EntityAggregate?

[dr-shorthair]

object aggregate (bfo:0000027) appears to be the class for aggregates or collections of things. However, it is limited to material entities.
Various has * part and is * part of allow for part-whole aggregations, but AFAICT except for object aggregate the 'whole' is conceived as a unitary thing.

So how does one represent a collection of 'planned acts', for example a set of observations or measurements? They do not compose a single planned act. In fact they may be assembled after the fact.

I've looked at #91 and #365 but can't see a way forward for non-ICE and non-material-entities.

[alanruttenberg]

For planned act you can make an Act and assert 'has occurrent part' <planned act 1>, <planned act 2>, ....
You can name it in a class expression as (Act and 'has occurrent part' min 2 planned act). If you want to tighten it you can also add and ('has occurrent part' only 'planned act' or ('occurrent part of' some 'planned act')).
You can SPARQL query for planned act parts.
Despite CCO's having a collection of 'aggregate' classes and relations, I don't think they are necessary or desirable. There are too many things that can aggregate and having them all would bloat the ontology. We can always compose these aggregates using the appropriate part relation.

But perhaps prove me wrong with a compelling example of why the above solution won't work.

[dr-shorthair]

My hesitation is that I don't see an arbitrary collection of observations, perhaps aggregated after the fact, as a single 'Act'.

I'm happy to define such an aggregate in an application ontology, but I thought perhaps a single generic entity aggregate, with perhaps the two key specializations continuant aggregate and occurrent aggregate might provide a useful base pattern.

(In SSN we have ExecutionCollection and SampleCollection as key cases.)

[alanruttenberg]

Not to my taste. We have parthood relations - use them. Others may have different opinions.

[arazdow]

Maybe it's helpful to distinguish among acts of aggregation. At least three examples come to mind:

• Extrinsic aggregates are made by reference to the things being aggregated.
• Equivalence classes that aggregate on properties of things are intrinsically aggregated.
• Objects connected by transitive connections between them (like parts of a machine) are aggregated by association.

These strike me as very different ontologically, and perhaps what drives our variations in taste and opinion on aggregates.

[alanruttenberg]

@dr-shorthair

My hesitation is that I don't see an arbitrary collection of observations, perhaps aggregated after the fact, as a single 'Act'.

Then define it as process that has those parts. Sums of processes are processes.

[dr-shorthair]

OK, I can accept that.
But then how is 'object aggregate' special?