Re: Ongoing objection to RDFa Profiles format (as XHTML+RDFa)

Ivan Herman wrote:
> Nathan,
> before you really go down that route...

just exploring what else may be possible, if RDF isn't the answer and we 
need something name-value then what are the options? We can't assert we 
need profiles but not define them.

> On Oct 8, 2010, at 14:10 , Nathan wrote:
>> Define a new HTML element <prefix> with two attributes @name and @value. to be used as such:
>> <prefix name="foaf" value="">
> Although very early versions of RDFa (if I remember well, I was not part of the group back then) did introduce new elements, the group made a very conscious decision to express everything in terms of attributes only. (Hence the name, RDF in attributes). The reason is that adding a new element to a host language is much more demanding. It is generally true that, say, a browser will simply ignore an attribute it does not understand but will add that attribute to the DOM tree, it may screw up the presentation algorithm of a browser if it hits an element it does not understand. On the other hand, convincing the HTML5 group to add a new, RDFa specific element to HTML5 is a lost cause.

Embarrassing, but I hadn't noticed the RDFa = "RDF in attributes" before!

I would point to my earlier mail [1] and the quote from the HTML WG that 
"it's good to allow other standards to define new HTML elements and 
attributes" in this case, but also I have a general notion that given 
all the focus on RDFa in both proposals [2][3] for the big open issue 
[4] they have about extensibility and this response [5] that we should 
probably be giving some official feedback from the RDFa WG and either 
work with them to come up with something or wait and see what they come 
up with as a general solution and see if we can tie in with it (or get 
them to come up with it), especially given some of the comments in [5].

Also interested to hear what Mark, or anybody in the WG thinks of the 
proposal/approach "in a perfect world" where we could standardize this 
without problem. And moreover, what are the alternatives to this or RDF?

Anyways, in some ways this is just me getting up to speed on the issue 
before the next meeting, so bar our previous round of need "rdfa 
profiles or not" don't take anything I say as having too much conviction 
behind it, just exploring :)


Received on Friday, 8 October 2010 13:07:26 UTC