- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Fri, 30 Nov 2007 07:37:43 +0200
- To: T.V Raman <raman@google.com>
- Cc: fielding@gbiv.com, annevk@opera.com, public-html@w3.org
On Nov 30, 2007, at 03:50, T.V Raman wrote: > What's more, two implementors who did that would almost certainly > come up with different answers. html5lib and the Validator.nu HTML parser are independent non-browser implementations. As far as I can tell, both agree on what the document tree should look like even when the document tree is not a DOM. As an implementor of a parser that can run without a DOM, I'm not complaining about the spec being written in terms of the DOM. > I think this thread is closely related to the question Henry > Thompson asked at the Tech Plenary which was: > "HTML5 says how all the incorrect cases are handled, but fails to > tell authors at the end of the day what the correct way of doing > things is" I answered that question at the Tech Plenary. There is a "Writing" section that describes what kinds of character sequences one may write without hitting the error conditions in the parsing algorithm. -- Henri Sivonen hsivonen@iki.fi http://hsivonen.iki.fi/
Received on Friday, 30 November 2007 05:37:58 UTC