Re: ISSUE-55: Re-enable @profile in HTML5 (draft 1)

Manu Sporny wrote:
> I volunteered for an action to produce a draft document that included
> @profile in HTML5:
> 
> http://www.w3.org/html/wg/tracker/actions/144
> 
> The document attempts to address the long standing ISSUE-55:
> 
> http://www.w3.org/html/wg/tracker/issues/55
> 
> After speaking with Julian over the weekend, we explored the possibility
> that perhaps the document should focus on the concept of extending
> processing behavior for user agents via @profile and @version. Perhaps
> even stating that we'd like to move away from @profile (because it
> hasn't been used as much as we'd like), toward something that is simpler
> and has been used in both XHTML and RDFa: a specifically formatted
> @version attribute.
> ...

It appears trying to do multiple things at once complicates things.

Thus, may I suggest to divide this into smaller chunks we can develop, 
and hopefully find consensus on?

For @profile the things to do include:

1) Agreeing on what HTML 4.01 should have said (acknowledging it's 
inconsistent, and thus contributes to the confusion)

2) Documenting @profile as conforming, either separately, or inside 
HTML5 (where I would prefer the latter)

3) Optionally find a better replacement for @profile, which *could* be 
link/@rel=profile. Note that deploying that will require updates to 
those specs that currently use @profile, such GRDDL or DC-HTML, so it 
will require a transition period in which @profile would stay conforming.

Attached is a proposal for 1). I realize that we're not doing errata for 
HTML 4.01, but IMHO having this phrased this way will be helpful for 
steps 2) and 3).

Best regards, Julian

PS: thanks to Björn for proofreading