Re: ssn ready for review

Hi Kerry and all,

It’s a big progress and nice work. Here are my comments after first round of review, I’m happy to discuss more and take some actions as pull requests on some of the points in the comments.


/sdw/ssn/rdf/sosa.ttl



1.      There is a language tagging for rdfs:label but not for skos:definition and rdfs:comment, is it intentionally?

2.      For sosa:hasValue definition in bellow,  it does not have meta:domainIncludes and meta:rangeIncludes like other properties. And also, it is not included in the diagram at sore-core-overview.pdf. Is there a reason for this?
sosa:hasValue
  rdf:type owl:DatatypeProperty ;
  rdfs:comment "The value of a Result, e.g., 23 or true." ;
  rdfs:label "has value"@en ;
  skos:definition "The value of a Result" ;
  skos:example "23 or true" ;


3.      There is a typo in sosa:Sensor definition
sosa:Sensor
  …
  skos:definition "Device, agent (including humans), or software (simulation) involved in, or implementing, a (Sensing) Procedure. Sensors responds to a stimulus, e.g., a change in the environment, and generate a Result. Sensors can be mounted on Platforms, e.g., a modern smart phone hosts multiple sensors."@en ;

/sdw/ssn/index.html


1.      In the abstract, we mention that  “SSN …is an OWL-2 DL ontology…” and based on OWL 2 Primer[1], section 9, OWL 2 DL is one way of assigning meaning to an OWL2 ontology that use OWL Direct semantics[2]. The rest of the document hasn’t touched on this aspect, so, I wonder if we consider the characteristics/features of OWL 2 DL is a common knowledge in the document. To me, it’s not trivial to say SSN is OWL 2 DL or OWL 2 Full from reading the SSN axioms together with the specs in [1] and [2].

Also, we have four modules in the specification that uses different OWL 2 and RDFS axioms. For example, SOSA uses RDF axioms and owl:inverseOf  which can categorized as RDFS++ ontology which can be efficiently reasoned using some commercial triple store such as Virtuoso or AllegroGraph supporting RDFS reasoning together OWL axioms relevant to following properties:

·         owl:sameAs

·         owl:inverseOf

·         owl:TransitiveProperty

Based on OWL 2 profile specification[3], there are 3 OWL profiles, i.e., OWL 2 EL, OWL 2 QL and OWL 2 RL for  categorizing  an ontology based on the axioms used on that ontology. I think we could  clarify which OWL profiles that  SSN, SSN-O&M and Docle-light alignment modules fall into, then, the complexity of reasoning over data annotated by these modules will be clear.


2.      Currently, SSN module (Section 4) is put before SOSA module (section 5), I think it’s less confusing if SOSA comes first.

3.      In SOSA spec of Section 5,  only OWL Class appears along with the class descriptions but in the file SOSA.ttl, we also added RDFS Class

4.      There is a typo in “Vertical Segmentation” section:  Vertical modules are build….

[1] https://www.w3.org/TR/owl2-primer/

[2] https://www.w3.org/TR/2012/REC-owl2-direct-semantics-20121211/

[3] https://www.w3.org/TR/owl2-profiles/


From: Kerry Taylor <kerry.taylor@anu.edu.au>
Date: Monday 12 December 2016 at 14:06
To: "public-sdw-wg@w3.org" <public-sdw-wg@w3.org>
Subject: ssn ready for review
Resent-From: <public-sdw-wg@w3.org>
Resent-Date: Monday 12 December 2016 at 18:44

To the best of my knowledge ssn is now stable and awaits your review prior to the vote to publish a fresh working draft  at the F2F.  In the last few days there
Has been work on tidying up issue-105 and the changes section (myself) , extending the section on modularity and sosa by Krzysztof, and the automated description of sosa together with relevant issue documentation by Armin.

Please have a look!
-Kerry