Re: Void elements in HTML (Was: ZIP-based packages and URI references into them ODF proposal)

On Wed, 31 Dec 2008, Julian Reschke wrote:
> Ian Hickson wrote:
> > ...
> > Right, but we disagree on this, as noted earlier in this thread. I have to
> > balance all the various needs here. Some of these needs are better satisfied
> > by introducing lots of new void elements frequently; other needs 
> 
> Which particular need requires introducing void elements?

<command> would be basically unusable if you had to include an end tag 
each time.


> > > Could you please explain (or point to an explanation) why exactly it 
> > > needs to be known beforehand whether something is block-level or 
> > > not? It's certainly not needed for serializing a DOM, so I assume 
> > > the parser needs to know for some reason?
> > 
> > Block-level elements have to be known by the parser so that optional 
> > </p> end tags can be implied.
> 
> Just make the </p> end tags required in this case. Problem solved.

This solves the problem for people who want to extend the language to add 
new block-level elements, and instead introduces a problem for authors who 
want to use the language (the language becomes inconsistent). There are 
orders of magnitude more users of the language than extenders or 
implementors of the language, so on the balance it's not a good solution.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 31 December 2008 12:25:01 UTC