[whatwg] RDFa Features

Kristof Zelechovski wrote:
> "Does not use QNames" is not an advantage any more than "does not require
> the user to be a USA citizen".  So you could have listed that as well.
> I would like to append the following to the disadvantages:
> The interface A[property] is very misleading.  You read it as "The property
> of this anchor is cc:creator".  That does not make any sense (an anchor has
> several properties and cc:creator is not one of them) and it does not
> reflect what you want to say, which is "this anchor represents a property
> named cc:creator of its enclosing element".  I am not sure how to fix this;
> "isProperty" would be better but still not very good.

Let's make sure not to discuss everything at once.

The first point is that embedded metadata with RDFa has clear use cases
and does not lead to any more logical issues than normal human-readble
web pages. I don't know if you agree with that yet, but we think we've
made some solid points to support this argument.

The second point is: what should the names of the actual attributes be?
Well, we can argue about that till the cows come home. No matter what
you choose, a lot of people will be unhappy because they "read" the HTML
in different ways. We think we've picked attribute names that are rather
harmless and that are "close enough" to what we mean. And we had an open
discussion about it for 2 years. At some point, you have to declare that
an issue like this one is "good enough."

Let's try keep those two points of discussions separate.

-Ben

Received on Friday, 29 August 2008 08:21:23 UTC