Re: Turtle data-type

On 23 May 2013, at 15:38, Steve Harris <steve.harris@garlik.com> wrote:

> On 2013-05-23, at 14:02, Henry Story <henry.story@bblfish.net> wrote:
>> 
>> On 23 May 2013, at 14:59, Steve Harris <steve.harris@garlik.com> wrote:
>> 
>>> On 2013-05-20, at 22:55, Henry Story <henry.story@bblfish.net> wrote:
>>> 
>>>> 
>>>> On 20 May 2013, at 23:46, Steve Harris <steve.harris@garlik.com> wrote:
>>>> 
>>>>> Yeah, it's been suggested many times - never seen it in the wild though.
>>>>> 
>>>>> If you want to use it, just define the datatype, it's not a Turtle feature per se - should work just as well in RDF/XML or whatever.
>>>> 
>>>> It would be useful to have a W3C URL for this. As we may need this in the LDP working group.
>>> 
>>> I think that would be a very bad idea, until the issues are shaken down.
>>> 
>>> Unless the LDP working group wants to define the semantics of the datatype, but I'd suggest that's a huge job.
>> 
>> 
>> It refers to the set of triples you get by interpreting the content as Turtle. Is that so complicated?
>> The only question would be what does an inconsistent turtle document refer to? But that is the same problem
>> for any literal datatype. For example what does "assds"^^xsd:integer refer to?
> 
> I've developed the habit of not thinking things like "how hard can it be" when dealing with semantics. The answer is generally "very".
> 
> You could well be right, but the cost of taking a stab, giving it a "W3C" URI, and getting it wrong is very high. We, as a community have done that a lot - RDF lists, reification, rdf:alt, rdf:XMLLiteral, plain literals, and probably more.

There is a pretty strong need for this, and it would be very helpful to help us transition from 
Turtle to N3.

It may be a lot of work, it may not be that much. We already have the tools for most of these
things: We have the syntax of Turtle, we have the mapping to semantics. So I am not suggesting
the LDP group do that, as it is beyond their competence. But perhaps another group can put it 
on their list of todos. Perhaps also we could have an initial estime of how much effort it really
would take to do this, with reasons for why.

Henry

> 
> - Steve
> 
> -- 
> Steve Harris
> Experian
> +44 20 3042 4132
> Registered in England and Wales 653331 VAT # 887 1335 93
> 80 Victoria Street, London, SW1E 5JL
> 

Social Web Architect
http://bblfish.net/

Received on Sunday, 26 May 2013 09:15:39 UTC