Re: Understanding 'chaining'

Mark Birbeck wrote:
> That doesn't sound like deadlock to me. Of course it might be, but I
> think we should give Ben a bit of space to reply before throwing our
> hands up, and declaring game over.

I didn't mean it to come across as if we should give up! :)

By all means, Ben should reply so we know exactly what each of you are
talking about... I didn't mean to short-circuit that process. I do
believe, however, that when Ben clarifies his position - we are going to
be in the same spot of having to pick one uncomfortable approach over
the other and accept all of the side-effects of that approach. I don't
see many side-effects with Ivan's approach - perhaps, I don't fully
understand the implications.

Keep in mind that I agree with your approach, Mark, but I'm starting to
think that it's not fair to assume that the XHTML author will be able to
grok all of the side-effects of either your or Ben's approach.

What I meant to express was that both you and Ben have made very
compelling cases for each approach. So compelling, that something seems
a bit off about having to pick one over the other.

When we've analyzed what each of these rules means in the big picture,
test cases included, the group has looked at the example and asked "What
does the author mean by this bit of XHTML+RDFa?"

The easiest concepts to understand in RDFa, IMHO, were those that seemed
to have very little disagreement on what the markup actually meant.

Contrast this with the @instanceof discussion. The last few weeks of
e-mails have helped everybody understand both your and Ben's positions.
 So while /understanding/ has been improving, I do not believe that we
are any closer to an /agreement/. Although, it may be that I am alone in
that opinion :)

>  * myself and Ben had reduced our 'differences' very precisely to the
>    behaviour of one attribute in one position;

While this is true, it is one attribute in one position that changes
quite a few things.

The elegance of Ivan's approach is that it eliminates both Ben and your
differences completely. It allows the XHTML author to be specific about
what they mean instead of us attempting to guess what they mean. Even if
we are correct for 70% of the authors, there will be 30% that won't like
that @instanceof applies to the "wrong thing".

The mere fact that there are several of us that look at the same markup
and see different things should be an indicator that other authors might
have slightly different mental models. Rather than force authors into
our implied mental model, we can provide a method that is more specific
and has very few side-effects. I believe Ivan's proposal has those two
benefits over both yours and Ben's.

That being said, I'll take some more time tonight to apply Ivan's
proposal to the test suite and see if any nasty side-effects emerge.

-- manu

-- 
Manu Sporny
President/CEO - Digital Bazaar, Inc.
blog: Over One Million Songs Available on Bitmunk
http://blog.digitalbazaar.com/2007/10/29/one-million-songs-on-bitmunk/

Received on Tuesday, 27 November 2007 22:55:37 UTC