Re: Spec parser

On Jun 10, 2013, at 17:04, James Graham <jgraham@opera.com> wrote:

> On 06/07/2013 05:52 PM, Tobie Langel wrote:
>> On Friday, June 7, 2013 at 9:06 AM, Takashi Hayakawa wrote:
>>> Hello Peter and all,
>>> 
>>> I started looking into the media element part of HTML-5 as a member of
>>> the small team at CableLabs that Bob Lund mentioned a couple of weeks
>>> ago.
>> 
>> You might find the following link useful: https://www.w3.org/2013/04/test-coverage/index.html?algos=2&assume-idl=100&assume-tooling=60&idl=4&level=3&propdef=8&reftest-factor=2&review-cost=30&review-success=50&rfc2119=4&show-details&spec=html&test-cost=60
> 
> Just so it is clear that the above method is very rough (but without suggesting that anyone has claimed otherwise), it is instructive to look at the syntax section. Based on the default settings, the coverage table claims that syntax/writing needs over 400 tests, but actually the only conformance criteria in that section are for users. On the other hand it claims that syntax/parsing is overtested by a factor of about 10. However experience shows that, although interop is dramatically better than it was a few years ago, we are still finding significant bugs and incompatibilities in implementations.

Yup. Any suggestions on how to improve that?

--tobie

Received on Monday, 10 June 2013 10:39:32 UTC