WCAG 2.0 Evaluation Methodology Task Force Teleconference

10 May 2012

See also: IRC log


Kathy, Don, Shadi, Sarah, Liz, Detlev, Samuel, Kerstin, Vivienne, Peter, Eric, Aurelien, Mike, Alistair, Maureen, Katie, Richard, Tim
Elle, Denis


<shadi> http://lists.w3.org/Archives/Public/public-wai-evaltf/2012Apr/0075.html

<shadi> http://www.w3.org/WAI/ER/conformance/ED-methodology-20120510

Comments Received

Public working draft has ended - so looking at 7 public comments

<shadi> http://lists.w3.org/Archives/Public/public-wcag-em-comments/

Eric - Interesting positive comments from the public, all of which will be placed in a document

<vivienne> +1

Eric - we will have a list of comments etc... which can be discussed in the next phone meeting

<ssirois> yes.

<Sarah_Swierenga> +1

<Aurelien_levy> +1

<ssirois> i loved that approach last time.

<Kathy> +1

<Kerstin> +1

<Mike_Elledge> +1

<MoeKraft> +1

<Liz> +1

Eric - this document should be ready for Monday

Eric - the next call will focus on the comments

Eric - how will the comments affect the draft

Peter - Not available next week

Eric - There will be more opportunity to talk about the comments on an on-going basis

Eric - You could enter any comments in the document, if you are unable to take part in the call

Peter - V.Busy

<richard> zakim ??P36 is richard

Eric - Point 3 - new version of editors draft

Eric - walking through changes

Eric - I made changes to earlier sections based on discussion - the changes have been documented in a email sent to the list

<Ryladog> Sorry for coming late

<Kerstin> shadi, no, don't think so

<Detlev> Do you have that echo, too?

<Aurelien_levy> Zakim P39 is Aurelien_levy

<Aurelien_levy> Zakim P39 is Aurelien_levy"

There is a feedback loop on the line...

<Detlev> maybe everyone should mute themselves?

<richard> p36 is richard

<korn> I've muted locally

Who is sitting in the Sydney Opera Hall???

<vivienne> there are a few people not muted - Sarah, Liz, Alistair and Mike

Locally muted

<korn> Shadi, can we simply mute EVERYONE except Eric?

<korn> Then folks can unmute when they get into queue?

mute me

<ssirois> oh! it worked...

<Detlev> much better

<ssirois> but now we don't know who it was! ;)

<Liz> mute me

<Kerstin> zakim is not my friend today

Eric - line ok

Eric - Added in section 3 - a web application could be a single webpage

Eric - Peter talked about a one page application

Eric - However, a single page could contain many components - so we looked at use cases

Eric - Difficult when looking at a web app to record urls

<Aurelien_levy> maybe we also can speak about component

<vivienne> We had to do that recently

Eric - Difficult to log the urls in a report

Eric - anyone have any ideas how this could be handled

Peter - web applications may only be on a single page

Peter - this might sum up the situation a bit better

Eric - Even if there are more than one pages - there still could be a several uses cases

Peter - you would have to describe the things which have been done on each page i.e. each use case followed

<Aurelien_levy> we can described the uses cases and we also can identify the component using xpath or id

Eric - How would you record this in a machine readable way

Shadi - Peter used the term screen instead of page - maybe this is better to talk about than page

Shadi - note - with a web application we could talk about screen instead of web page

Peter - forgotten comment, but like the idea

Peter - maybe we would want to consider a more formal definition of screen with respect to web apps

Richard - Not happy with doing evaluation on one or two pages - we should be looking in terms of websites

Eric - but what if the website has three pages which contain several applications

Eric - You could do 1000s of things through these pages - rather than just pages

Eric - we could add a definition of screen

Eric - do we know if this has been discussed in W3C

Katie - could be under the mobile one section

Peter - using mobile could be difficult - as it could relate to the concept of forms

Peter - the larger concept of 'screen' would be useful

Katie - in the old days the screen was called view port

Eric - send me a link if there is any further information

<Aurelien_levy> it's me again fighting with my connection

Detlev - Screen may not be the best term, what happens if you have hidden screens etc... could be confusing

Detlev - maybe state changes could cover it

Eric - could be worth considering it - maybe extending the definition of web page

Eric - step 5 a

Eric - Provide additional information about paths

<vivienne> Sorry Eric, where is that?

Eric - how do you describe paths which have been followed.



<vivienne> thanks, was not looking properly

<Aurelien_levy> I think Deque Fireeyes extension generate some scenario file

Peter - notion of screen raises important questions - what is a screen, accordion, refresh

<Mike_Elledge> Bye folks, have to go to another meeting. Ciao/Cheers/a bientot.

Peter - bullet at end of 5a - don't think we can define path until we more fully understand other concepts

<Sarah_Swierenga> 5a. last bullet makes sense to me

Peter - how important is machine readability

<Aurelien_levy> I think my mic is generating the echo so i can't unmute

Eric - machine readable reports allow easy comparisons

Peter - from my past in software testing - coverage / path testing - could be a large and time consuming thing

<shadi> AG: macros can record the steps taken

<shadi> ...where clicked and where that lead you to

Alistair - macros could be recorded to go along with use cases

Samuel - how can we continue to provide machine readable reports without being bond to a product

Samuel - there are products which could do this but it might be tricky

Eric - Remembered that machine readable report is optional

Eric - This is just looking at a way which would easily create a machine readable report

Samuel - Wondering if someone from Evaluation and Report Group could comment

Peter - good to hear that this is an optional area

Peter - as things change quickly

Peter - recording macros could be brittle, and not 100% accurate

Peter - we should ensure that this section is optional

<Aurelien_levy> one sort of marco command is : http://seleniumhq.org/docs/

Eric - it would still be good to look at how to do it


Shadi - chairs EARL working group

<ssirois> i can take the task of inviting someone from the ERT WG on our next meeting or i can discuss a little with them (since i'm registered to there mailing list) in order to understand how EARL 1.0 Schema could help us. (see http://www.w3.org/TR/EARL10-Schema/)

Shadi - EARL supports a number of ways to record page e.g. xpath, xpointer, exchange between server and client

Shadi - Peter was saying that it was not 100% accurate

<shadi> Step 4.d: Archive Web Pages for Reference (Optional)

Shadi - Another important point was archive web pages

Shadi - in the context of an application which show personal information could be tricky

Shadi - we might not want to record such information

Shadi - we would just have to outline the situation

Vivienne - Archiving pages - Needed to record websites - Can you save dynamic pages into blackbox

Peter - it depends upon the page and site

Peter - trying to record pages which rely on local code presents a problem

Eric - can we think of a standardised format

Shadi - you can record the DOM at best

Shadi - DOM plus steps to get to the DOM

Shadi - edge cases

Eric - can we have input from the EARL working group

Shadi - is that 5f optional machine readable reports

Shadi - what do we specifically want to record

Alistair - How does google cache pages?

Eric - will add to 5f - trying to put in information for people who would want to provide machine readable reports

Alistair - how many applications can read machine readable reports?

<vivienne> I can't understand Peter clearly. Is it the line?

Peter - we are using technologies which do not have a DOM

Peter - so we must remember this

Eric - Evaluator and commissioner can be asked for supporting material such as use cases

Eric - Check text to see if this is reflected

Eric - 5a bullet point 3 added - make clear what pages are part of a complete process

Eric - why, pages in a complete process must be checked in combination

Eric - Detlev is this OK

Detlev - All pages in a complete process must be part of a sample - do we need to do this in all cases. We need an exception rule

Detlev - 70 pages which contain similar content which you step through page by page

Eric - other important changes

Eric - definition of template added

Eric - first concept version of a definition

Eric - like or hate it?

Detlev - Comment about techniques from the email list

Detlev - 4B

Detlev - There is a great deal of talk relating to this point

<Aurelien_levy> +1 for Detlev

Eric - We should look at this next week

What is the command to make minutes?

Peter - can we have multiple templates on a single page - does it have to be HTML based?

<Detlev> I think it is "multiple samples on the same page"!


Shadi - how to make minutes?

Shadi - next week is a public holiday

<vivienne> not in Australia either?

Eric - we shall miss the meeting next week

<vivienne> okay, 24 May is good

Eric - 24th May next meeting

Eric - documents will come next week

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/05/16 13:29:24 $