Re: FPWD comment - literals, data types and language tags

Hi Randall,

On Mon, Dec 15, 2014 at 12:53 PM, Randall Leeds <randall@bleeds.info> wrote:
>
> On Mon, Dec 15, 2014 at 1:36 AM, Jacco van Ossenbruggen <
> Jacco.van.Ossenbruggen@cwi.nl> wrote:
>
>> Having to explain to developers with an RDF background why this triple
>> > <> oa:body "hi"@en .
>> is illegal in OA would be a royal pain.
>>
>
> Can anyone clarify for me? I am not understanding why this would be an
> illegal triple.
> Isn't this object the resource {"@language": "en", "@value": "hi"} ?
> What's illegal here? I only see that the resulting triple does not
> serialize as a literal in JSON-LD, but it does have a valid serialization.
>

That's a literal, not a resource (which demonstrates the confusion or at
least the intuitive terminology problem :) )

It's currently illegal according to the spec because you can't have a
language tagged literal as the body, only an xsd:string. All language
tagged literals have a special data type in RDF "
http://www.w3.org/1999/02/22-rdf-syntax-ns#langString" (see:
http://www.w3.org/TR/rdf11-concepts/#section-Graph-Literal) and thus cannot
be of any other data type.
You also can't give it any other properties, type or identity.

Allowing it would make all of the possibilities I listed earlier in the
thread valid, requiring systems check for the different variants for the
sake of making developers' lives ... easier?

Rob

-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305

Received on Monday, 15 December 2014 21:08:24 UTC