HajoRijgersberg/OM

Robot verification raises issues with the current draft

Closed this issue · 8 comments

rpgoldman commented

I checked OM-2 against the OWL DL profile using robot, as follows:

robot validate-profile --profile DL --input OM/om-2.0.rdf --output OM/validation.txt

Then, because there were a bunch of bibliography info constraints, I did the following filtering:

cat OM/validation.txt | fgrep -v bibo | fgrep -v foaf | fgrep -v purl.org | fgrep -v bibliography > OM/filtered-validation.txt

This gave me the attached file, which mentions 16 issues.
filtered-validation.txt

Trying to track this down, I see the following:

  1. I don't think robot likes the following, because it applies the Union operator to a collection with only one element:
  <owl:Class rdf:about="&om;FahrenheitTemperatureUnit">
    <rdfs:label xml:lang="en">Fahrenheit temperature unit</rdfs:label>
    <rdfs:subClassOf rdf:resource="&om;TemperatureUnit"/>
    <owl:equivalentClass>
      <owl:Class>
        <owl:unionOf rdf:parseType="Collection">
          <owl:Class>
            <owl:oneOf rdf:parseType="Collection">
              <om:Unit rdf:about="&om;degreeFahrenheit"/>
            </owl:oneOf>
          </owl:Class>
        </owl:unionOf>
      </owl:Class>
    </owl:equivalentClass>
  </owl:Class>

There are a number of similar cases like this, in om:FixedPoint, om:FahrenheitTemperatureUnit, om:gramPerSquareMetre, and om:ReaumurTemperatureUnit.

To be quite honest, it doesn't seem to me that having some degenerate Union's is actually bad, but it looks like they could be removed.

  1. There are a few "Use of undeclared class:" in the report, too, for: om:FahrenheitScale, om:CelsiusScale, om:RankineScale, om:ReaumurScale, om:KelvinScale.
HajoRijgersberg commented

Thanx for your action! :)

  1. I'll put this on my to-do list. Hope to solve this soon. I'll keep you updated. P.S.: I agree with you that these Unions actually is not bad. But I'm certainly willing to remove them.
  2. That's strange. These definitions are simply present in OM. For instance, om:CelsiusScale:
  <om:IntervalScale rdf:about="&om;CelsiusScale">
    <rdfs:label xml:lang="en">Celsius scale</rdfs:label>
    <rdfs:label xml:lang="nl">Celsiusschaal</rdfs:label>
    <rdfs:label xml:lang="zh">摄氏温标</rdfs:label>
    <om:alternativeLabel xml:lang="en">International Celsius Temperature scale of 1990</om:alternativeLabel>
    <om:hasDimension rdf:resource="&om;thermodynamicTemperature-Dimension"/>
  </om:IntervalScale>
rpgoldman commented

Actually, I think I see the problem with these items. Robot seems to be saying that it expects these to be classes, but in OM-2 they are individuals.

I can't really read OWL RDF, so I am working from a translation into Manchester form. What I see is this:

Class: <http://www.ontology-of-units-of-measure.org/resource/om-2/FahrenheitScale>

    Annotations: 
        rdfs:label "Fahrenheit scale"@en,
        rdfs:label "Fahrenheitschaal"@nl,
        rdfs:label "华氏温标"@zh
    
    EquivalentTo: 
        <http://www.ontology-of-units-of-measure.org/resource/om-2/FahrenheitTemperatureScale>

and I see this:

Individual: <http://www.ontology-of-units-of-measure.org/resource/om-2/FahrenheitScale>

    Annotations: 
        rdfs:label "Fahrenheit scale"@en,
        rdfs:label "Fahrenheitschaal"@nl,
        rdfs:label "华氏温标"@zh
    
    Types: 
        <http://www.ontology-of-units-of-measure.org/resource/om-2/IntervalScale>
    
    Facts:  
     <http://www.ontology-of-units-of-measure.org/resource/om-2/hasDimension>  <http://www.ontology-of-units-of-measure.org/resource/om-2/thermodynamicTemperature-Dimension>,
     <http://www.ontology-of-units-of-measure.org/resource/om-2/hasScale>  <http://www.ontology-of-units-of-measure.org/resource/om-2/KelvinScale>,
     <http://www.ontology-of-units-of-measure.org/resource/om-2/hasUnit>  <http://www.ontology-of-units-of-measure.org/resource/om-2/degreeFahrenheit>,
     <http://www.ontology-of-units-of-measure.org/resource/om-2/hasFactor>  1.8f,
     <http://www.ontology-of-units-of-measure.org/resource/om-2/hasOff-Set>  -459.67f

Is this intended punning? And, if so, is it robot's problem that it doesn't accept it?

HajoRijgersberg commented

Sorry for my late reply, busy times.
Thanks for checking. It's strange: FahrenheitScale is never defined as a class in OM. It is simply defined as an instance, just like CelsiusScale:

  <om:IntervalScale rdf:about="&om;FahrenheitScale">
    <rdfs:label xml:lang="en">Fahrenheit scale</rdfs:label>
    <rdfs:label xml:lang="nl">Fahrenheitschaal</rdfs:label>
    <rdfs:label xml:lang="zh">华氏温标</rdfs:label>
    <om:hasFactor rdf:datatype="&xsd;float">1.8</om:hasFactor>
    <om:hasOff-Set rdf:datatype="&xsd;float">-459.67</om:hasOff-Set>
    <om:hasScale rdf:resource="&om;KelvinScale"/>
    <om:hasUnit rdf:resource="&om;degreeFahrenheit"/>
    <om:hasDimension rdf:resource="&om;thermodynamicTemperature-Dimension"/>
  </om:IntervalScale>

Does that help?
It's no intended punning or anything. Do you know how/why these scales become classes for the robot?

HajoRijgersberg commented

With respect to the second item, maybe I should report this to the Robot folks as a possible bug.

Strange, I don't see this message anymore in the GitHub. Is that correct or is something wrong?

rpgoldman commented

With respect to the second item, maybe I should report this to the Robot folks as a possible bug.

Strange, I don't see this message anymore in the GitHub. Is that correct or is something wrong?

I'm not sure, to be honest.

rpgoldman commented

Sorry for my late reply, busy times. Thanks for checking. It's strange: FahrenheitScale is never defined as a class in OM. It is simply defined as an instance, just like CelsiusScale:

  <om:IntervalScale rdf:about="&om;FahrenheitScale">
    <rdfs:label xml:lang="en">Fahrenheit scale</rdfs:label>
    <rdfs:label xml:lang="nl">Fahrenheitschaal</rdfs:label>
    <rdfs:label xml:lang="zh">华氏温标</rdfs:label>
    <om:hasFactor rdf:datatype="&xsd;float">1.8</om:hasFactor>
    <om:hasOff-Set rdf:datatype="&xsd;float">-459.67</om:hasOff-Set>
    <om:hasScale rdf:resource="&om;KelvinScale"/>
    <om:hasUnit rdf:resource="&om;degreeFahrenheit"/>
    <om:hasDimension rdf:resource="&om;thermodynamicTemperature-Dimension"/>
  </om:IntervalScale>

Does that help? It's no intended punning or anything. Do you know how/why these scales become classes for the robot?

If that is right, does that imply that Robot has actually created the definition of FahrenheitScale as a class in the process of translating RDF to Manchester notation? I will check!

rpgoldman commented

I think I see the problem:

  <owl:Class rdf:about="&om;FahrenheitTemperatureScale">
    <rdfs:label xml:lang="en">Fahrenheit temperature scale</rdfs:label>
    <rdfs:subClassOf rdf:resource="&om;Scale"/> 
    <owl:equivalentClass>
      <om:IntervalScale rdf:about="&om;FahrenheitScale"/>
    </owl:equivalentClass>
  </owl:Class>

I believe that this should be:

  <owl:Class rdf:about="&om;FahrenheitTemperatureScale">
    <rdfs:label xml:lang="en">Fahrenheit temperature scale</rdfs:label>
    <rdfs:subClassOf rdf:resource="&om;Scale"/> 
    <owl:equivalentClass>
      <owl:Class>  <!-- Equivalent to the class whose only member is FahrenheitScale -->
	     <owl:oneOf rdf:parseType="Collection">
            <om:IntervalScale rdf:about="&om;FahrenheitScale"/>
         </owl:oneOf>
      </owl:Class>
    </owl:equivalentClass>
  </owl:Class>

Warning! That's just hand-written XML. No guarantee it will even parse!

HajoRijgersberg commented

Thanx, that was a very stupid mistake of mine. I'm going to look at your pull request right now!