Skip to content

Meeting Notes 26.03.2021

Jump to bottom
Casper Welzel Andersen edited this page Apr 9, 2021 · 1 revision

Participants:

  • Jesper Friis
  • Casper Andersen
  • Saulius Grazulis
  • Francesca Bleken
  • Emanuele Ghedini
  • James Hester
  • Rickard Armiento
  • Joana Morgado

Ontologizing CIF dictionaries

Presenting the newest version of the CIF ontologization.

Add further elucidation of CATEGORY that instances of it are categorized objects. Emanuele: While we have an isA relation between SPACE_GROUP_SYMOP and CATEGORY, this is not true. Rather, individuals of SPACE_GROUP_SYMOP are CIF categorized objects. It’s a higher order relation than what is represented currently in the diagram.

Jesper: 1 should be an individual of _space_group_symop.id, being a DATA_ITEM and a type. We need to specialize the current CIF files with the new CIFv2 types. James: This is doable, as the new types are equivalent to the older types. Hence, CIFv1 files should be interpretable using the CIFv2 dictionary (and ontology).

Jesper has developed a Python script to import CIF dictionaries, converting them to Turtle with the ontological relations as schematized by the v4 of the CIF ontology schema.

James: We shouldn’t lose sight of the reason we are ontologizing the CIF dictionaries; to be able to parse the CIF data content with high fidelity, while keeping and understanding the inter-relations. I.e., there is no need to implement all and every CIF quirk or ontological meta-naming.

Jesper: The essence now is to implement the CIFv2 types, to make all DATA_ITEMs relate to one of these specific types.

Saulius: We should be careful not to mix in syntactical definitions with the semantic concepts.

The DATA_ITEM is a confusing concept name and it’s more understandable to rename it to DATA_VALUE. This also refers directly to its sub-classes being both a type and DATA_VALUE.

There’s a bit confusion about ROW versus COLUMN. Mostly whether they are semantically different, and due to the naming being very close to syntactical concepts, it is difficult to understand the semantical difference.

The rest of this meeting went on to discuss the finer points of Emanuele's earlier statement that a SPACE_GROUP_SYMOP is not a CATEGORY and the implications of this idea.

Clone this wiki locally