# Meeting Notes - 19.11.2021

**Participants**:
- [X] Casper W. Andersen
- [X] Rickard Armiento
- [ ] Francesca L. Bleken
- [ ] Matthias Büschelberger
- [ ] Stuart Chalk
- [X] Kathrin Frei
- [X] Jesper Friis
- [X] Emanuele Ghedini
- [ ] Luca Ghiringhelli
- [ ] Gerhard Goldbeck
- [X] Saulius Grazulis
- [ ] James Hester
- [ ] Joana F. Morgado
- [X] Antanas Vaitkus

## CIF ontology generation tool

Emanuele: What is the status of the CIF ontology generation tool?
Casper: The tool itself should be finished(TM), but to me the main hurdle is how we ensure proper versioning when publishing and securing an IRI for the ontology.

Casper suggests to solve the versioning issue by publishing the generated ontologies in a dedicated branch in the GitHub repository, where they will be placed in a folder; the folder name being the ontology version.

Jesper: If we want to use the EMMO domain for the permanent IRIs, the redirect script needs to be updated, which redirects the permanent IRIs (URLs) to "raw" GitHub files.
Casper suggests to make the branch name a variable in the script, instead of hardcoding it to use `master`.

## Procedure for extending the crystallography domain ontology

Last time Emanuele suggested the following approach:

> Find 2-3 fundamental definitions from the dictionary of IUCr. **Unit cell, crystal, crystalline**. Discuss them and try to place them in the ontology. The ontology here being the EMMO that Emanuele showed parts of in the last meeting, where we tried to place “Crystal”.

The work so far is "published" in the EMMO version 1.0.0-beta2.

### Extending the domain ontology

We initially chose to place *Unit cell*.
This led to a longer discussion, first about where to start out a "Crystallographical/Crystallography" "umbrella"-concept. We decided to place it under the `Language` concept.

Then we discussed the role of the generated CIF ontology and `DictionaryDefinedItem`s.
Emanuele pointed out the meaning of the separation of parts in the CIF ontology, and that the `DictionaryDefinedItem`s should be placed under the "Crystallographical/Crystallography" `Language`.
This led to a long discussion about why a concept like `(...)symop.id` should be included, since the domain experts (Vaitkus, Rickard, Casper) vocally were against considering such a concept having any crystallographic semantic meaning.
In the end, it came down to a misunderstanding: Emanuele's idea was to include the `DictionaryDefinedItem`s under a `CIF` or `CORE_CIF` concept under "Crystallographical/Crystallography". I.e., it would become the semantical CIF way of representing crystallography. And indeed, in this perspective, the format-specific concepts have a right.

The intention is to have manually created crystallographical concepts (like `UnitCell`) exist directly under "Crystallographical/Crystallography", being at the same level as `CIF` or `CORE_CIF` and the likes.
Then we eventually need to match these manually created concepts up with the auto-generated concepts from the CIF dictionaries.

## Homework

- Casper, Jesper, and Emanuele will take a look at the EMMO URL (IRI) redirections to determine how to best setup ontology publication of the CIF ontology - if we are to use the `emmo.info` domain.

- Casper and Jesper will try to outline the intended connection between the CIF ontology and the EMMO according to what was suggested at today's meeting.