Additional Axiom in OWL saying that first page of a collection is part of the collection itself

[cristianolongoodhl]

Pull request #689

We report the definitions of the two properties partOf and first from the Activity Vocabulary

first

In a paged Collection, indicates the furthest preceeding page of items in the collection.

partOf

Identifies the Collection to which a CollectionPage objects items belong.

In other words, every collection page (say p) is connected to the collection it belongs to (say c) via the partOf property and, conversely, a collection c is linked to its first page trough the first property.

Here we propose to extends the OWL ontology with the additional axiom first ⊑ partOf ^-, stating that the first page of a collection is part of the collection itself. For example, it would allow to derive p partOf c from c first p (the dashed edge in the following graphical representation).

[evanp]

Would this apply to other properties of the Collection that indicate a page - last, current are the two that come to mind most easily.

I'm also a little concerned about verification -- I like when one object declares a relationship like first to be able to verify it by finding a partOf in the other object that has the first object as a value. Would this axiom obviate that kind of verification?

[evanp]

Would this change the text of first or other text in the activity vocabulary? Or is this just a change to the OWL file?

[evanp]

@cristianolongoodhl would you mind doing a PR for this?

[cristianolongoodhl]

Would this apply to other properties of the Collection that indicate a page - last, current are the two that come to mind most easily.

Yes, I have yet prepared two subsequent pull requests for both last and current.

I'm also a little concerned about verification -- I like when one object declares a relationship like first to be able to verify it by finding a partOf in the other object that has the first object as a value. Would this axiom obviate that kind of verification?

No, it makes it implicit. If you write c first p one can assume p partOf c.

@cristianolongoodhl would you mind doing a PR for this?

#689

[TallTed]

No, it makes it implicit. If you write c first p one can assume p partOf c.

But if the explicit statement is not made, then reasoning that infers the partof relationships based on the first relationships MUST be active in order to deliver the described verification.

I'm a big fan of active reasoning, inferring on the fly, but assuming that such is present and active is a bad plan, in the overall RDF world.

[cristianolongoodhl]

@TallTed reasoning for an online service is not feasible due to the huge amount of resources needed. And, reasoning over a distributed knowledge base is a very complex matter. I think that the definitions in the OWL version of activity streams should be used offline, at first to clarify usage and semantics of activity streams items. In other words, I hope we will be able to encode in a precise and machine readable language the specifications provided in natural language.