- From: Jason White <jasonw@ariel.its.unimelb.edu.au>
- Date: Mon, 30 Jul 2007 12:53:44 +1000
- To: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Cc: public-html@w3.org, wai-xtech@w3.org
On Sun, Jul 29, 2007 at 12:22:06PM -0400, Al Gilman wrote: > I think we've made some real progress in this thread. Yes. We're starting to move into the details, which is often a positive development. > > Let me just try to hit two points: one candidate agreements that we might > manage to discuss less; and one perceived point of difference that we might > manage to discuss more concretely: Your first point aptly summarizes an important design principle. Moving to your second point, > 2. [possible disagreement] One of the major reasons for creating HTML5 > is to clean up the mess of markup in use. > > Jason said something like this. I fear that if we took Jason's > paragraph, or post, and put it up to an opinion poll in HTML WG there > would be a majority of 'disagree' and 'strongly disagree' responses > (on a five point scale including those two, 'don't care', 'agree', > and 'strongly agree'). Supposing for the sake of discussion that this is correct, a different result might be reached if the kinds of clean-up envisaged were more precisely delineated. There are two important respects in which the current draft already leans toward greater cleanliness: 1. It specifies precisely the syntax of a conformant HTML document, with a view to making it easier to construct interoperable HTML processors. This is part of what I meant by arguing that the barrier to entry needs to be lowered so as to reduce the effort required to implement an HTML application. I expect this will lead to benefits across the entire range of HTML processors, including applications that transform and present HTML in ways that benefit particular groups of users (including users with specific types of disability). 2. As stated in section 3.3.1 of the current draft, authors must only use elements, attributes and attribute values for the appropriate semantic purposes. It is stated elsewhere that the conformance requirements applicable to wysiwyg editors and format converters are more lenient, but even so, any move in the direction of greater semantic adherence to the HTML specification will improve interoperability and enable user agents and other HTML processors to take the HTML document at face value, resorting less to unreliable means of attempting to derive semantics from erroneous markup. Again, this simplifies the application of styles and other transformations by which the document can be adapted to diverse delivery contexts. Thus, a clean-up is already within the scope of the HTML 5 effort. What is more likely to engender controversy, however, would be an attempt to regularize and systematize HTML through wholesale replacement of features inherited from HTML 4 and from existing user agents. Indeed, I think the design principles, as currently drafted, effectively rule this out, unless motivated by problems affecting users, authors or implementors, the solution of which entails backward-incompatible changes.
Received on Monday, 30 July 2007 02:54:04 UTC