Re: Functions in N3 (was: Re: What Happened to the Semantic Web?)

On 02/12/17 17:42, Gregory Williams wrote:
> On Dec 2, 2017, at 9:31 AM, Martynas Jusevičius <martynas@atomgraph.com> wrote:
>>
>> Not SPARQL versions such as STRSTARTS though? One could argue that these are distinct resources, and belong in the SPARQL namespace.
> 
> Yes. Some of the functions don’t have corresponding xsd/xpath functions, and most (I believe) have different restrictions on domain/range. What I had defined (but didn’t get published in the w3 namespace) was simply a list of `a sd:Function` (or sd:Aggregate) type statements for all of the functions/aggregates defined in the spec.
> 
> Thanks,
> Greg

Some don't - e.g. IRI and several others in 17.4.2.

STRSTART is fn:starts-with

But if it's F&O op: theer is no URI.

RIF functions are not identical to F&O or SPARQL - they handle argument 
errors differently making the evaluation undefined, whereas SPARQL and 
F&O signal an error.

e.g. last sentence of:

https://www.w3.org/TR/2013/REC-rif-dtb-20130205/#pred:numeric-equal_.28adapted_from_op:numeric-equal.29

     Andy

> 
> 

Received on Saturday, 2 December 2017 19:10:23 UTC