Generator

State held against the length of the window, with the exact structure beside it

Rendered here at the parameters it defaults to, with every essay that calls it — which is the same list as the blast radius of changing it.
State held against the length of the window, with the exact structure beside itEach series is an exponential histogram at one accuracy, measured over 40,000 arrivals of a bursty bit stream. The straight line at slope one is the exact answer, which is W bits and has no cleverness available to it. The summaries grow like the square of the logarithm, so the ratio between the two improves without limit — and BELOW a certain window length the summary is the larger of the two: at ε = 0.2 up to W = 500, at ε = 0.1 up to W = 500, at ε = 0.05 up to W = 1,000. An approximation that costs more than the exact answer is not a compromise, and the crossing point is the thing to know before choosing one. Both axes are logarithmic.1,00010³window length W, in arrivalsbits of state heldW bits, exactε = 0.2ε = 0.1ε = 0.05sliding-window model · 40,000 arrivals · state from the shape of the structure2,808 bits at W = 8,000

State held against the length of the window, with the exact structure beside it

Each series is an exponential histogram at one accuracy, measured over 40,000 arrivals of a bursty bit stream. The straight line at slope one is the exact answer, which is W bits and has no cleverness available to it. The summaries grow like the square of the logarithm, so the ratio between the two improves without limit — and BELOW a certain window length the summary is the larger of the two: at ε = 0.2 up to W = 500, at ε = 0.1 up to W = 500, at ε = 0.05 up to W = 1,000. An approximation that costs more than the exact answer is not a compromise, and the crossing point is the thing to know before choosing one. Both axes are logarithmic.

Drawn at 700 × 430, wide on the page. Everything above is what window-cost returns with no arguments; the caption is the generator's own, computed from the numbers in the drawing rather than written beside it.

3 essays call window-cost. The drawing above is what it returns with no arguments at all; every call below passes it something, because a placement that passes nothing draws whichever member of the family the generator happens to default to rather than the one its essay argues about — which is what optcheck and figfill exist to catch.

Where it is called

Changing this generator changes every one of these figures.

The library, page 5 of 5 — where window-cost sits