[whatwg] Allow trailing slash in always-empty HTML5 elements?

[Ian Hickson]

On Wed, 29 Nov 2006, Steve Runyon wrote:
>
> Sorry for being the dunce here, but is anybody saying otherwise?  
> Whereas XML _requires_ that you close every tag, HTML5 _should allow_ 
> you to close any tag.  I agree with what was said previously about 
> considering something like '<select /></select>' invalid, but if 
> somebody's suggesting that something like '<img src="..." />' or '<br 
> />' should also be invalid, I disagree.  Validators and UAs should 
> accept singleton tags _with or without_ the self-closer.
> 
> Am I totally misunderstanding or missing the point here?

The argument is that the self-closer "/" is an XMLism, and that HTML5 has 
nothing to do with XML, so there's no reason for it to apply here.

Note that in HTML, this:

   <p/> test 

...regardless of what this discussion results in, will always be treated 
exactly the same as:

   <p> test </p>

...because, for legacy reasons, there's no way we can treat "/" as a 
self-closer in any tag other than void tags (like <img> or <br>).

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