I an exploring making a formal ontology in an Obsidian vault.
I have to abide by some very strict rules to keep the conceptual organization clear.
Right now, I am requiring that a note corresponds to a “record”, like in a database, and a record should be an entity of some kind, like a specific instance of some class or type. So, you can have a note that is the representation of say, your teacher Ms. Robinson, but less so a note that is just you writing random thoughts on things.
A thing’s properties are given by the YAML frontmatter.
Like in SQL, RDF, and other relational data formats, an entity can have a data property or an object property. A data property is a literal value of one of a few established data types. In Obsidian, there are six types available: checkbox, date, date and time, number, list, and text.
An object property has another record as its value. In Obsidian, this can be achieved by linking to that record using the [[Wikilink]] syntax.
This means that you should choose your filenames carefully, as the become the canonical identifier for that resource / record.
I think it would be cool if I could set up the following.
Right now, I have a file called my email address. I have a text property called instance of with value [[email address]]. That file has property instance of type. That means that I have a type type, and email address is a specific type - so it is an instance of the type type.
In email address, I created a list property called has property. It includes the values [[domain (internet)]] and [[local part (email)]].
Ideally, Obsidan should understand that any note that is instance of email address should automatically have the property domain (internet) and local part (email). I can even make notes for those two entities, like domain (internet) instance of property, has range text, and so on.
I don’t know if you can run triggers that get run when you add a value to instance of automatically or something?
