search

BiodiversityOntologies / bco

Biological Collections Ontology
Creative Commons Zero v1.0 Universal
21 stars 3 forks source link

Properties of oberving process - SOSA alignment #93

Open dr-shorthair opened 5 years ago

dr-shorthair commented 5 years ago

To enable the transport of data between systems, could we check that BCO supports (i.e. has equivalents to) the essential properties identified in the W3C SOSA/SSN ontology:

I expect some of these might correspond with property-chains in OBO, in which case we might consider naming the shortcuts too.

To help orient, the alignment of SOSA to OBOE is at https://www.w3.org/TR/vocab-ssn/#OBOE_Alignment

ramonawalls commented 5 years ago

@dr-shorthair, the alignment that came out of our meeting in Santa Barbara includes BCO, OBOE, and SSNO. It probably needs some tweaking to align to SOSA, but it will give you an idea of the relationships. Also, the OBI model of planned processes has evolved/solidified since that meet, and my understanding of it has improved, so I think I could do a better job with this mapping now.

ramonawalls commented 5 years ago

Here is an image that summarizes the relations linked to OBI's 'planned process'.

OBI_planned_process.pdf

ramonawalls commented 5 years ago

We don't yet have a set way to distinguish between the ultimate feature of interest and the direct input, but I think we reached some agreement about how to do this on our last call.

dr-shorthair commented 5 years ago

So you are pointing at the more generic properties of a Planned Process.
It looks like there would be an approximate mapping something like

sosa:hasFeatureOfInterest sub-property-of hasParticipant? or hasSpecifiedInput?
sosa:observedProperty sub-property-of achievesSpecifiedObjective
sosa:usedProcedure equivalent realizes
sosa:madeBySensor sub-property-of hasParticipant?
sosa:hasResult ~ hasSpecifiedOutput
sosa:isSampleOf ~ ??
dr-shorthair commented 5 years ago

Ping @ramonawalls @robgur @pbuttigieg - am struggling with alignments to RO. Any insights?

cmungall commented 5 years ago

Note that by using logical axioms (subP/equivP) to RO you will induce ontological commitments upon the sosa classes, which may not be intentional. E.g. usedProcedure will induce the range to be a bfo Realizable, which will conflict with someone using procedure as a Process or information entity.

Sorry if this is obvious, don't have the overall context of the project - are logical axioms mapping sosa classes to BFO/OBO Classes/CommonCore available?

dr-shorthair commented 5 years ago

Thanks Chris - this is indeed the kind of guidance I am looking for.

I find the definitions and documentation of RO a bit difficult and incomplete, even though it clearly is the key module in the whole shebang. So am really just looking for some suggestions/orientation from more experienced hands at this stage. I've only highlighted 6 key properties in the table above, and my hunch is that someone who already knows RO (yourself? Ramona?) could likely knock those off in at least provisional fashion in < 1/2 hour, while it would take me many multiples of that and I'd likely still miss something. And I would definitely keep these alignments in a very separate graph so the commitments would only affect those who chose to load them.