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

This is also probably a good time to mention that if the working group documents a compatibility table, every single problem listed should point to a reported bug in each platforms respective bug tracker. Blanket statements like "Feature X doesn't work in Product Y" can be misleading if not associated with a version and bug link. I repeatedly run across hearsay or outdated compatibility reports that do not list a version or link to a bug report, even when the problem has been working for years. The link version number and to the bug report at least keep the report accurate.

Try this format: Expectation # of Feature X does not work as expected in Product Y version # build ###, on Platform Z version #. Reproduction steps and other details of Product X bug ### have been reported at URL: http://example.com/bugs/###

To all involved, please don't attempt to document a compatibility table unless you're willing to commit to the rigorous detail of specific versions and quality bug reporting with each listing of a compatibility problem. Anything less is potentially misleading and not worthy of the working group's time.

On Apr 7, 2014, at 6:11 AM, Birkir Gunnarsson <birkir.gunnarsson@deque.com> wrote:

> The format of the table looks decent (I am not sure why, but screen reader
> support for the colspan scope in tables is decent (minus Voiceover
> unfortunately),

As mentioned above, blanket statements like ~"scope doesn't work in VoiceOver" are not particularly helpful and potentially misleading. Does this mean scope with a value of col or colspan does not work? Does this mean scope applied on an element with a colspan does not work? Include a reduced test case to avoid ambiguity. What version doesn't it work in? What does "work" mean in this case: in other words, what did it do that you didn't expect it to do, or what didn't it do that you expected it should do? 

VoiceOver works reasonably well in that table despite the lack of *any* scope attributes, so your statement is vague or possibly incorrect. If there is a problem with some aspect of the scope attribute that does not behave as you'd expect, please file a bug with the browser and OS version numbers, a test case, reproduction steps, and expectation of what should have happened. Please make sure you test in a WebKit nightly. If you find an example where it is not supported or "works" insufficiently, please start with a new WebKit AX bug at http://webkit.org/new-ax-bug

If you document any bug in any location other than the specific product bug tracker (such as this planned compatibility table), please include a link to the bug from that location.

> but no screen reader supports the rowspan scope, but that is
> a topic for a different conversation).

The table mentioned does not use TH cells on the row or scope attributes. The left hand cells are plain TD elements, despite having a different background color.

> I see the need for this kind of up-to-date information every day in my job 

The intention to keep information like this up-to-date is worthwhile, but perhaps unrealistic. For this reason, it's critical to include version numbers and bug links. Even if the table does remain up-to-date, these details will help the table maintainers such as yourself keep track of current issues and know when to test again, and will ensure that the data remains accurate, even when subsequent versions of the software are released. Even if the problem is fixed in version #n.4, that does not negate the documentation that it did not work in version #n.3. If you didn't include the specific version number, your table would be out-of-date immediately.

> at Deque, and I would be very happy to contribute to this work.
> So you guys can definitely count on me for contributons of examles and
> testing.
> I am new to the group, so I do not know the processes of getting things
> started yet. *grin* but I have seat on my pants, so I can learn to figure it
> out.

Welcome to the working group, thanks for volunteering, and hold on to your pants seat. ;-)

Cheers,
James

Received on Monday, 7 April 2014 22:26:26 UTC