Re: What cool stuff have we done?

Changed rdf:XMLLiteral so that it doesn't require XML canonicalization. This means implementations and authors can use it in the obvious way.

Added xsd:duration. Added the new XSD 1.1 datatypes xsd:dayTimeDuration,xsd:yearMonthDuration and xsd:dateTimeStamp .

All literals, including language-tagged ones, now have a data type in the abstract syntax. Will make APIs more internally consistent.

Richard


On 9 Apr 2013, at 02:09, Sandro Hawke <sandro@w3.org> wrote:

> Tomorrow evening [1] I'm presenting about all the cool new stuff in RDF 1.1.   What should I put on the list?   (I'll also be doing this in June at SemTech.)
> 
> Here are the things I can think of, that I can say proudly to this audience, which I think has lots of SPARQL experience and some OWL experience:
> 
> - Turtle now meticulously defined, with extensive test suite (will be joining RDF/XML and RDFa as "Recommended" syntax for RDF)
> - Turtle no longer sometimes needs a space before the period
> - JSON-LD allows nice-looking JSON to be read/written as RDF (another "Recommended" syntax for RDF)
> - standard way to replace blank nodes with IRIs (genid)
> - rdf:HTML datatype provides a practical way to put HTML fragments in RDF
> - plain literals (w/o lang) and xs:string literals are now the same thing
> - sparql's named graph / dataset model is now clarified and part of RDF
> - trig provides a syntax for datasets (collections of separable graphs)
> - in general, documents are being rewritten to be simpler and clearer - for now, see http://www.w3.org/TR/rdf11-concepts/
> 
> What am I missing?   What else are you really pleased about or proud that we've done?   I only have 20 minutes, so I think the above list, with one slide per item, is about the right length.
> 
> Also, depending on time and the mood of the audience, I might ask about some of our remaining issues (really the At Risk ones).  The two that I think I can explain very clearly:
> 
> - should turtle have PREFIX & BASE like SPARQL or @prefix & @base like n3?  (Currently we're allowing both styles)
> 
> - can you use a blank node as the graph label in a dataset, or do you have to invent a URL for each graph?  (Currently JSON-LD and some SPARQL engines allow them, but the Trig and Concepts spec disallow them)
> 
> Input before 2pm Boston Time most welcome.
> 
>     -- Sandro
> 
> [1] http://www.meetup.com/The-Cambridge-Semantic-Web-Meetup-Group/events/104611672/ 
>