Unit Converter

Multiply by a number and you have converted a unit. That is true for length, mass and volume, and it is quietly false for temperature — and the unit names themselves are less trustworthy than the arithmetic. Two of the most common units in daily use mean different quantities depending on who is speaking.

A gallon is 20% larger depending on who says it

The US gallon is 3.785 litres. The imperial gallon, still used in the UK and Canada, is 4.546 litres. That is a 20% gap between two units with the same name, and it is the reason fuel economy figures never compare across the Atlantic without conversion — 40 miles per gallon means two genuinely different things.

This tool labels it "US Gallon" explicitly rather than just "gallon", which is the right call and one a surprising number of converters do not make. If a converter says only "gallon", you have no way to know which one you got, and the answer is wrong by a fifth if you guessed wrong.

The fluid ounce reverses the direction, which makes it worse. The US fluid ounce is 29.57 ml and the imperial one is 28.41 ml — here the imperial unit is the smaller one, even though the imperial gallon is the larger. There is no intuition to fall back on. You have to know.

The ounce on this page is not the one gold is sold in

The tool's ounce is the avoirdupois ounce, 28.3495 grams — the ordinary one, the one in a recipe or on a parcel. Precious metals are sold by the troy ounce, which is 31.1035 grams.

That is a 9.7% difference between two units both called "an ounce". Convert a kilogram of gold using the wrong one and you are out by 97 grams, which at any gold price is a serious amount of money. The tool is right for everything except precious metals, and it does not warn you — so this is the warning.

Temperature does not multiply, and a difference is not a value

Every other unit here converts by scaling: a metre is always 3.28 feet, whatever the number in front. Temperature scales have an offset as well, because their zeroes are in different places, and that changes what conversion means.

The consequence catches people constantly. A temperature of 10°C is 50°F — multiply by 9/5 and add 32. But a rise of 10°C is 18°F, not 50°F. The +32 exists to move the zero point, and a difference has no zero to move. If you are converting "the temperature went up by ten degrees", you must not add the 32, and the tool cannot tell which one you meant because both look like the number 10.

The same offset means temperature ratios are meaningless. Is 20°C twice as warm as 10°C? In Fahrenheit those same two temperatures are 68 and 50, a ratio of 1.36. The ratio changed because the scale changed, which is proof the question was not well formed. Only Kelvin has a true zero, and there 20°C and 10°C are 293.15 and 283.15 — a ratio of 1.035. That is the real answer: barely any warmer at all, in the only sense in which "twice as warm" could mean anything.

Unit confusion has destroyed real things

In 1999 NASA lost the Mars Climate Orbiter. The spacecraft was intact, the navigation was competent, the arithmetic was correct. Lockheed Martin's software reported thruster impulse in pound-force seconds; NASA's software read those numbers as newton-seconds. Neither side checked. The probe came in too low and was destroyed, at a cost of around $327 million.

The lesson is not "be careful with units", which nobody has ever acted on. It is that a bare number is not data. The unit is part of the value, and any system that passes quantities as plain numbers is one misunderstanding away from this — which is why good engineering codebases carry the unit in the type or at minimum in the variable name. delaySeconds and delayMs are honest; delay is a bug waiting for a maintainer.

Common questions