Re: Extension conflict/compatibility requirement

Group,
We have a new version of the extension framework/requirements document with edits made as a result of discussion on last week’s call.

The main changes can be seen in the diff version:
https://www.w3.org/WAI/GL/wiki/index.php?title=WCAG_Extensions_Framework&diff=5872&oldid=5867


Or you can view the new version only:
https://www.w3.org/WAI/GL/wiki/WCAG_Extensions_Framework


We will discuss this on the call today, but also on the list, so please take some time to review this and share comments as we would like to provide this for public review soon.

Thanks,
AWK

Andrew Kirkpatrick
Group Product Manager, Accessibility
Adobe

akirkpat@adobe.com
http://twitter.com/awkawk

http://blogs.adobe.com/accessibility



On Fri, Oct 16, 2015 at 5:28 AM, Joshue O Connor <josh@interaccess.ie<mailto:josh@interaccess.ie>> wrote:
Hi all,

On the working group call this week there were a couple of interesting points raised regarding extensions that require further discussion. We also wish to engage other people on the list who were not on the call, and make sure that they are aware of some of the finer points and able to express an opinion here on the list.

To sum up, two main 'themes' in our extension framework are extension compatibility, and the need to reduce, minimise or indeed remove any conflict between extensions.

NOTE: As a thought experiment, one possible way to do that would be to have a 'MonoSpec' extension which combined the output from all TFs (Mobile/Cognitive/Low Vision) in a single spec. Potentially where care is taken to ensure that these extension SCs are fully compatible with each other there may be less 'conflict'.

The 'PolySpec' extension approach would involve taking the SCs from each group and placing them in separate docs that conformance claims would be written against individually.

While in principle, the contents of these docs would be more or less the same, the potential for conflict if there is only a 'MonoSpec' may be reduced. If only because a valid conformance claim would need to be written against it in toto. Also this approach would mean that devs would have to satisfy the success criteria in the MonoSpec fully, even if some are outside of the developers immediate area of interest. So in short could be a good way of conditioning developers to consider other user needs - rather than thinking "I need to make my content conform to just mobile, or low vision success criteria etc".

Regarding extension conflict, in our current draft 'WCAG Extensions Framework' document it states: [2]

"Ensure that all WCAG extensions are compatible with each other
Extensions must not conflict with each other. This is important for the purpose of enabling content providers to implement support for more than one extension. For this reason will be critically important for group members working on different extensions to maintain good communication about extension work in progress."

There are a couple of questions/points that arise:

1) Should we explicitly call out the need within the framework that there must NOT be conflict between extensions? It has been pointed out (rather practically) that it just may not be possible to avoid conflict with our extensions.

2) If we do explicitly call out this issue in our framework, it may help focus working group attention on carefully finding where there are conflicts in extensions (between there own group and others).

3) On a more granular level how do you think the framework should even define conflict?

4) Obviously while spec fragmentation is a concern inherent in the extensions discussion a final thought is the basic question; Is conflict always inherently bad? Can positive conflict or friction between various user requirements result in the end in better content, better user experience etc?

What do you think?

[1] http://www.w3.org/2015/10/13-wai-wcag-minutes.html

[2] https://www.w3.org/WAI/GL/wiki/WCAG_Extensions_Framework