An ontology is a formal specification of what exists. It consists of an enumeration of entities and relations between them.

When you study mathematics, you familiarize yourself with a graph of axioms and theorems, connected by implication. There is a high level overview of mathematics to be had by looking at this graph.

When you study botany, you learn the cladogram of plants and fungi. You learn relations of evolutionary succession, distinguishing criteria for different branches on the tree.

When you study physics, you learn the pictures of the standard model. A story about particles and their fields and the principle of stationary action as a generator of the PDE's that govern the evolution of a system.

Every time someone writes some instructional material on a topic in the sciences, there's a way to connect it to a universal ontology. The human instructional artifact is related to various physical entities and scientific theories in a straightforward way. Excerpts from the text are explanations of individual phenomena or describing relations between objects of the some model.

The value here is not just in standardizing and universalizing (if it is at all), but of using a model of knowledge which is deeper than individual tables of content, of disparate curriculla, study guides, blog posts, and course sequences.

Universalizing all knowledge into a single model probably isn't that important. The Semantic Web is an academic area developing a protocol for interoperable ontologies on the internet which more or less culminates in the Web Ontology Language (OWL). OWL has never become very widespread, except perhaps in biology where its fascillitates standardized taxonomies, and the data collected which depends on that vocabulary.

Current visualizations of the semantic web are often just tangled balls of yarn. We can do better. The important thing is modeling the underlying domain, and spatializing the relations with intuitive representations. The cladogram, the timeline, these are terrific representations. They work well because they have the possibility of representing anything in a given class. They bound entities with potential mereology. A bad diagram is an indefinite list. An exhaustive list is useful though.

I'm tempted to try and make visualizations, but I'm not sure if that's what is needed. I want to be able to browse the web connecting what I see to the knowledge base, so that I have some fucking purpose, so that I'm not drowning in myself in drivel.