I like problems that resist the first description.

Practice

I tend to make the question operational. If it concerns a model, I want to inspect its errors. If it concerns an interface, I want to watch somebody use it. If it concerns an institution or category, I want to see what happens at the edge cases.

That produces a mixed bag: web applications, data notebooks, spatial analysis, simulations, diagrams, essays, and tools that began as a way to stop doing some tedious thing by hand.

Interests

Data and software are recurring materials, not the whole subject. I am also interested in how models carry judgment, how systems acquire obligations, how people learn a rule, and what a useful explanation owes its reader.

Publishing

This is not an inventory of everything I have done. It is a record of things worth returning to. Entries can change when the work changes. Uncertainty belongs in the record; it is not something to sand off before publication.