Prefix Definition

[marqh]

Should a prefix definition include the double underscore character, so that cat prefix alias results in a URI?

• this is current behaviour

Or

Should the prefix separator be part of the standard, and not be included in the prefix definition

• this would require a code change and changes to the examples

[marqh]

for example:
https://github.com/binary-array-ld/bald/blob/master/lib/bald/tests/integration/CDL/array_reference.cdl#L13

    prefix_list:bald__ = "http://binary-array-ld.net/latest/" ;
    prefix_list:rdf__ = "http://www.w3.org/1999/02/22-rdf-syntax-ns#" ;

should the double underscore __ always be implied? Such that

    prefix_list:bald = "http://binary-array-ld.net/latest/" ;
    prefix_list:rdf = "http://www.w3.org/1999/02/22-rdf-syntax-ns#" ;

[marqh]

For comparison, the terse triple language (ttl) includes the prefix separator (':') in the prefix definition:
https://github.com/binary-array-ld/bald/blob/master/lib/bald/tests/integration/TTL/array_reference.ttl#L1

@prefix bald: <http://binary-array-ld.net/latest/> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

I suspect it was this pattern that we initially adopted.

[jyucsiro]

The other pattern relevant here for the alternative approach is taken from JSON-LD - see https://json-ld.org/spec/latest/json-ld/

Example:
{
"@context": {
"foaf": "http://xmlns.com/foaf/0.1/",
"xsd": "http://www.w3.org/2001/XMLSchema#",
"name": "foaf:name",
"foaf:age": {
"@id": "foaf:age",
"@type": "xsd:integer"
},
"http://xmlns.com/foaf/0.1/homepage": {
"@type": "@id"
}
},
...
}

[jyucsiro]

The current proposal is to provide prefixes in a variable such as "bald_prefixes". In this case the parser knows where to get the prefixes from so not having the double underscore in a prefix declaration is not a problem.

Pros might be:

• Requires less characters and therefore might result in less user entered errors.

Cons:

• Does mean a bit more processing by the bald library (and a code change/change to examples).

The double underscore pattern - pros:

• movable to other parts of the netCDF headers, i.e. outside of a variable which solely contains "bald_prefixes".
• current behaviour

Cons

• Could result in user errors? Extraneous characters.