20180501 Ontology Change Improvement Call

2018.05.01

Attendees: Marijane White, Muhammad Javed, Ralph O'Flinn, Mike Conlon, Damaris Murry, Juliane Schneider

Agenda: https://wiki.duraspace.org/display/VIVO/2018-05-01+Ontology+Improvement+Call

Ralph wondered if Duraspace is going to pay for Slack now that we're nearing our 10,000 message free-slack limit. * He suggested setting up spaces at UAB's Microsoft Teams.
* Marijane suggested Gitter, but Ralph raised concerns about new GDPR regulations and the fact that Gitter has not been recently updated.

Sprint 1 Ontology work update:

Javed reports that we are almost done with what we've set out to do.

• Tatiana finished her VCARD evaluation and it is waiting for Javed and Marijane to review it.
• Javed has also completed work on the ontology modules analysis.
  • He sent out a message to vivo-community to see if anyone is currently using the OCRe study design/clinical trial entities to see if they could be removed from the core VIVO ontology.
  • Ralph notes that UAB has clinical trials, which means perhaps only the top-level class remains in core, and the rest would be moved into a module.
  • Mike believes we should be able to add terms to vivo.owl from the ISF, which might be a replacement for terms from OCRe.
  • Javed says the issue is not replacement -- the concept is coming from eagle-i and it is currently in vivo.owl.
  • Mike notes that the "has phase" relationship seems wrong, because clinical trials themselves are phases, they don't have them. We need to analyze what we can say now, what do we want to be able to say, and how do we want to be able to say it/what is the best way to say it?
  • The VIVO Scientific Research namespace is misnamed, it's all about Clinical Research

Continuing discussion of modules:

• Mike notes that eagle-i goes into much more detail than VIVO needs.
• Melissa would like to be able to attach specific concepts to publications, like organisms.
• Being able to do this would require loading relevant ontologies, but many of them are very large, and it would be better if they were a service, like VIVO does with FACT. So sometimes eagle-i is a module, sometimes it's a service.
• Javed thinks we should be able to provide that link, the "about" connection between the publication and the thing.
• Regarding Humanities concepts, Mike isn't sure if there's a module here, because VIVO needs to be able to talk about all the scholarship of all the faculty, including things like directing symphony performances. There is likely a level of ontology for Humanities that is analogous to the level of ontology about the sciences. We want to treat the very detailed concepts like the clinical trial concepts in OCRe.
• Damaris notes that it's not even a Humanities issue anymore, everyone is creating videos and the like these days. Mike agrees.
• Mike notes that we need to figure out how to do some homework around this, particularly around finding work already done by other groups.
• Javed notes that LD4L has done a lot of work in this area, but that librarians tend to model things differently than VIVO does.
• Mike notes that rather than a document-oriented view, BFO is on what happened. (Realist ontology)
• Javed notes that librarians are focused on works, manifestations, etc. (FRBR?)
• Mike thinks we need to have a workshop, with everyone sharing the work they already have beforehand to set the stage for a conversation at the workshop.

Conversation about cleaning and Unconference sessions at the conference:

• Javed views this as being about removing obsolete concepts
• Mike notes that the VCARD cleanup is very tactical and it would be good to expose people to those issues, it would be good to have a session about that.
• If we also had scientific research and humanities sessions, we're covering all the bases
• Javed suggests we have one ontology session that gets subdivided
• Javed thinks we should do some homework first
• Mike says we can send out a call for work in the humanities to get people to start sharing it
• Damaris says Duke refers to the Getty AAT a lot: http://www.getty.edu/vow/AATHierarchy?find=dance&logic=AND&note=&page=1&subjectid=300054137
• Mike heard about a Humanities ontology at OR and will dig it up. "Humanities" may not be the right terminology.

Javed has some comments about inference:

• Not sure if any institution or the software is using the upper level ontology classes. Which raises questions:
  • If no one is using them, why are they part of the ontology?
  • And why are we generating all the triples for them?
• Mike says he would have asked this exact question in 2013-2015. Had a conversation with Shahim Essaid about it. From a practical POV, Javed is correct, we don't do anything with them and we don't need them. On the other hand, it is 100% important that as ontologists, that we understand what these things are, and it is critically important. These are two completely different ideas. It is definitely possible to consider not having all these inferred classes in the software, and this is something talked about at VIVO camp. There are actually some VIVO classes that need to be hooked into this hierarchy, and some that are even wrong, like Relationships and Academic Degrees. But none of this needs to end up as assertions in the triplestore.
• Marijane notes that one way to deal with this would be to do dynamic inference rather than static inference. Mike says that would be a good conversation to have with Brian Lowe -- what inferencers are available, etc.