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

On Fri, 28 Aug 2009, Jonas Sicking wrote:
> On Fri, Aug 28, 2009 at 1:23 AM, Ian Hickson<ian@hixie.ch> wrote:
> > On Fri, 28 Aug 2009, Julian Reschke wrote:
> >> Ian Hickson wrote:
> >> >
> >> > If you think it's obvious, maybe you would be willing to explain it 
> >> > to me?
> >> >
> >> > The only interpretation that I can see is that Mike means 
> >> > non-normative text giving an introduction to the feature to help 
> >> > authors use it. That, however, is not a definition, and would in 
> >> > any case be inappropriate for obsolete features such as those being 
> >> > discussed here.
> >>
> >> Defining what an element or attribute means isn't "informative"; it's 
> >> an essential part of specifying a vocabulary.
> >
> > If by "defining" you mean text with no normative conformance criteria, 
> > that is untestable, and whose only purpose is to help authors work out 
> > what the feature is for, then no, that's an essential part of 
> > specifying a tutorial and is basically only fluff at the specification 
> > level. It's useful, important even, for features we want authors to 
> > use, but it has no use whatsoever for obsolete features that authors 
> > aren't allowed to use.
> 
> I assume that you don't include specifying semantic meaning in "fluff"? 
> I.e. the spec has always defined that <em> has the semantic meaning of 
> "emphasis".

Yes, but it does so in a precisely defined way that normatively references 
the word "represents" which is defined in the rendering section. It's a 
testable stattement (however weakly so).


> Like-wise the outline algorithm specifies semantic meaning of headers 
> and what they cover.

Indeed, these are normatively testable statements.


> I *think* what is being asked for is defining the semantic meaning of <a 
> name="...">.

That kind of normative statement already exist for <a name=""> in HTML5. 
There's even a past-tense statement saying what it could be used for in 
the past. What doesn't exist is a present-tense statement saying what <a 
name=""> is to be used for in HTML5, since its use is a "should not".


On Fri, 28 Aug 2009, Julian Reschke wrote:
> > >
> > > Defining what an element or attribute means isn't "informative"; 
> > > it's an essential part of specifying a vocabulary.
> > 
> > If by "defining" you mean text with no normative conformance criteria, 
> > that is untestable, and whose only purpose is to help authors work out 
> > what the feature is for, then no, that's an essential part of 
> > specifying a tutorial and is basically only fluff at the specification 
> > level. It's useful, important even, for features we want authors to 
> > use, but it has no use whatsoever for obsolete features that authors 
> > aren't allowed to use.
> 
> Well, in that case HTML5 is unsuitable as the *only* specification 
> referenced by the text/html media type registration.

I disagree.

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

Received on Friday, 28 August 2009 18:07:03 UTC