- From: Daniel Glazman <daniel.glazman@disruptive-innovations.com>
- Date: Thu, 17 Jun 2010 19:09:31 +0200
- To: Edward O'Connor <hober0@gmail.com>
- Cc: HTML WG <public-html@w3.org>, Steven Faulkner <faulkner.steve@gmail.com>
Le 17/06/10 18:45, Edward O'Connor a écrit : > Only some author conformance requirements are machine-checkable. This Then they cannot be enforced and are useless. They can even be harmful from a REC track point of view. > is one of the ones that isn't. Are you objecting to the presence of > any non-machine-checkable author conformance criteria? If so, please > see this comment on the relevant Bugzilla bug: > > http://www.w3.org/Bugs/Public/show_bug.cgi?id=7034#c35 Yes I am. Let me give you such an author-related criterium from the HTML4 spec: authors should not use the class attribute to simulate their own markup. Fine. "should not" is fine here. But it's not a conformance criterium because it cannot be one. The "non-conforming" statement previously mentioned on table layout is exactly of the same kind: it should be a strong authoring recommendation but it cannot be a criterium of conformance ; because it's not machine- checkable, nobody will care and editing tools will be unable to deal with it. There will also be edge cases where one says it's non-conforming and others say it is. Chaos. </Daniel>
Received on Thursday, 17 June 2010 17:10:07 UTC