fluidnet — the teaching cycle
Your work splits into shapes a net already owns, which come back repaired and proved, and
open-ended work you solve yourself and may then teach, feeding the vocabulary. Seven of your own classes
fit in a guard, and the cap is per process — so the vocabulary grows by adding nets.
the teaching cycle Your vocabulary sits at the top. Work splits into taught shapes, which a net repairs and proves without your attention, and open-ended work, which you solve and may then teach, feeding back into the vocabulary.
Your vocabulary 7 shapes taught, across 1 net Your work a failing test Taught shape the net owns this Repaired, with a proof no attention from you Open-ended nothing taught matches You solve it then teach the shape
Shapes taught
7
real fixes that are 1 line
37%
The seven-slot arithmetic is exact — a dictionary holds 16 kinds, 7 of them reserved for your own
classes, and the cap is per process. The 37 / 70 / 91 figures are measured on real maintainer fixes across
click, arrow, rich and python-sortedcontainers. How fast your own split shifts from the lower lane to the
upper one is not measured — that depends on how repetitive your work is.