Re: Re-registration of text/html

On 10.03.2010 18:25, Kornel Lesinski wrote:
> On 10 Mar 2010, at 16:36, Tantek Çelik wrote:
>>>
>>> Isn't that statement true for the majority of existing usage of the profile
>>> attribute? And therefore a SHOULD requirement is adequate?
>>
>> If it's true for the majority implementations of profile attribute,
>> then a MAY requirement is sufficient.
>>
>> No reason to encourage (which is what a "SHOULD" is, an encouragement)
>> breaking of existing implementations such as GRDDL processors.
>>
>> MAY wording here also seems more compatible with allowing user agents
>> to implement the HTML5 Profile attribute extension.
>
> Is GRDDL supposed to work with text/html? The GRDDL spec seems to focus on XHTML. If it's for XHTML only, then it doesn't matter what text/html registration says.

The registration for application/xhtml+xml is identical with respect to 
this.

> Do GRDDL processors actually rely on profile attribute?
> When researching pages with microformats (on 'anecdotal' scale) I found that that few of them used profile at all, and some of them had irrelevant/invalid profile, e.g. XFN profile on page with hCard. In case of microformats I think that tool which respected profiles would be worse off than tool that ignores profiles completely (I don't know if that's true for GRDDL in general).
>
> Are there conflicting vocabularies used in the wild that couldn't be processed correctly without disambiguation with profile? Couldn't<link rel=transformation>  be used instead?

For instance, eRDF and DC-HTML use the same extension point. If they 
ever become incompatible, the @profile URI would be needed to 
distinguish the intent of the author.

Even if this is an edge case: why break it (except for political reasons)?

Best regards, Julian

Received on Wednesday, 10 March 2010 17:48:21 UTC