Re: HTML5 design philosophy disconnect

Hi Benjamin,

>- the implementation status note for WebKit explains that WebKit trunk
>does not pass all the relevant test cases.

this is of little use to a developer as there is no indication of what
it does pass and what it does not and what the relevant test cases
are, and in fact all the browsers have the same note.

I would be surprised if the accessibility API mappings are considered
relevant test cases for HTML5 as these are for the most part not
specified in HTML5.

> I'd hope to see similar such notes in the accessibility API mapping
> specification, again based on test cases.

Yes i agree and we will work on that aspect, but will endeavour to
make the information more meaningful for developers.

The approach I have taken in the alt spec is to provide information to
developers about the advantages and disadvantages of various features,
for example:
http://dev.w3.org/html5/alt-techniques/#altmethod

regards
stevef

On 6 January 2012 12:00, Benjamin Hawkes-Lewis
<bhawkeslewis@googlemail.com> wrote:
> On Fri, Jan 6, 2012 at 11:18 AM, Steve Faulkner
> <faulkner.steve@gmail.com> wrote:
>> When will specifiers give users and developers the same level of
>> consideration in the provision of useful practical information as is
>> afforded implementors?
>
> Assuming that implementors are not unwilling to implement a given
> feature, its status can reflected directly in the spec based on test
> cases. For example at -
>
> http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#the-label-element
>
> - the implementation status note for WebKit explains that WebKit trunk
> does not pass all the relevant test cases.
>
> I'd hope to see similar such notes in the accessibility API mapping
> specification, again based on test cases.
>
> https://dvcs.w3.org/hg/html/file/c807d2d4da84/tests/ is where the HTML
> test suite is currently living by the way.
>
> In summary, accurate information about implementation status can (and
> I think should) be surfaced in specifications but depends on test
> suites.
>
> --
> Benjamin Hawkes-Lewis

Received on Friday, 6 January 2012 16:04:04 UTC