W3C home > Mailing lists > Public > public-html@w3.org > November 2007

Re: Request for clarification on HTML 5 publication status (ISSUE-19)

From: Ian Hickson <ian@hixie.ch>
Date: Fri, 30 Nov 2007 00:27:41 +0000 (UTC)
To: "Roy T. Fielding" <fielding@gbiv.com>
Cc: Maciej Stachowiak <mjs@apple.com>, Dan Connolly <connolly@w3.org>, "public-html@w3.org WG" <public-html@w3.org>
Message-ID: <Pine.LNX.4.62.0711300018520.3737@hixie.dreamhostps.com>

On Thu, 29 Nov 2007, Roy T. Fielding wrote:
>
> HTML is a declarative data format for the interoperable exchange of 
> hypertext.

HTML was a declarative data format for the interoperable exchange of 
hypertext. However, it grew beyond that some years ago. Now it's also an 
application development and deployment framework.

Going forward, I don't think it is productive for us to really split our 
efforts into making two specifications, one for browsers and one for 
everything else. I think we would find that splitting the normative 
conformance criteria across multiple specifications would lead to more 
errors in the specification.

However, I have tried to structure the spec in a way that somewhat 
mitigates the problem you are describing. For example, most of the Web- 
browser-specific parts are shunted off to section 4 (unambiguously titled 
"Web browsers" so as to leave no doubt). Similarly, the syntax section 
clearly separates the syntax prose (author-aimed) from the parsing prose
(implementor-aimed). The DOM section is also kept separate from the 
semantics section.

On the long run, when the spec is more complete, I intend to go through 
the spec annotating what parts apply to whom, and making it possible to 
hide the parts you are not interested in, using alternative style sheets. 
However, I don't think we're really ready for that yet (it's a lot of 
editing work to maintain such annotations).

I understand if you don't believe this is satisfactory -- I similarly 
don't believe it is satisfactory for HTTP to _not_ address the needs of 
all its target conformance classes in the same specification (for example, 
HTTP doesn't precisely define how to handle multiple Content-Type headers, 
something which is causing interoperability problems in Web browsers and, 
presumably, any user agent that honours those headers on arbitrary Web 
content). I'm not sure that it will be possible to satisfy both of us at 
the same time, though.

If you have any suggestions on how we can address your desires while still 
addressing the other desires that have been put forward, I'm certainly 
open to suggestions.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Friday, 30 November 2007 00:27:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:10 GMT