- From: Robin Berjon <robin@berjon.com>
- Date: Wed, 29 Feb 2012 20:20:21 +0100
- To: Arthur Barstow <art.barstow@nokia.com>
- Cc: public-coremob@w3.org
On Feb 29, 2012, at 19:06 , Arthur Barstow wrote:
> Well I hope we're all on the same page here. So given the following in the spec text:
>
> [[
> All normative content will be specified exclusively by reference to the original standard defining the feature. Additional non-normative implementation guidance may be included.
> ]]
>
> Is the expectation:
>
> * The CG {will | expects to} create tests for its own "non-normative guidance" statements
>
> * If the CG identifies any testing related issues e.g. gaps, bugs, etc. for "normative references" the CG {will | expects to} contribute tests, patches, etc. to the relevant WG
I think the answer to that is "Yes". I reckon that for tests that are clearly gaps in other specs' suites the logical thing is for them to be contributed. For non-normative guidance, the problem is that these things may be hard to test. For instance, a typical example of non-normative guidance could be "Usage of the Canvas2D API SHOULD be hardware accelerated". If we did succeed in building a test for that, it would still not be clear that it would find a comfortable home in existing suites since it's more a matter of implementation quality than conformance.
--
Robin Berjon - http://berjon.com/ - @robinberjon
Coming up soon: I'm teaching a W3C online course on Mobile Web Apps
http://www.w3devcampus.com/writing-great-web-applications-for-mobile/
Received on Wednesday, 29 February 2012 19:20:47 UTC