Re: [RDFa] ISSUE-3 @class and @role for rdf:type

Hi Ben,

> > I pressed send to early....anyway I don't think it's surprising that
> > I'm confused on what we're voting on, if you look at your first email
> > that kicked off this thread.
>
> I see what you're saying now. The way I phrased it, it sounded like
> everything is up in the air. In my mind, I was thinking that we were
> simply formalizing what we had already informally agreed upon.
>
> So, I'm still hoping we can formalize what we informally mentioned in
> the telecon on 5/31, which is that @class is rdf:type and @role is left
> to XHTML2, because otherwise we have a moving target.

If we're supposed to be voting on the resolution of issues, then
that's what we should do. I don't see--outside of the former Soviet
Union--how you can say, 'well we informally agreed on this, so please
vote this way'. I don't want to keep raising this point...but you keep
bringing it back up again. :)

Anyway, where I'd reached in my thread of discussion was that I said I
would _prefer_ a new attribute to represent rdf:type, but I wouldn't
be crying if we settled on @class. However, given that there are
people who have been against using the use of @class (to represent
rdf:type) right from the start--in particular, Steven--I did also say
that I'd like to hear what those people thought on this resolution,
before we finally ticked it off as 'use @class'.

The reason for wanting to hear what they thought is this; up until
recently, the debate has been framed as a choice between using @role
or @class to represent rdf:type. Steven felt @class was wrong, and
therefore argued *for* @role. On the other hand, I felt that @role was
wrong, and therefore argued *for* @class.

But neither of us were arguing for our respective attributes
_positively_. That is, we weren't particularly in favour of the
attribute we were promoting, it was simply that we thought the
alternative was not a good choice, for whatever reasons.

If you think about it, this was the wrong approach, and we should have
started from the requirement, which would have been something like 'we
need a shorthand way to express rdf:type'. And when you come at it
from that way round, you no longer have only two choices--use @class
or @role--you have whatever everyone can agree meets the requirement.

Anyway, although I thought this a while ago I had decided not to raise
the idea of a 'third way'--until Ivan mentioned that he also didn't
like @class. When he said that, I thought it was worth trying to find
a compromise, and proposing the use of a new attribute seemed like a
good one.

But...but...but...I know that sounds like a long-winded justification
for creating a new attribute, but actually it's not. As I said above
and in the previous thread, I can live with @class, so all I'm trying
to justify is keeping the debate open long enough for Steven to say
what he thinks. Ivan has already said he can live with @class, so if
Steven says he can too, then @class it is, and we can mark this
closed.

But if Steven says that he is glad to now have the possibility of a
third choice, then I think it is only fair that what we ask everyone
to vote on is the resolution of the broader *requirement*, and the
possible solutions to that should include the creation of a new
attribute. (I guess the vote would also need to include the
possibility of leaving the whole requirement to a future version of
RDFa, although I would have thought we wouldn't need to do that.)

Regards,

Mark

-- 
  Mark Birbeck, formsPlayer

  mark.birbeck@x-port.net | +44 (0) 20 7689 9232
  http://www.formsPlayer.com | http://internet-apps.blogspot.com

  standards. innovation.

Received on Wednesday, 4 July 2007 21:40:04 UTC