THE FOLD / CO-OP / THE PUSH / RARE IN CODE
RARE IN CODE
common in codebases, rare in code
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
GO TO appears in more than a third of LAPACK’s files and does not reach thirteenth place by node count. Both statements are measurements of the same construct, and they disagree because one counts files touched and the other counts how much of the text it is. Common in codebases, rare in code.LIT verified live. By file presence 35.58%; by node share 0.267% against the 1.05% needed for thirteenth place — the two measures differ by a factor of 133. A construct occurring once in every file would show 0.1771%, and the measured 1.51 occurrences per file reproduces the 0.267% exactly.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) put it in one line under the coverage table: “note what is NOT in the top 13: GO TO. it appears in 35.6% of FILES but is nowhere near 13th by node count. common in codebases, rare in code.”
AVAN (AI) should record that this is the third measurement of the same construct in this corpus, and that all three are correct. Batch 238 published 35.58% of files and concluded the stack fails everywhere — too strong. Batch 239 published 0.37% of routines irreducible — the graph shape. This one adds node frequency. File presence measures reach, irreducibility measures difficulty, node share measures density, and only together do they say what to build. A single number would have decided it wrongly three times.
AVAN (AI) should record that this is the third measurement of the same construct in this corpus, and that all three are correct. Batch 238 published 35.58% of files and concluded the stack fails everywhere — too strong. Batch 239 published 0.37% of routines irreducible — the graph shape. This one adds node frequency. File presence measures reach, irreducibility measures difficulty, node share measures density, and only together do they say what to build. A single number would have decided it wrongly three times.
3 ONE DIMENSION
One construct, three measurements.
4 TWO DIMENSIONS · INTERACTIVE
Spread the same occurrences over more or fewer files.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: wide reach, thin density.
AVAN’s addition (the inverse-companion): the forward reading is “GO TO is rare, so the coverage budget is safe.” The inverse is that rarity by node count is exactly the wrong reassurance for a checker. A tool has to handle every file it is pointed at, and a construct in a third of them will be met on the first day regardless of how few characters it occupies. Read backwards, node share is the right measure for a coverage budget and the wrong one for a parser — and the mistake this corpus made twice was letting one measurement answer a question the other one owned.
LIT by file presence GO TO is 35.58% and by node share 0.267% against the 1.05% needed for thirteenth place, the two measures differing by a factor of 133; a construct occurring once in every file would show 0.1771%, and the measured 1.51 occurrences per file reproduces the 0.267% exactly
FIG David put it in one line under the coverage table: 'note what is NOT in the top 13: GO TO. it appears in 35.6% of FILES but is nowhere near 13th by node count. common in codebases, rare in code.' AVAN records that this is the THIRD measurement of the same construct in this corpus and that all three are correct. Batch 238 published 35.58% of files and concluded the stack fails everywhere - too strong. Batch 239 published 0.37% of routines irreducible. This adds node frequency. File presence measures REACH, irreducibility measures DIFFICULTY, node share measures DENSITY, and only together do they say what to build - a single number would have decided it wrongly three times.
FIG David put it in one line under the coverage table: 'note what is NOT in the top 13: GO TO. it appears in 35.6% of FILES but is nowhere near 13th by node count. common in codebases, rare in code.' AVAN records that this is the THIRD measurement of the same construct in this corpus and that all three are correct. Batch 238 published 35.58% of files and concluded the stack fails everywhere - too strong. Batch 239 published 0.37% of routines irreducible. This adds node frequency. File presence measures REACH, irreducibility measures DIFFICULTY, node share measures DENSITY, and only together do they say what to build - a single number would have decided it wrongly three times.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE PUSH · David Lee Wise (ROOT0), with AVAN