JSON values

[pfps]

With rdf:JSON there are now JSON values. But what is a JSON value?

For example, in SPARQL should

"7"^^rdf:JSON = 7

be true or false?

[afs]

Good question.

It's type error. The dispatches are based on datatype.

Other points:

XSD does not define a type conversion from rdf:JSON in the same way it does for xsd:integer to xsd:double.

xsd:double is "patterned after IEEE 754-2008", and has been made to "mirror IEEE 754-2008", but XSD 1.1 does not say the value space "is" IEEE 754.

You could ask about booleans where the value space (all defs) is true and false.

[pfps]

How did xsd:double get in here?

[pfps]

How about

' "7" '^^rdf:JSON = "7"

[gkellogg]

I think it's the same point here, there is no defined way to compare rdf:JSON literals with xsd:string.

[afs]

How did xsd:double get in here?

F&O defines operations on numerics by applying subtype substitution and numeric type promotion.

(double, integer) becomes (double, double).

JSON numbers are IEEE 754 , xsd:double are related to IEEE 754.

[pfps]

@gkellogg But isn't the value of ' "7" '^^rdf:JSON a string?

[pfps]

@afs But isn't the value of "7"^^rdf:JSON just a number, not an IEEE floating point number?

[gkellogg]

@gkellogg But isn't the value of ' "7" '^^rdf:JSON a string?

Is an INFRA/ECMAScript string the same as an XSD string? At a certain level, but where is that defined? Is a JSON number the same as an xsd:float/double? Or boolean?

I'm note sure we need to define comparison between JSON values and any other.

[afs]

These one line comments without background leaving a lot open as to what the point being made actually is.

But isn't the value of "7"^^rdf:JSON just a number,

JSON number value is "corresponding" -- https://tc39.es/ecma262/#sec-terms-and-definitions-number-value

[afs]

Just because two values space contain e.g. strings does not mean they are comparable for two reasons:

1:
The value space is the range of the datatype mapping function.

https://www.w3.org/TR/xmlschema11-2/#value-space

The value space of a datatype is the set of values for that datatype.
...
The relations of identity and equality are required for each value space.

RDF Semantics:

The value space of a datatype is the range of the [lexical-to-value mapping](https://www.w3.org/TR/rdf12-concepts/#dfn-lexical-> to-value-mapping).

"The range" is specific to the datatype.

2:
This is reflected in how operators are mapped to (XQuery/XPath) functions.

Operators are matched to functions, in SPARQL, by dispatch, and the functions are datatype specific.
There is no assumption that the two sets of values of unrelated datatypes can be compared for = which is equality by datatype.

JSON-LD makes this clear:

Despite being defined as a set of strings, this value space is considered distinct from the value space of xsd:string, in order to avoid side effects with existing specifications.

[afs]

Is an INFRA/ECMAScript string the same as an XSD string?

No - https://tc39.es/ecma262/#sec-terms-and-definitions-string-value are encoding in UTF-16.

xsd:strings are characters, after decoding.

Infra:

The Infra Standard aims to define the fundamental concepts upon which standards are built.

should be read as specific to whatwg.

JSON Numbers are binary patterns of IEEE 754, not rational numbers - encoding vs abstract value.

[pchampin]

can we close this issue as superceded by w3c/rdf-concepts#116 ?

[pfps]

Agreed, closing