RE: ACTION-95, ISSUE-65: Plan to publish a new WD of HTML-5

On Tue, 27 Jan 2009, Larry Masinter wrote:
> 
> The technical issue here, the serious technical issue, is whether it is 
> actually possible to independently define HTML as a markup *language* 
> with a semantics that are independent of the operational processing 
> rules defined in the current document.

This question is a non-sequitur here. The draft that Mike is proposing, as 
I've noted before, would not be such a document. Given its current 
audience statement, and assuming that statement to be accurate, the draft 
Mike is proposing could not be independent of the implementation 
conformance criteria.

If the audience statement is inaccurate, as you seem to be implying, then 
it should be fixed, and I do think we should do that before publishing the 
draft as a WD, since this point has caused so much controversy.


> Can that independent language be read, written, interpreted, without 
> making reference to somebody's really cool open source implementation of 
> a parser that everybody just has to use.

The answer to _that_, at least insofar as the parser is concerned, is 
obviously yes, as we've already done it in the HTML5 draft and had this 
verified with multiple independent implementations tested against the 
start of a pretty substantial test suite.


> Can it be mapped to other processing methods, compact XML, XML 
> databases, and other elements that the non-browser part of the web 
> community actually care about?

That too is already done in the HTML5 draft.


> [...] you actually need to try to write the language specification [...] 
> Mike's done a valiant job of trying to draft such a language 
> specification, and I think it would be a damned good idea to let the 
> community actually evaluate that.

Mike has *not* done this, at least not according to him. He has, according 
to what he has said, merely started writing the first part of a document 
with the same audience as the HTML5 draft. Such a draft would have to 
include all manner of implementation conformance criteria, DOM APIs, 
parser rules, etc.


> [...] you actually need to try to write the language specification and 
> evaluate it for compatibility both with the parsing rules as well as 
> with the browser specification [...]

Could you clarify what you mean by "the browser specification"?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 28 January 2009 01:05:05 UTC