Re: Combined proposed IDL for RDFa API

Toby Inkster wrote:
> On Mon, 2010-11-01 at 12:30 +0000, Nathan wrote:
>> I'm aware that with all the open issues and conversations etc around the 
>> API that it's getting a bit confusing, thus here's a combined IDL from 
>> all proposals, issues, comments etc and I guess how I'm personally 
>> seeing the API at the minute should all issues be resolved in line with 
>> this...
> 
> Left you some comments in IRC:
> http://chatlogs.planetrdf.com/swig/2010-11-03.html#T16-35-28
> 
> Also some comments for Ivan there too.

Hi Toby,

Thanks,

re: "have you considered dropping DataContext and moving all its members 
to DocumentData?"
yes, strongly, but was thinking it would be good to clear all the other 
issues first and see if it's still needed.

re: " my other suggestion is to add 
SubjectOrigin/propertyOrigin/objectOrigin as readonly attributes on 
RDFTriple with type "any". For data parsed from RDFa, these would return 
the XML elements where the subject, predicate and object of the triple 
were gleaned from. For triples parsed from other sources, these might 
have other values (e.g. for Turtle, line numbers) or be null."

kind of, was thinking it may not actually be needed at all, but time 
will tell - I've still got to implement the document extensions so will 
know then, if it turns out it really is required and can be made to 
work, then I'd probably propose an RDFaTriple which extended RDFTriple 
and was only returned by methods on the document interface, as not all 
parsers will even have a DOM.

Finally, probably worth mentioning that all of the proposed IDL and a 
lot more is implemented in:
   http://webr3.org/apps/play/api/rdfapi.js

details here:
   http://webr3.org/apps/play/api/lib

and all tested / proven, but ironically missing RDFa support at the minute.

Best,

Nathan

Received on Wednesday, 3 November 2010 18:02:08 UTC