Re: RDFa API: ondata="javascript: foo();" attribute for <body> element?

Manu Sporny wrote:
> I haven't thought this through very well, but was wondering if it would
> be advisable to have an ondata attribute for the <body> element? This
> would work very much like the onload attribute that is already on the
> <body> element.
> 
> I can see a few implementers complaining that the RDFa processing of a
> document could take a long time (as in more than 100ms), especially if
> certain RDFa profiles take a while to retrieve or timeout.
> 
> The worst case is that a bad implementation of the RDFa API would cause
> the document not to render because the RDFa profiles are being
> fetched/processed. Ideally, RDFa processing should happen asynchronously.

Perhaps then all parsing should happen asynchronously.

DataStore could implement EventTarget and provide and dispatch the 
onload event whenever processing has completed.

   var store = document.data.createStore();
   store.onload = function(event) {
     // triggered when the store is ready
     // i.e. once parsing is finished
   }
   var parser = document.data.createParser("rdfa1.1", store)
   parser.parse( somedocument );


Actually, thinking about this more, an XMLHttpRequest style readystate 
may be needed, ie add the following:

interface DataStore : EventTarget {

   attribute Function onerror
   attribute Function onload
   attribute Function onreadystatechange;

   const unsigned short EMPTY = 0;
   const unsigned short LOADING = 1;
   const unsigned short DONE = 2;

   readonly attribute unsigned short readyState;

}

Ack, this brings up a whole host of (sadly unavoidable) questions - I'll 
mail them under separate cover - sorry.

However, it seems to make sense for DataStore to work async and dispatch 
events - that can be the take away - well caught Manu.

Best,

Nathan

> However, if it happens asynchronously, how do we signal developer code
> that the RDFa/Microformats/Microdata has been extracted and is ready to
> be read.
> 
> The only approach that I could think of (in the 10 minutes that I've
> been thinking about the problem) is to have an "ondata" attribute on the
> BODY element that is called whenever all data extraction has completed
> for the document and there is data available in document.data.store.
> 
> Have we thought about this problem before?
> 
> -- manu
>

Received on Saturday, 4 September 2010 19:09:33 UTC