W3C home > Mailing lists > Public > public-html@w3.org > November 2008

Re: Comments on HTML WG face to face meetings in France Oct 08

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Tue, 18 Nov 2008 09:28:40 -0500
Message-ID: <4922D118.8060703@mit.edu>
To: elharo@metalab.unc.edu
CC: public-html <public-html@w3.org>, www-tag@w3.org

Elliotte Harold wrote:
> Actually there's a lot more inconsistency than that. The XML 
> specification doesn't require anything in particular, even if the close 
> tag were present. The assumption that DOM is what you want or will get 
> is a big leap.

Sure.  I'm not saying a DOM is required.  The same information can be 
expressed in other ways (in the case of XML, SAX callbacks are one way 
as you pointed out).

XML does define what callbacks happen up to the point where the error is 
detected.  It does not define what should happen with the data after 
that point.  We seem to agree on this, right?

In the case of HTML5, all that's happening is that there is error 
recovery, so the sequence of callbacks after an error is detected is 
still well-defined.  No more, no less.

> Beyond error handling, a big concern I have with HTML 5 as written is 
> that it mixes together multiple layers that should be separate.

That sounds like a theoretical purity concern.  While I'm sympathetic to 
those, see the parallel discussion of the working group's charter as 
regards such concerns.

> The object model by which one accesses a document is very different from the 
> syntax of the document, the semantics of the document, and the display 
> of the document in a particular medium. HTML 5 treats these as more or 
> less  different views of the same thing rather than as separable, 
> independent pieces that connect to but do not depend on each other.

Unfortunately, in HTML as it is written on the web the interconnections 
are very substantial and very deep.  It doesn't help that HTML4 was 
already in significant part a presentational language, heavily 
conflating semantics and display.

While I think it would be possible to specify the parts separately, the 
result would be very difficult to read and even more difficult to 
implement.  In my opinion, of course.  If someone wants to take a shot 
at it, no one is stopping them!

Note that the specification does address the case of HTML consumers who 
do not have a DOM or don't support scripting.

-Bori
Received on Tuesday, 18 November 2008 14:42:25 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:59 UTC