align SSN:Observation to OML:observation

[ramonawalls]

Should they be equivalent classes?

[dr-shorthair]

Nope. ssn:Observation is an information object that describes an observation. oml:Observation is an observation event or activity.

The original derivation of SSN from DOLCE hinted at this issue, and [1] confirms my interpretation using PROV core classes. In fact, Compton et al. needed to introduce a new class 'ActivityOfSensing', which looks like it corresponds with oml:Observation. Here's the PROV alignments (red arrows=disjoint classes):

and my current interpretation of the implications for ssn-oml alignment:

:ActivityOfSensing owl:equivalentClass oml:Observation . ssn:featureOfInterest owl:equivalentProperty oml:featureOfInterest . ssn:observationResult owl:equivalentProperty oml:result . ssn:observationResultTime owl:equivalentProperty oml:resultTime . ssn:observedProperty owl:equivalentProperty oml:observedProperty . ssn:Sensing rdfs:subClassOf oml:Process . ssn:Sensor rdfs:subClassOf oml:Process . ssn:observedBy rdfs:subPropertyOf oml:procedure . ssn:sensingMethodUsed rdfs:subPropertyOf oml:procedure .

[1] M. Compton, D. Corsar, K. Taylor, Sensor Data Provenance: SSNO and PROV-O Together at Last, in: 7th Int. Work. Semant. Sens. Networks, 2014: p. 16. http://knoesis.org/ssn2014/paper_9.pdf (accessed April 2, 2015).

[dr-shorthair]

(BTW - am currently revising the OML paper to include this material.)

[ramonawalls]

Thanks for clarifying, Simon. We will still need to decide whether or not we include SSN in our mapping. I suggest that the main focus of the mapping (e.g., the figure) be OBOE, OML, and BCO, and that we provide other mappings as auxiliary findings (because the authors of SSN and PROV were not part of the mapping process).

[wernerkuhn]

It would seem an omission to me if SSN was not part of this effort. The fact that no direct authors were present at our workshop was unfortunate and should not affect the overall alignment effort. Having done the original research underlying SSN, I will try my best to account for it, pulling in Jano and others where appropriate. 

--  Werner 

[ramonawalls]

Thanks, Werner. Sorry for spacing out on that. Yes, it does indeed make sense to include it in that case.

[wernerkuhn]

Thanks, Ramona, could you point me to the evolving paper draft again please?

--  Werner 

[ramonawalls]

Looks like you found it!

https://docs.google.com/document/d/1Yn1RscG1_EOKenMs3QCYDfNVTU2LTHZryOs57_VMxbo/edit

[wernerkuhn]

Ah, I was not sure whether it had moved elsewhere, thanks!

[dr-shorthair]

Note that the figure above is from [2](revised and re-submitted to Semantic Web Journal), where the SSN/PROV/OML alignment is presented at a little more depth.

[2] Cox, SJD "Ontology for observations and sampling features, with alignments to existing models" Submitted to Semantic Web Journal, http://www.semantic-web-journal.net/content/ontology-observations-and-sampling-features-alignments-existing-models

[dr-shorthair]

An updated version of [2] is now on the Semantic Web Journal site. No significant differences in the paper for our purposes.

http://semantic-web-journal.net/content/ontology-observations-and-sampling-features-alignments-existing-models-0

[dr-shorthair]

Semantic Web Journal paper officially accepted for publication (2015-12-12) http://semantic-web-journal.net/content/ontology-observations-and-sampling-features-alignments-existing-models-0

[dr-shorthair]

Also see this presentation from AGU: http://www.slideshare.net/drshorthair/pitfalls-in-alignment-of-observation-models-resolved-using-prov-as-an-upper-ontology

[ramonawalls]

Great Simon! Glad to see it published.