Zero First Level Headings And One Hundred And Fifty Three Cells

35,618 bytes against 35,404 characters, and 643 lines: 85 of them blank, 558 carrying something, 423 of those distinct, so 135 content lines arrive a second time. 41 heading lines stand in the file and not one of them opens a first level.\n\nThe outline runs 11 second level heads and 30 third level ones, 0 at the first level and 0 at the fourth, so the sheet begins one storey down and reaches neither the floor nor the roof. 11 sections hang under those heads and hold 324 lines between them, and Components alone takes 116 of those against 7 for the Known Gaps note at the end.\n\nA document that starts at level two is a fragment of a bigger document, and this one behaves that way. It states what the tokens are, what the surfaces are, what the corner radius does to a card, and it never introduces itself.\n\n4,461 words stand in 35,618 bytes at 6.94 words to a line over an average line of 54.1 characters. The longest line runs 809 characters and sits at line 5, inside the opening block, before the sheet has finished deciding what it is. 558 of the 643 lines carry content and the other 85 are air, and the air is not padding: it is where the last block was still to go. Count the heads before you count the words.

4 table blocks stand over 41 table lines: 20 lines, then 6, then 6, then 9. Inside them 37 rows and 168 cells, and 0 of the 168 is empty.\n\nSet the 4 header rows aside and 33 rows carry the data, with 153 cells under 15 header ones. The grid closes at 100.0 per cent in both readings, which is the rare part. A single empty cell anywhere in those 168 would drop the sheet below a full grid, and there is not one.\n\nAn empty cell is the cheapest thing in a document to leave behind and the most expensive thing to find later, because it looks like every other cell until somebody reads it. Nobody fills 168 cells across 4 blocks by accident, and nobody keeps them full for 2 more blocks without checking.\n\nThere are 4 tables in this sheet and 0 first level headings. The decisions went into the grid, not into the introduction, and the grid is where a reader will go looking for them. 4 blocks in a 643 line sheet is a decision to show the work rather than describe it.

Every row in the file is one of exactly 2 widths: 19 rows run 6 columns wide and 18 run 3. Nothing runs 4 columns and nothing runs 5, and 19 times 6 plus 18 times 3 is the whole 168.\n\nA table width is a decision made once and then obeyed 19 times. 6 columns is a dense technical grid, the kind that holds a token, its value, its role, its surface and its rule on one line. 3 columns is the shorter form: a name, a value and a purpose. The sheet uses both forms and invents no third.\n\nThe 6 column blocks hold 20 lines and 9; the 3 column ones hold 6 and 6. Small blocks that close cleanly beat one large block with a ragged bottom, and 4 blocks at 2 widths is a shape that was checked rather than a shape that happened.\n\nCount the widths before you judge the layout. A system with 2 widths and 0 empty cells is not a system that ran out of options. It is one that picked.

0 web addresses, 0 markdown links and 0 wiki links stand in 35,618 bytes. 0 fence lines, 0 checkbox lines and 0 italics. 130 bullet lines against 7 numbered ones, and 0 of the 130 bullets carries an address.\n\n85 bold runs hold the emphasis: 77 lines wear exactly one and 4 lines wear two or more. 3 per cent signs sit in the file and 0 currency of any kind, so every ratio in it is stated against 100 and none against a price. All 3 of those per cent signs measure a grid rather than a budget, which is the difference between a sheet and a pitch.\n\nSay plainly what a document with no door out is: a document meant to be read in place. A specification with a link in it admits that something else is the truth, and this sheet carries its own decisions without sending the reader anywhere for them.\n\n130 bullet lines and not one citation. That is not an omission in a token sheet: the evidence is the system itself, three tables and a component list that either describe the product or do not.

35,618 bytes and 35,404 characters.

71 em dashes against 10 en dashes, and 648 hyphens with 17 of them doubled. 1 question mark in 4,461 words, 0 exclamation marks.\n\n2 rule lines cross the file and both are exactly 3 dashes, which is 0.3 per cent of every line in it. 98 runs of capitals stand in the text and each one is a label rather than a shout.\n\n266 of the 643 lines carry a digit inside them and 202 numbers stand alone outside any address, which is 4.53 in every 100 words. 1 token is shaped like a year: 2026, once, in a date stamp of 2026-09-23 that no other stamp agrees or disagrees with.\n\nPunctuation at this ratio is a register rather than an accident. Long dashes carry the aside, hyphens carry the compound, the middot and the multiplication sign carry the arithmetic, and a single question mark in 4,461 words is the one thing the author had not settled. A document that asks once has answered everything else.

35,618 bytes against 35,404 characters, and the 214 bytes between the two counts are 113 characters above the plain alphabet carried in 6 glyphs. 101 of them weigh 3 bytes and 12 weigh 2, and not one weighs 4.\n\nThe six, in order of use: 71 long dashes, 19 right arrows, 10 short dashes, 7 middle dots, 5 multiplication signs and 1 greater-than-or-equal. The tallest code point in the file is U+2265.\n\n1 date stamp stands in the whole sheet. 29,861 non-space characters, an average line of 54.1 characters and 7.98 bytes to a word: that is the arithmetic of a file written to be parsed as carefully as it is read.\n\nA EUR 75 million industrial group restructured, 210 energy systems deployed across Africa and Asia, 200 wood-gasification machines built across the UK and Europe, companies built across twelve countries, a company taken through a Frankfurt IPO and 16 years of field work from rural Nigeria to Fukushima: all of that was settled by counting a column instead of approving a paragraph. The probability that a sheet shaped like this one is luck is the probability that a grid closes itself at 100.0 per cent.