- From: Maciej Stachowiak <mjs@apple.com>
- Date: Tue, 17 Mar 2009 16:02:20 -0700
- To: "T.V Raman" <raman@google.com>
- Cc: jonas@sicking.cc, robin@berjon.com, forums@david-woolley.me.uk, www-svg@w3.org, public-html@w3.org
On Mar 17, 2009, at 9:08 AM, T.V Raman wrote: > Correct. > > I believe in its final incarnation, the HTML5 work needs to > define the following: > > A) The collection of algorithmic hacks that it takes to go from > the garbage found on the legacy Web to a meaningful DOM. HTML5 does define this, as part of the parsing algorithm. > B) Define *everything* else -- in terms of that canonical DOM, > including clean serialization. HTML5 does this as well - once the parsing algorithm has produced a DOM, everything else is defined in term of the DOM. Note however that it does not prevent the DOM from having validity errors. Both the parser and DOM APIs can produce DOMs that violate content model requirements. But the DOM abstracts out lower-level syntax errors or syntax quirks. I think defining the parsing rules and serialization rules are properly in scope. What is not in scope is mandating when and how the browser UI should expose these things. Regards, Maciej
Received on Tuesday, 17 March 2009 23:03:12 UTC