[whatwg] [WF2] Readonly and default pseudoclass matching

[Boris Zbarsky]

Matthew Raymond wrote:
> Take a look at the following URL:
> 
>    http://www.w3.org/TR/2003/REC-xforms-20031014/sliceF.html#id2644859

That has no bearing on how :read-only is to be applied outside the context of 
the XForms namespace.  Like to HTML, say.  Which brings us back to WF2, which is 
working with HTML...

>    So, clearly, when :read-only was first introduced for XForms, it was
> meant to be used only with form controls that are not only set to
> read-only, but are actually capable of being set to read-only in markup.

Which makes some sense in the context of XForms, where form controls are what 
you care about styling.  Outside of that context, that seems like a very 
unreasonable restriction.

>    The XForms spec clearly states :read-only selects a form control, so
> if :read-only is a "way to style elements which are in the respective
> states as defined by XForms", then it can't apply to a non-control element.

Sure it can, if the non-control element is not in the XForms namespace (if 
nothing else, you can then just style XForms-namespace content that matches 
:read-only, if desired).

>> WF2 is claiming to be doing exactly such clarification, if you note.
> 
>    WF2 can suggest how styling should be handled, as XForms did, but it
> needs to ultimately be defined by CSS.

Actually, no.  CSS defers to document languages on a number of issues; HTML5 and 
specifically the Web Forms 2 part of it is such a language.  XForms is another 
language.  CSS just defines that a :read-only psuedo-class exists and leaves it 
up to the document language to define what is matched by it.  XForms has such a 
definition.  So does Web Forms 2, but the Web Forms 2 definition seems 
inadequate to me in the context of HTML5.  If Web Forms 2 were somehow separate 
from HTML5 that might be OK, but it's not.

>    The width of the checkbox is 100 pixels. You should have used the
> :disabled pseudo-class from CSS3-UI:

I realize :disabled would match there.  The question is why :read-only should 
not match -- the checkbox is readonly in this case; the user can't change its value.

Again, this comes back to the basic question of "what does :read-only select?" 
Is it "read-only elements" or "form controls that have a readonly attribute in 
the DTD and have it set"?  The former seems more useful to me from a general 
user-interface basis.  You seem to be convinced that it should be the latter, 
with "that's what XForms does" as the argument.  I think that this is a case 
where HTML5 shouldn't copy XForms.

>> You seem to be confusing the "readonly" attribute and the :read-only CSS 
>> pseudo-class...
> 
>    Not at all. See the following URLs:
> 
> http://whatwg.org/specs/web-forms/current-work/#readonly
> http://www.w3.org/TR/html4/interact/forms.html#adef-readonly

Those both talk about the "readonly" attribute.  They don't have any mention of 
:read-only.  I stand by my original statement.

-Boris