A point is not a spot on a grid or a number. It is a thing with an identity. Without an identity it is only potential, an abstraction waiting to become real. Give it an identity and it becomes a point you can hold, move, and work with. Everything in this way of thinking is a point, and the first rule is that a point must be a some-thing, not a maybe-thing.
Established In software this is the difference between a type, which is potential, and an instance, which is a real thing with its own identity.
Everything runs on two moves, and they are just the shapes of dividing and multiplying.
Separate a point and it blooms into its parts. One thing opening into many. That is division, seen as a shape.
Gather the parts and they collapse back into one point. Many drawn into one. That is multiplication, seen as a shape. Gather an identity together with how it behaves and you get its state, and that state is the next identity you carry forward. The work moves ahead by gathering.
Breaking a whole into parts and composing parts into a whole is everyday computing (Established). Naming the two as bloom and collapse, and treating the gathered state as the next thing you act on, is my framing (Mine).
When this says a thing has a shape, it means a real surface, not a sum on paper. The gather of an identity and its behavior has one: a saddle, a twisted square, a surface that rises along one direction and falls along the other, with a ridge running down its diagonal and a half twist through it.
You do not read a number off it. You lay a lens over it, and the lens reads one property at the place you are standing: a height, a slope, a colour, a tone. Change the lens and the very same surface gives a different reading. The shape decides what is possible; the lens decides what you see.
So gathering is not arithmetic. It is landing on a real surface and reading it through a chosen lens. That is the whole reason this is told in shapes and not in symbols: the saddle is the thing, and the symbols are only its shadow.
The fuller picture of that surface, and why a formula can have a body, is in The SEAT and the Shadow.
There are two kinds of motion, and this is the heart of why it is called dimensional.
Adding and taking away keep you on the line. They move you within a level, back and forth, same scale.
Gathering and separating make area. They move you between levels, up into a whole or down into the parts. Those two are the dimensional moves; adding and subtracting are not.
And when a separation does not come out clean, what is left over is the residual, the little bit of disorder you have to carry. (Computing calls this the remainder, or modulo.)
Established Treating adding as sliding along a length and multiplying as making an area is the oldest picture of arithmetic, older than the symbols. Calling one "within a level" and the other "between levels" is the framing here (Mine).
The same two moves build one ladder, and you climb the very same ladder whether you are shaping data, thinking through a problem, or asking a question.
That last rung is the trick. Once a thing is fully understood, it collapses back into a single point one level higher, a closed box again, reusable and measurable like any other point. Then you can build on it without re-opening it. You went all the way out into the volume and came back with a point.
A closed box is also a security boundary. The world sees only the box, the point, the one value you expose; everything inside is contained, reachable only by opening it with the right access. The same shape that makes a point reusable makes it safe: outsiders touch the face, not the inside.
A finished, fully-understood thing becoming a reusable closed box is abstraction, encapsulation, and information hiding (Established), and information hiding is a security principle: you can only act on what you hold a reference to. Running the one ladder across data, thinking, and prompts with a single vocabulary is the contribution here (Mine).
Call it the Ms Kravitz rule: every point knows the points right next to it. From that small, local knowledge you get two large things almost for free. News travels, because each point passes word to its neighbors until it has reached everywhere. And damage repairs itself, because a point that loses its way can be patched from what the points around it still hold. Nothing needs to see the whole; the whole takes care of itself from the edges in.
Established Local-neighbor knowledge giving you propagation and self-healing is how gossip protocols, mesh networks, and resilient distributed systems already work. The nosy-neighbor name is just to make it stick.
Nothing separates until it is observed. A point stays gathered and quiet until you ask for its parts, and then it blooms just enough to answer the question, and no more. You hold the small definition and let the detail appear on demand, rather than keeping it all spread out and stored.
Established Deriving what you need only when you need it is lazy evaluation, and computing it from a small definition instead of storing it is derive-not-store. Both are ordinary, sturdy ideas.
So dimensional programming is two moves and one rule. Separate to bloom a point into its parts. Gather to collapse parts into a point. And let every point keep its identity and know its neighbors. Told in shapes, it stays small enough to hold in your head, which is the whole point.