Re: ISSUE-55: Re-enable @profile in HTML5 (draft 2)

Jonas Sicking On 09-11-07 10.08:

> On Wed, Nov 4, 2009 at 7:08 AM, Julian Reschke <julian.reschke@gmx.de> wrote:
>>>> it's an indicator that a certain extension is used. This is relevant
>>>> when multiple extensions occupy the same extension point in the
>>>> syntax.
>>> (I think a better solution is not occupying the same syntax as
>>> something else is occupying.)
>> Yes, that would be better. But it's hard to achieve when there aren't sufficient extension points.
> 
> Is microdata a sufficient extension point? If 'no', why not?


@profile is a solution that, as you said, "is not occupying the 
same syntax as something else is occupying". The HTML 4 invalid

	<head class="htt://myprofile.example.com">

and the HTML 5 draft invalid

	<head profile="http://myprofile.example.com">

differs in that we know that the purpose of the URI is to link to 
the page that explains the semantic conventions applying to the 
document. While a URI inside @class is without specific meaning.

As to microdata, then it could probably be possible to use it 
inside the profile page at "http://myprofile.example.com", much 
the same way RDFa is already used in e.g. the XHTML vocabulary 
profile at "http://www.w3.org/1999/xhtml/vocab".

But, for those pages that were using the "myprofile.example.com" 
profile or the XHTML vocabulary profile, then they could not get 
the same as they get from @profile whether from RDFa or from 
microdata.

RDFa has, however, been defined so that it can make use of 
profiles. And one should think that microdata could as well (be 
defined to) make use of them. (Thus I don't understand why, you 
continue to be against @profile.)

The profile concept is much wider than both RDFa and microdata.

RDFa and microdata are not class names, meta element names or rel 
values. They can be used to define a meaning of particular value 
names, but they are not themselves those values. Thus, unless e.g. 
the class names are *purely* (in some sense) semantical, then one 
cannot replace them with RDFa or microdata.

A profile can for instance apply a particular meaning to the class 
name "particular-UA". Using that class name, I may for instance 
target or exclude a particular user agent based on its CSS 
capabilities or CSS incapabilities.

Like Wikipedia says: [1] "In standardization, a profile consists 
of an agreed-upon subset and interpretation of a specification."

This is not something one can use RDFa or microdata for.

[1] http://en.wikipedia.org/wiki/Profile_(engineering)
-- 
leif halvard silli

Received on Saturday, 7 November 2009 20:53:14 UTC