Why I Still Sketch Algorithms on Paper

I still sketch algorithms on paper because paper resists premature precision. An editor invites completion. Paper invites exploration. Boxes can stay vague, arrows can contradict one another, and an uncertain invariant can sit in the margin without pretending it is already code.

For geometric work especially, drawing matters. Once you place points, boundaries, and motion on a page, the problem stops being a paragraph and becomes a spatial object. That shift is often the difference between confusion and progress.

There is also a memory component. I remember ideas better when I have physically arranged them. The page becomes a tiny topography of the thought process.

I do not romanticize analog tools. The final work belongs in code, tests, and version control. But a blank sheet is still one of the cheapest debugging environments I know.

Computers are excellent at exactness. Humans are still better at gesturing toward structure before it is fully named.