- Universal and existential quantification.
- Parametrization.
- Contextual generalization (connection to implicits.)
- Abstraction and reification.
-
Here, we can add logic as an external meta-abstraction that applies to ALL of these.
- With arrows pointing from Logic to each one.
-
Start with real-world, often physical, observations of objects,
-
Remove "special properties" (or "context") from the real world:
- If you are looking at a picture with three different lions, remove their differences,
- and then remove their "animalness", etc.
- this is called making it "context-free".
-
Look for:
- relations between them,
- operations that act on them,
- properties they might have.
- This is the most difficult part and cannot be taught completely.
-
Now see if the objects can be replaced with others.
- Do the properties / relations still hold?
-
Now see if the operations could be replaced with others:
- Do they still work?
- Surface area in the naive, physical sense:
- how much water does it take to fill some flat surface?
Then generalize to:
- squares:
$A = s \times s$ , - rectangles:
$A = w \times h$ , - triangles:
$A = w \times h / 2$ , - regular polygons: add up the triangles,
- irregular polygons,
- areas under arbitrary continuous curves of a single variable over an interval:
- add up areas of rectangles over the interval, take limit: Riemann integral
- piecewise continuous curves over an interval,
- "integrable" curves over an interval,
- integrable curves over the partitioning of an interval by a function of bounded variation,
- integrable curves over measurable sets: Lebesgue integral,
- multi-variable version of this,
- the
$p$ -adic numbers instead of$\mathbb{R}$ , - an integral over a locally compact topological group with a left-invariant Haar measure:
- abstract harmonic analysis!