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

Re: Rethinking HTML 5 (Was: Re: Semicolon after entities)

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Wed, 02 May 2007 03:26:43 +1000
Message-ID: <46377853.1020608@lachy.id.au>
To: tina@greytower.net
CC: www-html@w3.org, public-html@w3.org

Tina Holmboe wrote:
> On  1 May, Lachlan Hunt wrote:
> 
>>> The spec authors looked at what browsers supported, documented it,
>>> and called it a standard.
>> Could you please explain how you think the process should work?
> 
>     "This group will maintain and produce incremental revisions to
>      the HTML specification."
> 
>   I'm sure you recognise the quote.
> 
>   So lets do just that. Start with the currently HTML 4.01 Strict, and
>   /revise/ it...

Starting over with HTML4 would effectively discard 3 years of good, 
solid work that has gone into HTML5, and take us right back to where we 
were in 1999.  That is not acceptable.

>>>   In other words the WG is to study, and document, both good and bad
>>>   practices of browsers and authors today, and tailor the specification
>>>   to allow them?
>> The specification will be tailored to *support* the bad things, which 
>> browsers are required to do regardless of what the spec says.  However, 
>> the spec will not necessarily allow such things in conforming documents.
> 
> The specification will /support/ bad ideas, but not /allow/ bad ideas?

I think you misinterpreted my use of the word "support".  I meant as in 
how to handle it, not approve of its use.  Let me try to rephrase:

   The specification's user agent requirements will define precisely
   how browsers need to handle markup, whether it's good or bad, but
   the authoring requirements will not allow the use of bad markup
   in conforming documents.

This seems to be the source of contention in the current debate.  For 
the spec to be implementable, it needs to define conformance 
requirements for UAs, including error handling and how to handle both 
existing and future content.

It also needs to define authoring requirements.  This is where you 
should be focussing your attention, as they are the requirements that 
define what markup is considered good and can be used, but is just a 
subset of what browsers are required to support.

However, it appears that you, and several others, are objecting to the 
inclusion of user agent conformance requirements, using arguments based 
around authoring requirements.  We need to be careful not to conflate 
these issues when discussing them.

Personally, I agree with you in principle that HTML5 should not 
encourage the further proliferation of purely presentational markup. 
The separation of semantics and presentation is certainly a useful goal, 
and is something we should strive for where practical, though I don't 
take such a hard line stance on it as you do.

We can get the authoring requirements right this time around, but 
defining authoring requirements that disallow the use of erroneous 
markup, does not remove the requirements for user agents to support it. 
  We must not let our ideals about developing a good markup language for 
authors get in the way of UA conformance criteria.

> The /specification of a markup language/ should /not/ "support" bad 
> practice. /Browsers/ should support bad practice.

We clearly agree that browsers need to support existing content, but the 
problem is that there is no spec that defines how.  Without such a spec, 
browsers are forced to guess and reverse engineer each other.  That 
approach clearly has not worked.  Browser are not interoperable, yet 
that is precisely what we need to achieve.

-- 
Lachlan Hunt
http://lachy.id.au/
Received on Tuesday, 1 May 2007 17:26:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:15:58 GMT