- From: Joshue O Connor <joshue.oconnor@cfit.ie>
- Date: Fri, 18 Apr 2008 11:47:36 +0100
- To: Karl Dubost <karl@w3.org>
- Cc: James Graham <jg307@cam.ac.uk>, "Bonner, Matt (IPG)" <matt.bonner@hp.com>, Ian Hickson <ian@hixie.ch>, Anne van Kesteren <annevk@opera.com>, Steven Faulkner <faulkner.steve@gmail.com>, "public-html@w3.org" <public-html@w3.org>, "wai-xtech@w3.org" <wai-xtech@w3.org>, "wai-liaison@w3.org" <wai-liaison@w3.org>, Henri Sivonen <hsivonen@iki.fi>
The following is a post from Jason White to the HTML4all list that I think it a very interesting idea. Jason has given his permission for me to share the post with this group: <quote> Ideally, HTML validators would be like compilers, dividing errors into different classes of severity - minimally, "warning" and "fatal error". There would be a specification that describes what to treat as warnings, and what as validation errors. then <img src="http://example.org/mystery.png"> could be a warning, or at least we could have a nice debate about its severity level. Maybe the two options presently open, "fatal error" and "no error at all" simply don't capture the spectrum of possibilities adequately. It also appears to me that, if HTML 5 is committed to extensive user-agent error handling, then the case of a missing @alt is one for which it would be desirable to prescribe consistent fall-back behaviour. Once that is done, the discussion seems to become a question of what validators (including validating authoring tools) should accept as correct document instances. </quote> Josh
Received on Friday, 18 April 2008 10:50:40 UTC