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

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

From: Geoffrey Sneddon <foolistbar@googlemail.com>
Date: Tue, 1 May 2007 16:57:17 +0100
Message-Id: <1F173715-2CAE-4769-8B65-9838F59880F2@googlemail.com>
Cc: public-html@w3.org
To: "Philip Taylor (Webmaster)" <P.Taylor@Rhul.Ac.Uk>

On 1 May 2007, at 16:49, Philip Taylor (Webmaster) wrote:

> Geoffrey Sneddon wrote:
>> How will the browsers know how to support bad practice? The  
>> purpose of putting it in the spec is so that all browsers support  
>> invalid HTML consistently.
> /That/ has little or nothing to do with defining HTML5, which
> is what I believe to be the remit of this group.  By all means
> let there be a working group to define how browsers should behave
> when faced with invalid HTML 4.01- (a.k.a. tag soup), but do not
> let their work interfere with the task in hand, which is to define
> what the next iteration of HTML should be, and how browsers should
> behave when presented with a /valid/ instance of it (and, perhaps,
> how to fail gracefully when presented with something that claims to
> be [valid] HTML5 but is not).

So you are proposing error handling remains undefined, despite the  
fact we need to define parsing anyway? If we're going to define  
parsing, we may as well define parsing for all cases, both valid and  
invalid documents. As someone implementing HTML, I don't want to go  
go around reverse engineering other browsers. If we include tag soup  
in "classic HTML" (as it's unclear what "classic" means in that  
context), we are chartered to define parsing requirements.

- Geoffrey Sneddon
Received on Tuesday, 1 May 2007 15:58:02 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:20 UTC