20170911 Ontology Change Improvement Call

Date: 11 Sept 2017

Attendees: Marijane White, Graham Triggs, Muhammad Javed

Agenda:

• Building a single vivo ontology file out of Mike's filegraph.owl

Marijane: Mike is probably not joining us today because of Hurricane Irma.

Javed: I looked at the filegraph a little bit, in two ways.

1. What's declared, and what's not declared.
2. What is declared but defined differently in two files.

see notes at : (starting on page two for Javed's notes) https://goo.gl/pxAVH1

Equivalent class definition differences: https://drive.google.com/open?id=1zNmugbuGYEk-3qUTRlpqi_r8Vv7GWknGuopvClDzR20

Question for Graham, where are the Vitro annotations used? In the TBox, when defining something, or...?

Graham: I think some of the things about link suppression are used in the interface, defining what level you can do certain things at.

Javed: This reminds me of something we're doing now at Cornell. We're using Symplectic Elements at the backend and VIVO on the frontend. There are things like this in Elements, and we wanted to push it to VIVO as well, so publications can be ranked/sorted, users can set their favorite. We kept those kind of properties, which are very specific to the application and representation, not the model, in a different file, so we can keep them separate.

Graham: The Vitro properties are about the application, how things are displayed, what people can do with them. But some of the things you are bringing over from Elements, and that you use to affect how things are displayed in the interface, I think there's a valid argument that these things should be part of the model. A researcher expressing that some set of publications are their favorites, or publications they don't wish to have highlighted, that is actually useful knowledge, it's not just about the application, there is an argument for storing that part of the model.

Javed: Yes, you're right about that.

Marijane: I would agree. So do you have more work left on this, Javed?

Javed: Yes, it's not done yet. There are more notes to be made on the axioms, and more color highlighting by type, to help us navigate. I wanted to put everything in one place so we can all have a look and see the differences, and then understand what to do about it. But of course, the first step is to create a single file, and then fix that.

So I don't understand how the files became different. An example from the translator property: ObjectPropertyDomain(http://purl.org/ontology/bibo/translator ObjectUnionOf(http://purl.org/ontology/bibo/Collection http://purl.org/ontology/bibo/Document))

Graham: so with the differences, we don't know how they got there?

Javed: no, I found them, let me open them up and show you with screen sharing.

• shows some examples of different definitions between filegraph.owl 1.6 *

Marijane: This is making me curious about what's in the source.owl, in the ISF repository, and whether we can treat that as the source of truth. Also, the person who seems to know the most about how 1.6 came to be is Brian Lowe, so I wonder if he might have any insight into these differences.

At the very least, it's good to have found these issues.

Javed: we should have a sprint planning meeting and just decide what to do with them, one by one.

So maybe I'll continue working on that, and maybe by the next meeting we can start looking into these.