Re: ISSUE-53: mediatypereg - suggest closing on 2009-09-03

On Tue, 25 Aug 2009, Roy T. Fielding wrote:
> 
> Real content management systems are like fine-grained mash-up systems.  
> Sometimes the input is legacy and is farmed out in various formats.  
> Sometimes the output is legacy (sequence of bytes provided to the public 
> must not change) even though the inputs have been "managed" into more 
> reusable parts.  Both the input and the output are valid "text/html" 
> even if neither is conforming HTML5.

The whole point is that "valid text/html" and "conforming HTML5" should be 
the same, and that legacy documents that were valid text/html in the past 
but use now obsolete features should no longer be considered valid 
text/html.


> HTML5 draft says (in a section about validators):

(That section isn't about validators, the next one is.)

>    Authors should not specify the name attribute on a elements.
>    If the attribute is present, its value must not be the empty string.
>    In earlier versions of the language, this attribute served a similar
>    role as the id attribute. The id attribute should be used instead.
> 
> HTML 4.01 says (in the section on <a>):
> 
>    Attribute definitions
>      name = cdata[CS]
>        This attribute names the current anchor so that it may be the
>        destination of another link.

That's inaccurate, by the way. The attributes does more than that, e.g. it 
also affects HTMLCollection processing.

>                                     The value of this attribute must
>        be a unique anchor name. The scope of this name is the current
>        document. Note that this attribute shares the same name space
>        as the id attribute.
> 
> HTML5 draft doesn't define what the attribute means -- it only says
> it once had a role similar to id.

The attribute doesn't mean anything any more (though it has some authoring 
conformance criteria).


> It doesn't specify that its value must be a unique anchor name, which is 
> a significant statement for link checking software that verifies such 
> things as destinations.

Fixed the part about uniqueness. The anchor name part is already defined 
as part of the "text/html" MIME type definition in HTML5.


> It doesn't specify that it shares the same name space as the id 
> attributes, which again is significant for both link checkers and 
> content management.

Fixed.


> Moreover, what it does say about the subject is placed far away from 
> where a reader would be expected to look up a definition for this funky 
> "name" attribute they happened to see on an anchor in some "text/html".

This will be resolved when we fill in the attribute index.


Thank you for giving actual concrete feedback that I could address.

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

Received on Tuesday, 25 August 2009 23:44:44 UTC