[whatwg] [wf2] 2.3. Changes to existing controls

[Ian Hickson]

On Sat, 29 Oct 2005, Anne van Kesteren wrote:
> Quoting Ian Hickson <ian at hixie.ch>:
> > > > > # A control is said to have no value selected if its value is the
> > > > > # empty string.
> > 
> > Apparently... but I don't see how it could have any effect. Nothing in 
> > the spec says anything about "no value selected" having any effect on 
> > radio buttons or check boxes, does it?
> 
> It says "A control". Perhaps it should say, "A form control, with the 
> exception of radio buttons, checkboxes and file controls, is said to 
> have no value selected if its value is the empty string."

I don't understaqnd why it matters. This sentence is just a definition, 
which is then used in normative criteria. None of those normative criteria 
apply to radio buttons. So why should the definition mention radio 
buttons? You would never get to that definition if you were reading the 
spec to find out what radio buttons should do.


> > It's a problem in HTML too (depends on the UA, typically, some are 
> > more compliant than others).
> 
> Right... You mean the one that is based on SGML?

No, there is considerable variation in how attributes are normalised 
across various common HTML implementations (sometimes it varies 
per-attribute, too).

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