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

Re: Semicolon after entities

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Tue, 01 May 2007 23:09:31 +1000
Message-ID: <46373C0B.8080803@lachy.id.au>
To: David Woolley <forums@david-woolley.me.uk>
CC: www-html@w3.org

David Woolley wrote:
> Lachlan Hunt wrote:
>> Tina Holmboe wrote:
>>> Can you enumerate the features that /cannot/ be implemented?
>>
>> There are many, but ok.
>
> Most of the list /can/ be implemented.  I believe what Tina was asking 
> was which features of HTML cannot be implemented because they are either 
> theoretically impossible (self contradictory, halting problem, etc.) or 
> because they would have performance problems, or an inordinately high 
> implementation cost.

No, Tina just asked for a list of features that cannot be implemented, 
regardless of the reason.  However, my list included features that are 
poorly defined, difficult to implement and even cited the self 
contradictory conformance criteria.

> I'm sure she knows that the majority of web sites would break if used 
> with a user agent that fully validated against the specification.
> 
> (I suspect that she was not putting a high priority on the features that 
> are only implied by the formal SGML definition, rather than the narrative.)

You can't claim that a spec can be implemented, and then selectively 
exclude sections that prove otherwise.  Without SGML, HTML4 *does not* 
define any parsing, syntax or error handling requirements, which makes 
it impossible to implement from the start.

Now, consider that there is little practical difference between a spec 
that doesn't define those things, and one that does by referencing a 
spec that cannot be implemented in the real world anyway.  The fact is 
that HTML4 *cannot* be implemented as defined, period.

-- 
Lachlan Hunt
http://lachy.id.au/
Received on Tuesday, 1 May 2007 13:10:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:09 GMT