Technical note 10

TN-10: Millimeters

Why this project measures in integer millimeters, and why that is a design decision rather than a habit. First draft 18 Aug 2026, from KU's account of his own conversion and the project's practice. Status: DRAFT.

The claim

The millimeter is the natural quantum of human-scale making. One millimeter is close to the real resolution of any reasonable manual wood process, so rounding every dimension to 1 mm costs nothing in accuracy and buys integer arithmetic everywhere: no fractions, no compound units, no conversion, and a cut list you can check in your head. This project's chairs are designed, drawn, and built in integer millimeters, and the Open Chair Standard should say so normatively.

How this project got here, a testimony

KU's account, recorded 18 Aug 2026, kept because the conversion of one American maker is the argument in miniature. He is American and grew up building in inches. At Xootr, his first manufacturing business, his brother insisted on SI units. His father, Gael D. Ulrich, a strong willed professor of chemical engineering, wrote his textbook (A Guide to Chemical Engineering Process Design and Economics, Wiley, 1984 [VERIFIED, the edition and date, 18 Aug 2026]) in SI units at a time when that was unpopular in the field; the SI-only character of the book is family testimony here and is queued for a page check. Product design work followed, all SI. But building chairs in Australia is what forced the last holdout, the shop itself, into millimeters: the tools are set up that way and the tape measure has no inches on it. The adaptation took about two minutes. The one place it still hurts is the American table saw, whose fence reads inches; more on that below.

The resolution argument

One millimeter is 0.039 inch: a bit bigger than 1/32 (0.794 mm) and a bit smaller than 1/16 (1.588 mm) [DERIVED; the inch has been defined as exactly 25.4 mm since the international yard and pound agreement of 1959, STANDARD]. That interval is exactly where manual woodworking lives. A sharp pencil line is about half a millimeter wide; a handsaw wanders more than that; seasonal movement in solid wood erases finer precision within the year. So 1 mm is the coarsest quantum that never limits the work and the finest one the work can actually hold, which is precisely what a quantum should be.

And here the project found delightful corroboration. When Australia converted, the building industry completed the change within two years, by January 1976, the first major industry group to finish, and the conversion rules named the units for building as the meter and the millimeter, with the centimeter explicitly excluded: "the centimetre (cm) shall not be used" (the quote keeps its own spelling). The reasoning recorded for that choice is the same one stated above: building tolerances rarely fall below one millimeter, so the millimeter is the right submultiple, and a centimeter would only reintroduce a decimal point [VERIFIED against the metrication record, 18 Aug 2026]. An entire country's construction sector ran this note's experiment fifty years ago and reached the same conclusion.

The arithmetic argument

Fractional inches have exactly one virtue: halving. A binary fraction halves forever without a remainder, which mattered when layout meant folding a rule. Everything else about them is a tax. Adding 3 5/8 and 2 11/16 requires a common denominator; adding 92 and 68 does not. And construction inches arrive compounded, feet plus inches plus a binary fraction, three number systems stacked inside a single dimension.

Millimeters add like money. Every dimension on a chair-sized object is a two or three digit integer, and the project's cut list shows what that buys: the keel's stations are 70, 335 and 405, and the fact that 405 = 70 + 335, the third station being the second struck again from the first mark, is visible by inspection. Single datum dimensioning plus integer stations is the whole story pole method, and it is only mental arithmetic because the numbers are integers.

One honest nuance: the halves and quarters

The Keel Chair's DERIVED geometry is not all integers: the two triangle lines sit 17.5 apart, and the keel board's wing is 16.25. No contradiction, but a rule worth stating. Dimensions a maker puts a tape on are integers by design. Derived layout values may carry decimals, because they are struck once, with a drill at a vertex or a fence setting, not measured repeatedly; the drawing should say which numbers are for the tape and which are for the record. The saw cannot see a quarter millimeter, and the layout does not ask it to: the vertex is drilled at 1.5 and the cuts run to the hole.

The machines already agree

G-code runs in millimeters. STEP files ship in millimeters. Structural plywood is 18, hardware is M6, and the barrel nut of TN-8 was printed from a model dimensioned in millimeters with printed fits in fractions of one. The bits half of this project has never used an inch. Bringing the shop to millimeters does not modernize the craft so much as reunite it with its own drawings.

The one place it still hurts

An American table saw with an imperial fence scale is the single genuine annoyance in a converted life. Three mitigations, in ascending order of principle: stick an adhesive metric tape over the fence scale; fit a digital readout, which reads either unit and cares about neither; or note that the project's own keel doctrine dissolves the problem, because a story pole has no units at all. A scribed member is the length of the gap it fills, in no unit whatsoever. The endgame of the millimeter argument is that the best dimension is the one nobody reads off a scale.

For the standard

The Open Chair Standard should carry a units clause: all normative dimensions in integer millimeters; centimeters not used, following the Australian building convention; imperial equivalents, where given, are informative conversions only and never definitions.

Open questions

1. Page-check the SI-only character of the 1984 textbook against a copy.
2. Whether the standard's conformance tests should state tolerances in integer millimeters too (likely yes: a tolerance finer than 1 mm is a claim the shop cannot audit).
3. The claim that 1 mm matches manual process resolution is argued here from pencil lines and saw practice; a measured test (ten crosscuts to a knife line, measure the scatter) would pin it and would take an hour.

Revisions

18 Aug 2026: first draft, from KU's testimony of the same day, with the 1959 inch definition and the Australian building conversion verified.

Technical Notes of the Hundred-Year Chair: public from first draft, refined iteratively, forever. A claim is VERIFIED, STANDARD, DERIVED, or UNVERIFIED, and says which. The register · the verification ledger · hundredyearchair.com