Euclidean plane geometry

The formula (the core equation(s), each with a one-line reading of what it says)

Triangle angle sum:   α + β + γ = 180°

Reads: the three interior angles of any triangle on a flat plane always total two right angles — the single fact that lets you find the third angle from the other two.

Pythagorean theorem:   c² = a² + b²

Reads: in a right triangle, the square on the hypotenuse equals the squares on the legs combined — the reason a 3-4-5 layout gives you a guaranteed square corner (9 + 16 = 25).

Law of cosines:   c² = a² + b² − 2·a·b·cos(C)

Reads: the Pythagorean theorem with a correction term for corners that aren't square; set C = 90° and the cosine dies, handing back c² = a² + b².

Similar triangles:   a/a′ = b/b′ = c/c′

Reads: same shape at a different size — every ratio of corresponding sides is the same number, which is why a scale drawing works at all.

Circle:   C = 2·π·r     A = π·r²     chord = 2·R·sin(θ/2)

Reads: circumference grows with the radius, area with its square, and the straight-line distance between two points on a bolt circle comes from half the included angle.

Triangle area:   A = ½·b·h     or     A = √( s·(s−a)·(s−b)·(s−c) ),  s = (a+b+c)/2

Reads: half of base times height when you have a height; Heron's form when all you have is three side lengths. For the 3-4-5 triangle both give 6.

Where you meet it (2-4 concrete engineering situations, specific: bench, test stand, review board)

Laying out a bolt pattern at the machine. Eight holes on a 200 mm bolt circle: the included angle is 45°, so the hole-to-hole chord is 2·100·sin(22.5°) = 76.537 mm. Punch that number into a caliper check after drilling and you catch an indexing error before the mating part shows up. Six holes on a 100 mm circle is the party trick — the chord equals the radius, 50 mm exactly, because six equilateral triangles tile the hexagon.

Squaring a fixture or a slab without a square. Measure 3 units along one edge, 4 along the other, and adjust until the diagonal tapes 5. Concrete crews, weld fixture builders, and anyone anchoring a test stand baseplate uses the converse of Pythagoras daily, usually without naming it. Any multiple works — 6-8-10 spans more of the slab and halves the angular error for the same tape resolution.

The tolerance stack review. A stack-up across a bracket assembly is a chain of lengths and angles living on an assumed flat plane. When a datum surface tilts by a small angle, the displacement at the far end is that angle times the lever arm — similar triangles doing the conversion. Half the arguments at a stack-up review are really arguments about which triangle somebody drew wrong.

The strain-gauge and survey bench. Rosette reduction, antenna mast guy-wire lengths, sling angles on a lift plan, the run of a conduit across two offsets — all law-of-cosines problems. Legs of 60 and 80 units meeting at 120° span √(60² + 80² − 2·60·80·cos 120°) = 121.66 units, and no amount of Pythagoras alone gets you there because the corner isn't square.

How it works (the real substance — behavior, gotchas, limits of validity, the mistake people make)

Everything above follows from five assumptions Euclid wrote down: you can draw a line between any two points, extend it, draw a circle of any center and radius, all right angles are equal, and — the load-bearing one — through a point not on a line there is exactly one parallel to that line (Playfair's restatement of the fifth postulate). The 180° angle sum, similar triangles, and the Pythagorean theorem are all consequences of that parallel assumption. Change it and they change: on a sphere, triangle angles sum to more than 180°, and there is no such thing as two similar triangles of different sizes.

That is not academic trivia; it is the limit of validity. The Earth is the sphere in question. Flat-plane geometry is exact on your drawing and very slightly wrong on the ground: curvature drops the surface about d²/(2·R) below the tangent plane — 0.78 mm at 100 m, 78 mm at 1 km. Machine shops and building sites never notice. Long baseline surveys, runway layouts, and range instrumentation do, which is why geodetic work uses map projections and why a surveyed triangle covering 1000 km² closes about 5 arcseconds over 180° instead of on it. If a long-baseline traverse keeps failing closure by a consistent small amount, stop blaming the instrument and check whether someone is doing plane geometry on a curved planet.

The everyday gotchas are humbler:

Chords are not arcs. The distance between two bolt holes is the chord, 2R·sin(θ/2), not the arc length R·θ. At small angles they agree to within a fraction of a percent, which is exactly why the error survives checking — until the hole count gets low and the angles get big.
Angles don't add across a bent chain the way lengths do. A stack of angular tolerances multiplies through lever arms; two half-degree tilts at opposite ends of a 500 mm part are not "one degree of tilt," they are whatever the triangle says they are.
Similar-triangle scaling only holds if every dimension scales. Scale a bracket's outline by 2 and keep the same sheet thickness, and nothing about its stiffness or stress scales the way the drawing implies. Geometry scales; physics carries units.
A construction is a proof; a CAD sketch is a request. Euclid's figures are guaranteed by the postulates. A sketch that "looks constrained" in CAD can still have a degree of freedom left, and it will find it at the worst moment. Fully-defined or it isn't done.

The deeper habit plane geometry teaches is the axiomatic one: state your assumptions, then only claim what follows from them. Every tolerance stack-up is a small theorem — givens (datums, basic dimensions), permitted moves (the geometry above), conclusion (the gap at the interface). When a stack-up is wrong, it is wrong the way a proof is wrong: a step got used that the givens don't support.

History (who derived it and when, told as a short story with inline [n] citations)

Around 300 BC, in Alexandria under the first Ptolemy, Euclid assembled the Elements — thirteen books, 465 propositions, built from definitions, five postulates, and common notions ^[1]^[3]. Almost nothing reliable is known about the man himself; the book is the biography ^[1]. Book I alone carries most of what an engineer uses: triangle congruence, parallels, areas, the angle sum as Proposition 32, and the Pythagorean theorem as Proposition 47 ^[2]^[3]. It stayed the standard geometry text for roughly two thousand years, which no other technical book has managed ^[1]^[3].

The irritant the whole time was the fifth postulate. It reads like a theorem that should be provable from the other four, and for two millennia people tried — every attempt smuggling in an assumption equivalent to the thing being proved ^[4]^[5]. In 1795 John Playfair recast it in the form taught today: through a point not on a line, exactly one parallel ^[4]^[5]. In the early 1700s Girolamo Saccheri assumed it false and ground out theorem after theorem of a strange consistent geometry, then declared he had vindicated Euclid anyway ^[4]^[5]. Gauss privately concluded the postulate was independent and said nothing publicly ^[4]^[5]. Lobachevsky published the alternative geometry in 1829; János Bolyai published his own version in 1832, as an appendix to his father's textbook ^[4]^[5]^[6]. In 1868 Beltrami produced concrete models showing the new geometry is exactly as consistent as Euclid's — if one has a contradiction, so does the other ^[5]^[6].

The postmortem came in 1899. David Hilbert's Grundlagen der Geometrie rebuilt Euclid on a complete slate of about twenty axioms in five groups — incidence, order, congruence, parallels, continuity — filling the gaps where Euclid had quietly leaned on the diagram ^[7]^[8]. The engineering moral survived the mathematics: Euclid's geometry wasn't wrong, it was conditional. Know the assumptions your result rides on, because someone eventually builds the case where they don't hold.

Related tools (bullet list of HE calculator slugs that use or neighbor this topic, as /tools/ links)

/tools/convert-angle — degrees, radians, and the chord formula's θ/2 all live here
/tools/convert-area-volume — the area formulas above, in whatever units the drawing demands
/tools/moment-of-inertia-shapes — plane-figure geometry one integral deeper
/tools/section-modulus-rect — rectangle geometry driving bending stress
/tools/concrete-slab-volume — plane areas times depth, plus the 3-4-5 check before the pour
/tools/tank-volume-horizontal — circular-segment area doing paid work

Sources

Written by HE in our own words from the cited sources — engineering judgment included, your stamp still required. All entries →