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

Re: HTML version issue summary?

From: Jeff Schiller <codedread@gmail.com>
Date: Wed, 25 Apr 2007 07:51:39 -0500
Message-ID: <da131fde0704250551i655e6935x7f5f9380500ccb8a@mail.gmail.com>
To: "David Hyatt" <hyatt@apple.com>
Cc: "Dan Connolly" <connolly@w3.org>, public-html@w3.org

On 4/25/07, David Hyatt <hyatt@apple.com> wrote:
> On Apr 24, 2007, at 9:01 PM, Jeff Schiller wrote:
>
> > On 4/24/07, David Hyatt <hyatt@apple.com> wrote:
> >>
> >> (4) I think IE's opt-in should be independent of DOCTYPE until such
> >> time as they are confident that they have HTML5.0 fully implemented
> >> and supported.  Then one could imagine the doctype being used as the
> >> opt-in.
> >
> > I'm trying to understand this suggestion, but still don't get it.  I
> > think I have the same concern as Maciej in that every HTML5 document
> > that comes out would have to include the IE-specific opt-in mechanism
> > when a "perfectly good" opt-in solution (whatever the proposed HTML5
> > DOCTYPE or version attribute) would exist for Microsoft to use right
> > out of the gate.
> >
>
> My assumption is that IE would only have to use such an opt-in (a
> custom one) if their implementation was really far off, e.g., only
> 20% done, or full of enough experimental features/dangerous bugs that
> they'd be reluctant to brand it as official.  If what they produce is
> great and close enough to the standard, then I'd expect they could
> use the doctype switch right out of the gate instead.
>
> I guess what I'm getting at is that as far as the HTML WG is
> concerned, all we have to do is define a way of identifying HTML5
> (either a doctype or attribute) and then state that a fully
> conforming HTML5 browser uses this version switch as a definitive
> trigger for "HTML5 mode."  The custom opt-in idea is not something
> that people should really latch on to here, since that is really
> completely out of our hands.
>
> I only suggested it as a failsafe option if MSFT is worried that
> people would rush out and adopt a possibly broken implementation if
> they release something too early.

David,

Your suggestion makes sense now, thank you.  I would prefer to give
Microsoft the benefit of the doubt here, but as you say, if their
release is "too early" then they would have the option of using a
vendor-specific opt-in.

I think the point raised about HTML fragments is a good one that I
hadn't thought about before and in which the DOCTYPE w/version does
not solve.  I suspect maybe that it won't matter as much, but I worry
(like David) that I'm being naive.  In that light, what about this:

- HTML5 spec suggests to use <!DOCTYPE html> (essential to trigger
standards mode, and really nice and easy to remember to type), but
this is only a recommendation
- the HTML5 spec states that a version="5.0" attribute must be present
on the html element.  This is the web author way of telling user
agents that the content is HTML5 (not HTML 4.01).  This version
attribute supersedes any version identifier present in the DOCTYPE.

At some point in time, it may become clear that the HTML5 spec truly
does define a fully backward compatible specification that does not
break the existing web in any major way on any major browser.  I've
stated that I have my doubts on this one, but I'm willing to consider
that it is a possibility.

On this day, user agents may decide that the default rendering mode is
HTML5 even if the version attribute is not present and an older
DOCTYPE is used.  On that day, all existing web content automatically
becomes HTML 5.0 content for those browsers and the existing web
content is indeed "versionless" in that respect.  It sounds like the
likelihood of Microsoft doing this is slim in the short term because
there would still be SOME risk of breaking content out there and they
got burnt from IE7.  This is their prerogative, of course, but later
experiments, changes in market share, etc, may convince them
otherwise.

But if that day comes, where all major browsers are ignoring the
version attribute and de-facto treating all web content as HTML5, then
the final version of the HTML 5.0 spec could remove the version
attribute requirement altogether - thereby achieving the goal that Ian
Hickson outlined:  Crafting a spec that concisely outlines how a user
agent should render all existing web content.
Received on Wednesday, 25 April 2007 12:51:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:15:53 GMT