search

w3c / sdw-sosa-ssn

Repository of the Spatial Data on the Web Working Group for the SOSA/SSN vocabulary
7 stars 5 forks source link

Update OBOE-SOSA mapping to use ObservationCollection #153

Closed dr-shorthair closed 5 months ago

dr-shorthair commented 6 months ago

Closes #151

ldesousa commented 5 months ago

There is something I don't understand (therefore not blocking the merge): why are oboe-core:Characteristic and oboe-core:Measurement mapped with owl:equivalentClass, while oboe-core:Observation and oboe-core:ObservationCollection are mapped with rdfs:subClassOf? I understand the concept of the latter two being specialisations of sosa:ObservationCollection, I just wonder if there are other implications in using different predicates for the alignment.

dr-shorthair commented 5 months ago

Thanks for looking at this carefully @ldesousa .

The difference in the mappings was carefully considered.

The mappings are best understood using set theory. owl:equivalentClass means that the class extension (membership) is the same [1], while rdfs:subClassOf means that the class extension is a subset. Don't confuse owl:equivalentClass with owl:sameAs as the latter implies that actual graphs can be merged, which is a different thing.

OBOE provides a general model for observations, very comparable to the SOSA Observation and ObservationCollection structures.

Looking at OBOE and SOSA, every individual oboe-core:Measurement could be expressed as an individual sosa:Observation, and vice versa. Ditto for oboe-core:Characteristic and sosa:ObservableProperty. Hence oboe-core:Characteristic and oboe-core:Measurement are equivalentClass to sosa:ObservableProperty and sosa:Observation, respectively.

In contrast, two classes from OBOE (oboe-core:Observation and oboe-core:ObservationCollection) are implemented by only one class from SOSA (sosa:ObservationCollection), with some restrictions on collection membership. i.e. while every oboe-core:Observation is also a sosa:ObservationCollection, and every oboe-core:ObservationCollection is also a sosa:ObservationCollection, only some sosa:ObservationCollections are oboe-core:ObservationCollections and only some sosa:ObservationCollections are oboe-core:Observations, and some are neither. Furthermore, the two OBOE classes are disjoint, ie there is no intersection between these. Thus the mapping is a subclass relationship.

Finally, note that this mapping is labeled 'non-normative'.

[1] https://www.w3.org/TR/owl-ref/#equivalentClass-def

dr-shorthair commented 5 months ago

Could one more person review this PR so the changes can be merged

dr-shorthair commented 5 months ago

Note that this is a straightforward alignment, originally documented in SSN-ext. I've just re-used the figures from there.

It was one of the things that led to the ObservationCollection proposal.

Need one more review to merge.

dr-shorthair commented 5 months ago

As usual, the browser view for this branch can be seen at https://raw.githack.com/w3c/sdw-sosa-ssn/OBOE-mapping/ssn/index.html