Closed igrangel closed 6 years ago
Technical specification is different from a standard in the fact that it has not been passed through all the approval stages to be called a standard.
Yes, agree, but how can this fact be represented in the ontology in terms of properties and axioms? For example, are the classes disjoint (owl:disjointWith), which properties they share and which they do not?
sto:TechnicalSpecification owl:disjointWith sto:Standard;
Can you provide a reason for supporting this design decision?
Technical Specifications have the same attributes as a Standard does. In the IEC Webstore, one needs to look carefully what the publication type is otherwise you could easily not notice that it is a Technical Specification. In real world, this means that the Technical Specifications are yet to go through various approval stages to be called a Standard. The decision to introduce sto:TechnicalSpecification owl:disjointWith sto:Standard; relationship was to clarify the difference against the visual similarity in the way they are both represented in Webstore.
I agree that they are sufficiantly different, nevertheless sto:TechnicalSpecification owl:disjointWith sto:Standard;
is dangerous. Let's say:
A a sto:TechnicalSpecification . A sto:hasClassification x .
leads to:
owl:Thing equals owl:Nothing .
because:
sto:hasClassification rdfs:domain sto:Standard .
I am not saying that we do not need the separation. But the properties are not (yet) modeled accordingly.
A connected topic is the treatments of IETF RFCs (request for comment). While being a proposal in the original sense, RFCs are defacto standards. How do we want to model them? example
sto:RFC_1122 rdf:type sto:Standard ;
or
sto:RFC_1122 rdf:type sto:TechnicalSpecification;
About the sto:hasClassification rdfs:domain sto:Standard We initially worked like this. The Properties we created had sto:Standard as its domain. But once we created the class sto:Publication, this was an intermediate logical level between foaf:Document and sto:Standard. sto:TechnicalSpecification and sto:Standard are both a sto:Publication type.
and the Domains of properties changed to sto:Publication, as in the example sto:PublicationDate or sto:hasPublisher.
Please check if the Domain of sto:hasClassification can be changed to sto:Publication and it should solve the issue.
About the RFCs. We could model as follows
sto:RFC_1122 rdf:type sto:Publication or sto:RFC_1122 rdf:type foaf:Document
Since they are neither Standards nor Technical Specification. They are a different Kind of Document. But still specifies information similar to standards. I would prefer for it to be sto:Publication but, sto:Publication as defined by IEC-Publication which is our reference ,talks only about International Standards,Technical Specifications,Publicly Available Specifications,Amendments,Technical Corrigenda, Interpretation Sheets,Technical Reports,Guides but not RFCs. I guess we could modify our sto:Publication to include RFCs and model them as Publications
Please have a look at this push. We distinguish Standards from specifications like this.
In my view, RFCs fit into the definition of sto:Publication . You are right, IEC does not directly use the term 'RFC' but the further explanation perfectly fit. Therefore, let's put it one hierarchy level up and we are fine.
If we do so, we definitly need to adjust the domain statements as you suggested in your comment
And this will be our Technical Report. Another kind of Publication/Document
If we agree on this, we will make sure to update the definition of what a Publication means to us and make all the new kinds of Standard documents that we encounter further as a Publication type.
What makes then similar and what different, in terms of OWL properties and axioms, from Standards and Technical Specifications?