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

Re: Scope and obsolescence of other specs (was Re: minor copy edits...)

From: Ian Hickson <ian@hixie.ch>
Date: Mon, 18 Jun 2007 23:25:28 +0000 (UTC)
To: Mike Brown <mike@skew.org>
Cc: public-html@w3.org
Message-ID: <Pine.LNX.4.64.0706182310480.8482@dhalsim.dreamhost.com>

On Mon, 18 Jun 2007, Mike Brown wrote:
> 
> This is tangled up in the fact that I can't accept, even informally, the 
> notion of HTML 5 "replacing" these other specs. Discouraging HTML 
> authors, in general, from producing HTML 4 or XHTML 1.1 documents is one 
> thing, but do you really mean to suggest that once HTML 5 is done, there 
> will never be a need to produce XHTML 1.1 documents or process documents 
> according to that spec's requirements for UAs?

Exactly. The intend is for the HTML5 spec's requirements to be a complete 
superset of the XHTML 1.x, HTML 2, HTML3.2, HTML4, DOM1 HTML, and DOM2 
HTML specs' requirements, and for the language to be a better version such 
that there should be no good reason to prefer an older language or API 
version over the one now defined.


> You say (in the Intro, but not in the Status) that it's not a normative 
> statement, but it's just too strong a statement to be informative. Again 
> this falls under the "pretend we didn't say This; we really meant That" 
> gesticulating that I find too sloppy and not in line with the way the 
> referenced specs have handled things.

I'm not sure exactly which statement you're referring to, but assuming you 
mean either of these:

 * This specification is intended to replace (be the new version of) what 
   was previously the HTML4, XHTML 1.x, and DOM2 HTML specifications.

 * This specification represents a new version of HTML4 and XHTML1, along 
   with a new version of the associated DOM2 HTML API.

...then I don't understand what it would mean for those statements to be 
normative. They aren't requirements. How would you test them?

Those statements are indeed meant to be true, there's no gesticulating 
intended. They are also intended to be informative (in the literal sense).


> The HTML 5 spec's scope is, as I see it, to provide requirements for 
> authors of "HTML 5" documents and implementers of UAs that process such 
> documents. Likewise, the XHTML 1.1 spec is focused on providing 
> requirements for authors of "XHTML 1.1" documents and implementers of 
> XHTML 1.1 processors. Same for HTML 4.x and DOM Level 2 HTML -- these 
> specs are authoritative for their own domains, and documents and UAs can 
> claim conformance as they see fit. None replaces the other; they all 
> coexist, fulfilling separate purposes. There is no single "HTML" 
> document type or UA which these specs compete for authority over.

The relationship between HTML5 and XHTML 1.x and HTML 4 is the same as the 
relationship between DOM2 HTML and DOM1 HTML, or CSS2 and CSS1. The more 
recent version of the spec completely supplants the earlier version. You 
would never need to write a CSS1 stylesheet, or use the DOM1 HTML 
specification when referring to the DOM HTML APIs. Similarly, there should 
be no need to refer to an earlier version of the HTML and XHTML specs when 
implementing or using HTML -- HTML5 covers everything required.

(At least that's the intent. If you find holes in the spec, let me know.)


> In XHTML 1.1's case, consider its Modularization of XHTML basis, its 
> "strictly conforming" document & UA requirements, and its inclusion of 
> ruby markup. None of these are features of HTML 5, perhaps 
> intentionally.

Ruby will be added to HTML5, its omission is temporary. I don't understand 
what you mean about the lack of ""strictly conforming" document & UA 
requirements" in HTML5, indeed i thought HTML5 was more strict and precise 
than XHTML 1.1 regarding conformance. The "Modularization of XHTML" basis 
is a specification-level concern and doesn't affect authors or 
implementors, so I don't really understand its relevance.


> Also, I believe XHTML 1.1 is currently undergoing 2nd Edition work, the 
> contributors to which would probably like to know is all in vain, from 
> your spec's perspective.

I am not aware of this work. Could you elaborate on this?


> > The intent is to cover everything that the earlier specs did, in far 
> > more detail. (Some of the features, most notably the presentational 
> > elements in HTML4, aren't yet in the spec, but they will be in due 
> > course.)
> 
> That's fine, but covering everything isn't the same thing as rendering 
> the previous specs obsolete.

Isn't it?


> The HTML 4.x specs were fairly delicate with the way they made reference 
> to HTML 3.2, and the XHTML specs have also avoided making any sweeping 
> declarations of ascension. HTML 5 should tread just as lightly.

I don't really see the advantage of subtlty in this matter. Our task here 
is to be very clear as to our intent, not to try and avoid hurting 
people's feelings.


> The spec should not imply that in order to be a conforming HTML 4.x or 
> XHTML 1.x document processor, a UA must be an HTML 5 processor, which is 
> what I feel would be the case if the older specs were truly replaced.

I don't think it implies that. It implies that being a conforming HTML 4.x 
or XHTM 1.x document processor is irrelevant, but replacing those specs 
doesn't stop those specs from existing. You can implement CSS1 without 
miplementing CSS2, but CSS2 obsoletes CSS1, even if it doesn't say it 
explicitly.

If HTML5 is never widely supported, then it can just be ignored, in which 
case the statement that it replaces those other specs becomes irrelevant. 
It's only if HTML5 is supported that the other specs become obsolete, at 
which point the statement is true and making it clear is, IMHO, desirable.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 18 June 2007 23:25:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:01 GMT