10 Comments
User's avatar
Tiankai Feng's avatar

What an insightful article by two awesome people, and I love the diversity of the examples!

Jessica Talisman's avatar

Thank you, Tiankai!

Ed Wolfe's avatar

Ontologies: For ONCE the term ‘analytics’ is used to mean something other than a ‘sexy’ catchphrase!

Some ‘tidbits’ of mine:

‘To use an analogy, a data model is to the data engineer or analyst what a compass and map are to a sailor.’

‘keys and attributes’: This is where my fascination with databases blends in well with my fascination with (I’ll call it) ‘Bill Inmon stuff’.

‘In many ways the ontology is to text what the data model is to structured data.’

‘“the manager had to fire the employee after the office fire,” …so many times I was that “employee” who caused the “office fire.” Please notice the difference between “...office fire,” and “office fire.”.’ Correct English?

‘In an increasingly complex information landscape—where text, structured data, images, audio, and sensor data all contribute to enterprise knowledge—the ontology provides the semantic foundation that makes integration possible. It’s not metaphysical. It’s not mystical. It’s the essential infrastructure for organizations that want their knowledge to be findable, understandable, and usable across systems and over time.’ ‘Ahh…that’s the ticket’, as Studs Terkel would say. . . .

And so what is the difference ‘tween ‘database’ and ‘data base’?

Jessica Talisman's avatar

Love the Studs Turkel reference 😉

Ed Wolfe's avatar

I used to listen to him a lot.

Carl Widigsson's avatar

Just love this post! Humble title since it covers almost ALL perspectives :-)

Torsten's avatar

As a long-time follower of your work, I you point setting the line between the internal-facing Data Model and the external-facing Ontology.

Regarding the 'context engine' and Jessica’s point on semantic grounding: Do you feel we finally see more people realizing that they need to work on their business terms and actually manage their most important asset (aka data)? Not only to fix hallucinations, but also to properly manage their IT in general I would claim. I'm curious if you are seeing a real shift in adoption where companies are prioritizing the ontology before the AI implementation this time.

Matthew Mullins's avatar

lol "philosophy’s appropriation of the term" ontology, I think someone has confused the order of appropriation! The term was coined by a pair of German philosopher's in the 17th century and was only appropriated into information science in the 20th century. Anyway, enough pedantry for the day

Jessica Talisman's avatar

Thank you, Matthew. Appropriation does not imply order. The term ontology is used in both disciplines (philosophy and information science), but with different meanings. In linguistics, this is called a homonym, and happens throughout the English language.