RE: Is there an official place to document current role compatibility differences?

Where can the ARIA test harness be accessed? This sounds like a good
resource for conducting many of these tests.

The standard user DB sounds good, but the problem with general user data is
that it can often be sprinkled with user error data, such as the user simply
not liking the intended interaction design of a widget, or the way that
feedback is conveyed. I had that recently where a user said that ARIA Tabs
were wrong, because he couldn't access them in the Links list.

I always report bugs when I encounter something that I can prove is a bug,
or reasonably so until proven otherwise, so no worries there.

My main interest is to build out a single page high level overview resource
that can be edited as needed to give a good overview of where the breakdown
between browser support vs AT support differs, because then it will be
easier to figure out where the main issues are located. I think this still
will take a bit of playing with to get organized where version data can be
included and variation notes can be added, since one way of doing things
often varies greatly between another even if both ways are technically
compliant.

Apologies if I haven't replied earlier to others, my main comp died and Win
is being reinstalled, so I'm using a backup for the time being.


-----Original Message-----
From: James Craig [mailto:jcraig@apple.com] 
Sent: Tuesday, April 08, 2014 11:40 AM
To: Foliot, John
Cc: Birkir Gunnarsson; Bryan Garaventa; Cynthia Shelly; public-pfwg@w3.org
Subject: Re: Is there an official place to document current role
compatibility differences?

On Apr 8, 2014, at 11:25 AM, Foliot, John <john.foliot@chase.com> wrote:

> I think we are talking past each other. There won't be "a list" - there
will be a data-store of test results of ARIA attributes (and combinations
when required) tested against [Operating system + versions] + [Browser +
versions] + [Assistive Technology + versions]. It will all be captured in a
single database - the end.

It sounds a lot like the ARIA test harness. Why are you duplicating this
effort instead of contributing to the existing one? We've got 700+ test
cases already.

> No, asking the crowd to be that granular is a recipe for low
participation.

At a minimum then, the tool should not prevent them from entering this
critical information.

> Once the data store starts growing, it would be relatively easy for an
engineer at any company to run a query to ferret out bugs in their current
builds

Browser and AT manufacturers deal in bug numbers of thousands or tens of
thousands with every release and have work flows revolving around the
official bug databases for those products. It's naïve to assume an engineer
would manually compare and resolve the results of one database query against
another. If your database included a place for a bug number, plus other
useful information such as "last tested version", this would not have to be
manual.

Received on Tuesday, 8 April 2014 19:45:28 UTC