mmisw / udunits2rdf Goto Github PK

We want the vocabulary for prefixes as well. They aren't units, they are prefixes. I captured them (shown here in n3 form) as

@prefix :        <http://mmisw.org/ont/mmitest/udunits2-prefixes/> .
@prefix u2:      <http://mmisw.org/ont/mmitest/udunits2/> .

:_782a1369 
  a           u2:prefix  ;
  u2:hasValue:    1e24   ;  // they are all numbers, do I need quotes?
  u2:hasName:     :micro  ;
  u2:hasSymbol:  "&#xB5"  ;  // can n3/RDF add an annotation 'MICRO SIGN' (for the comment) ?
  u2:hasSymbol:  "&#x3BC" ;  // can n3/RDF add an annotation 'Greek small letter "mu"' ?
  u2:hasSymbol:  "u"   .

Not sure if these minor points can be managed via udunits2rdf, but maybe some can:

• Can the order of columns be name, type, namesUnit, hasName, hasAlias (then the rest)?
• Can the right side of the table have a border?

http://mmisw.org/orr/#http://mmisw.org/ont/mmitest/udunits2-base only has the classes, none of the actual units

am guessing that's because the unit tags are under <unit-system> tag.

Note also there is a tag, indicating this is a base unit. This is a valuable attribute to capture, IMHO. (Don't know what the property type should be. Perhaps

unit hasBasis "base"

Then, later we could make the units in the other files

unit hasBasis "common" / "derived" / "accepted"

udunits2-base has unneeded Prefix class, udunits2-prefixes has unneeded Unit and UnitName classes. (Won't bother me if you choose to wontfix this.)

In the presented vocabularies, unit names are also shown in the hasAlias column. This doesn't match the n3 example in the model, in which the name is not an alias.

udunits2-derived.xml contains:

    <unit>
        <dimensionless/>
        <name><singular>radian</singular></name>
        <symbol>rad</symbol>
    </unit>

How should the <dimensionless/> part be translated?

Per Carlos request, create an example or two of the desired relations for UDUNITS based on the XML. (Issue created so we can track progress easily.)

Per #10 , document the purpose somewhere -- I think in the udunits2rdf wiki is good -- and refer to it from the ontology metadata.

Parse the comments from the original file and include (where appropriate) for the unit or unitName

The Steering Committee raised the following issues, which I will discuss with Julie and split out as needed:

• There's something odd about mixing both opaque units and transparent names; not what users would expect/understand
• Unit descriptions are entirely missing (because they are in comments)
• Table has lots of blanks (because it is showing two types of concept, unit and alias)
• What's the expectation for usability of these unit URIs in external system? document the answer
• What is up with the coupling of units to names?

It's doubtful that the difference between symbols and alias symbols carries much importance in UDUNITS. Mostly this seems to reflect whether the symbol is for an alias or for a name; but in at least one place it looks like the alias symbol is just less important. (Or, maybe the alias originally existed but was deprecated, and the symbol remains.)

A fully faithful rendition of the vocabulary should capture this distinction (e.g., by giving the alias symbols a hasRole "alias" property).

I'd call this pretty low priority.

Steve Emmerson remarked, in the context of an issue with dram:

1. Each name should unambiguously identify a unit;
2. Each symbol should unambiguously identify a unit;
3. Long names (e.g., "avoirdupois_dram") should be supported.

I think there's at least one case where the same name is used in two different places, one of them in the derived vocabulary and one elsewhere. I'm not sure how we dealt with it; this is a reminder to me to find out and confirm it makes sense.