<trackbot> Date: 30 May 2013

<allanj> scribe: KimPatch

<allanj> open item 1

Jeanne: showing mobile accessibility examples in UAAG – how would apply to mobile devices. Judy wanted something she could show. Original idea was pull them out to put them in a separate document.
... we published last week
... Brussels conference, want to start requiring mobile UAAG requirements. People really want specific mobile guidelines. Separate document is a problem because people are going to start thinking this is all there is to UAAG. So now it will just be a view of UAAG: show me mobile, media player etc.
... working on that. The good news is we have a lot of interest in UAAG right now. We can't give them what they want in terms of specific mobile guidelines. But I think our visibility is improving, particularly around this area.
... XSL transfer skills needed. I'm probably going to be publishing again on Tuesday. At the moment the plan is to publish with the re-coding I'm doing now.

Kelly: attention is nice, it would be nice at the same time if they could help – throw bodies at it

Jeanne: we do have people asking if they can help. We were talking about crowdsourcing

<allanj> close item 1

crowd source Priority Sheet implementation list

Jeanne: we have one person who was really really persistent and wanted to help with UAAG, I asked him to look at the mobile examples, he did. We have four other people right now would like to help and we could give them something to do. Kim and I were talking about whether we could get people to start working on the implementation list, and how can we break up that work into manageable...
... chunks reasonably securely and starts crowdsourcing getting people to contribute implementation for interesting features

<allanj> kp: key to ask folks for a small favor. doing a shared document, here is what we need...be very directed

Jeanne: there are about 30 new iPhone browsers.
... so we are much more relevant and we do need help looking at all these browsers
... do we have ideas or ways we could do things. Spreadsheet is clumsy. Looking at programming skills of people on the call, what could we do

<allanj> have a much simpler/smaller spreadsheet to use very specific to finding implementations

Greg: spreadsheet is clumsy, but the alternative would be to have some shared non-spreadsheet documents that we could programmatically convert to a spreadsheet to manipulate because eventually we are going to want to have a spreadsheet to manipulate it. I guess it wouldn't be that hard to take a document – exported from whatever shared authoring platform we are using, and convert it.

Jeanne: we could do it on a wiki

<allanj> kp: simple spreadsheet. wikis are problematic.

<allanj> kp: use forms?

Greg: advantage to keeping it in a spreadsheet is sorting it in different ways

<allanj> kp: google forms, are very easy and fast

Greg: it would be important in whatever we do to have separation between imagine a spreadsheet for each browser and a row for every SC. Two columns, one standard and then free-form comments.so what

Jim: I have a survey tool

<Greg> That is, one column with a standardized code with answer (e.g. "Pass", "Fail", "Partial", "NA-Platform", etc.), and the other with free-form comment/description.

Jeanne: we can use the WC3 survey

Jim: can we make it public

Jeanne: yes
... one thing we have to watch out for, atag example misunderstood guidelines and failed

Jim: we have to verify answers

Jeanne: we have to do the testing anyway, but this is a nice lead-in to say my browser does this, my browser does this

Jim: subgroup to put one of these together?

Greg and Kim can help

<jeanne> ACTION: Jeanne to schedule a meeting for a subgroup to work on crowdsourcing project. [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action01]

<trackbot> Created ACTION-832 - Schedule a meeting for a subgroup to work on crowdsourcing project. [on Jeanne F Spellman - due 2013-06-06].

<allanj> wbs-group -- jim, jeanne, greg, kim

<allanj> were to advertise that the form is available?

Jim: when we get this done how to get the word out

Jeanne: Twitter, LinkedIn, Google plus

Jim: Greg, you seem to have a way to get the data down – if you could scribble that down for us

bug report from A11y Task force https://www.w3.org/Bugs/Public/show_bug.cgi?id=13624, feed back needed

Jim: bug report left over from HTML 5. this is about media players. There's the standard pause feature, there was some discussion that there needs to be an additional pause that some AT might set – like a screen reader might have an audio description and need a special pause for that
... some have said not necessary. To my knowledge I don't see the screen reader is ever going to talk to a media player and say stop while my user reads this. Do we have any specific issues with this particular bug – it currently stands that they are going to drop it unless we have some strong objections – that this needs to be in HTML 5

Jeanne: Kelly, if you pause a description, does it pause the player as well

Jim: that's my understanding

Jeanne: what would this be used for?

Jan: video going and text going on another part of the page that you might want to read up on. I don't think it's necessary these days

Kelly: imagine a webpage listing audio description and aria live updates with text

Jim: but they can still pause the video and read the text to their hearts content and then go back to the video

Greg: reading the bug I don't understand it. There are technical things referenced in the bug report that I don't feel necessarily correspond with what I think were signing it means

Jim: I'm checking out some of those references in there, they are saying this may work but we might have to change some of the accessibility APIs, now we're getting deep
... any objections to having them close the bug?

Greg: I don't feel confident saying it's not necessary because I don't understand it, therefore I don't feel comfortable giving my approval closing it – but that's just me.

Jeanne: I don't think we should say it's not necessary – I agree with Greg that's going too far. What's a good way of saying – we could say that the user agent working group didn't find a compelling use case for making this a requirement
... I think we could add that if there's a compelling use case to bring to our attention we can certainly consider it

Jim: this is been kicked around for almost 2 years and no additional comments on this bug. And getting accessibility APIs to change –
... so Greg, I think we've noted your objection and they will craft some wording and send it to the group before I send it off to the HTML 5 group

<allanj> ACTION: jim to craft a draft response to HTMLWG about the bug 13624 for UAWG review. [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action02]

<trackbot> Created ACTION-833 - Craft a draft response to HTMLWG about the bug 13624 for UAWG review. [on Jim Allan - due 2013-06-06].

Definition of browser - see "User Agent definition" thread http://lists.w3.org/Archives/Public/w3c-wai-ua/2013AprJun/0056.html

<Jan> http://lists.w3.org/Archives/Public/w3c-wai-ua/2013AprJun/0069.html

<Greg> I don't think "a software" is grammatically correct; perhaps change to just "software".

Jan: first change, because you get the situations I could do a calculator on my desktop that sends informational and renders an answer – that's not really a user agent
... Components are offered by different programs. They could be evaluated on their own merits by themselves and then you could then build on top of them so you wouldn't have to do a UAAG evaluation on them every single time. If the company was good at taking WCAG content it could do that for any component on it that uses WCAG content

<allanj> Note: There is an important distinction to be made between user agents that retrieve, render and facilitate end user interaction with arbitrary web content (in a given technology) and applications that retrieve, render, and facilitate end-user interaction with more restricted data sets. For example, a financial application that retrieves, renders and facilitates user interaction with an...

<allanj> ...XML-encoded stock price history would not be considered a user agent. Instead, if the application is constructed using web technologies, it would make more sense to evaluate it using WCAG 2.0.

Jan: just trying to find a way that every database backed web application does not become a user agent

Jeanne: should we not do that – there's a fair amount of pressure that we should do that

Jan: I'm just talking about web applications

<Greg> In the Note, should clarify that WCAG would only web-based apps, not native apps.

Greg: doesn't say web app, so applies to standalone as well, say Quicken

Jan: that can be tightened up
... this web component is pretty good for meeting UAAG and can be used as the basis for things

Jim: we have to be able to evaluate the thing

Jan: these are componentized user agents

Jim: my next question was does Apple have one of these for the iPad of the iPhone, and apparently they do

Greg: this is something that people can build a browser around

Jan: these are little browsers. To make a new browser you grab WebKit and you build the application around it, but it's doing 90% of the stuff itself

Greg: so is it an example of a web complement

Jeanne: I would call that a rendering engine not a view component

Greg: distinction between a rendering engine and a web view component?

Jan: allows you to run it natively on the OS with no other browser in sight

Greg: all these browsers today are baked around some rendering engine, correct – that is usually a separate component – usually integrated into, but can be used separately as well

Jan: I don't want to be the one person pushing this – if this isn't the right way to go

Greg: I don't understand it – what is the difference between a web view component and a rendering engine

Jan: my understanding issue open a shell application which can do almost nothing and this could be the only component that you drop in and it's basically like having a small stripped-down browser inside your application frame which makes your application basically a browser.

Jim: and then you write the widgets that say go get my airline application, but it's still the rendering engine that's getting the information

Greg: and that's a simpler process than having your app use WebKit?

Jan: my understanding is it is very similar to this.

Jim: Apple has one, Microsoft has one, android has one

<allanj> developer.chrome.com/dev/apps/webview_tag.html‎

<allanj> docs.oracle.com/javafx/2/api/javafx/scene/web/WebView.html

<allanj> msdn.microsoft.com/library/windows/apps/BR227702

Jan: or if UAAG just wants to talk about mainline browsers we could do that

Jim: pressure about requiring these little ditties to be browsers?

<allanj> https://developer.apple.com/library/mac/#documentation/cocoa/Reference/WebKit/Classes/WebView_Class/Reference/Reference.html

<allanj> developer.android.com/guide/webapps/webview.html‎

<allanj> developer.android.com/reference/android/webkit/WebView.html‎

Jeanne: no, pressure in policymaking right now is guidelines that are oriented to mobile apps, and that the mobile apps are not covered by WCAG because their interface is standalone. Discussion: if something is written in HTML 5 even if it's a standalone app whether that would be considered web content for the purposes of WCAG

<allanj> http://developer.blackberry.com/cascades/reference/bb__cascades__webview.html

Jeanne: I do think we've written guidelines that address the user interface – is actually probably a vulnerability in what Jan is talking about the user component because most of our guidelines are talking about the interface and not the rendering. About 75% user interface, 25% applies to rendering engine.
... what's pushing us to make this decision, glossary entry

Jim: it's at the top of the document

Jeanne: do we need this to be finalized to go forward, or do we leave it in a vague state

Jim: I think we leave it in a vague state until we get some direction from others

Jeanne: I don't want to lock out the mobile apps, I think what we have is valuable – I would still like to see us include mobile applications

<allanj> kp: mobile apps should be UA.

Jim: so WCAG may not put mobile apps in their charter?

Jeanne: if they take it on, they would have to start writing techniques which is a big deal. If they do it's not as important for us to

Jeane: I believe they are looking at both standalone web based applications and the standalone web applications that are written in HTML 5 but run on multiple platforms

Jan: their definition – any software that retrieves and presents web content
... UAAGs great – a superhero, but if you don't need that it's overkill. If you point to someone that is just a small application and say you are a user agent, not necessary

Jim: 5.1.1 – we also say the buttons in my airline app have to meet WCAG

Jeanne: can we postpone the discussion until after the CG meeting on the 19th

Jim: yes. Somehow we should be able to approach this so that we and WCAG can be in agreement about this. Based on what Jan was saying we are closer to that we just have to get our words right

<allanj> ACTION: jallan to resume the discussion of Definition of user Agent on June 13 and further discussion to follow [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action03]

<trackbot> Created ACTION-834 - Resume the discussion of Definition of user Agent on June 13 and further discussion to follow [on Jim Allan - due 2013-06-06].

Testing sub-group/writing tests

Jim: would be useful if someone would step up and say I will handle this

Jeanne: I wanted to do a sample test this week, but got caught up in other projects.

Greg: if I remember correctly email going around the different proposals for what these tests would look like

Jim: perhaps if we decided on a format

Greg: lot of good arguments for using what atag was setting

Jeanne: it shouldn't have to change much to work with the W3C testing tool but the point was to write tests that were clear instructions to a tester and then if we had to change how was formatted we could do that – it's easy to change the formatting, it's hard to write the language especially to make it clear and understandable

Jim: Jan, you said that atag had a lot of if, then's
... others a lot of videos etc. and you could see if you could turn them on etc.

Jan: for some things you need that

Jim: for a lot of the user interface things we would need something like that
... split out user interface things from a tag if then things

Jeanne: you did that and it was mailed out

<allanj> ACTION: jim to create a breakout list of UI components etc. that need a test page, and other SC that need IF/THEN type of testing statements [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action04]

<trackbot> Created ACTION-835 - Create a breakout list of UI components etc. that need a test page, and other SC that need IF/THEN type of testing statements [on Jim Allan - due 2013-06-06].

Jeanne: we've had a W3C fellow for a year who is working on testing is leading the whole testing effort mostly for HTML 5 and CSS3. One thing he said that was very important was if there are not a lot of areas that can be automated it's important not to automate them because people will tend to only test the things that can be automated and ignore the rest. And we all saw that in the early...
... WCAG where people could only test for four things and therefore those were the only things they tested for. Cautionary note before we start trying to create test pages.
... they are doing away with the manual tests
... had a long conversation with him, conclusion is we are not going to be able to automate it because we can't run automated tests on the user interface because it's a violation of security to do it – they can do it internally, but we can't

Jim: I'll finish that breakout lists, and perhaps some creative people will create tests.

Summary of Action Items

[NEW] ACTION: jallan to resume the discussion of Definition of user Agent on June 13 and further discussion to follow [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action03]
[NEW] ACTION: Jeanne to schedule a meeting for a subgroup to work on crowdsourcing project. [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action01]
[NEW] ACTION: jim to craft a draft response to HTMLWG about the bug 13624 for UAWG review. [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action02]
[NEW] ACTION: jim to create a breakout list of UI components etc. that need a test page, and other SC that need IF/THEN type of testing statements [recorded in http://www.w3.org/2013/05/30-ua-minutes.html#action04]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2013-05-30 18:32:10 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138  of Date: 2013-04-25 13:59:11  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found Scribe: KimPatch
Inferring ScribeNick: KimPatch
Default Present: Jeanne, Jim_Allan, kford, Kim_Patch, Jan, +1.609.734.aaaa, Eric, Greg_Lowney
Present: Jeanne Jim_Allan kford Kim_Patch Jan +1.609.734.aaaa Eric Greg_Lowney
Regrets: Simon
Found Date: 30 May 2013
Guessing minutes URL: http://www.w3.org/2013/05/30-ua-minutes.html
People with action items: jallan jeanne jim

[End of scribe.perl diagnostic output]