Conversion calculations are inconsistent

[JohnathanGander]

Your readme says:
"(10 inches - 0) / 0.0254 = 0.254 metres"
Which is incorrect.

This is correct:
(10 inches * 0.0254) + 0 = 0.254 metres

It appears that this is true for almost every other conversion (eg: all measurements in length, temperature is ok). Simply inverting the multiplier (1/multiplier) should resolve the issues.
1/0.0254 ~= 39.37
(10 inches - 0) / 39.37 ~= 0.254 metres

[JohnathanGander]

Upon further investigation it would be easier to change temperature rather than all other units. I double checked your dimensions for time, length, mass, electric current, amount of substance, luminous intensity, solid angle, and plane angle.

The new formulas for conversion should be:
base = ( original * multiplier ) + offset
new = ( base - offset ) / multiplier

I've updated and checked the conversions for temperature.

		"temperature": {
			"name": "temperature",
			"symbol": "Θ",
			"baseUnit": "kelvin",
			"vector": false,
			"units": {
				"kelvin": {
					"name": "kelvin",
					"otherNames": ["degree kelvin", "degrees kelvin"],
					"symbol": "K",
					"type": "si",
					"systems": ["si"],
					"notes": "As a postpositional adjective, the plural kelvin is used instead of kelvins"
				},
				"celsius": {
					"name": "celsius",
					"otherNames": ["degree celsius", "degrees celsius"],
					"symbol": "°C",
					"type": "si",
					"systems": ["siCommon"],
					"tags": ["cooking"],
					"offset": 273.15
				},
				"fahrenheit": {
					"name": "fahrenheit",
					"otherNames": ["degree fahrenheit", "degrees fahrenheit"],
					"symbol": "°F",
					"type": "customary",
					"systems": ["englishUnits", "imperial", "usCustomary"],
					"tags": ["cooking"],
					"multiplier": 0.555555555555555555555,
					"offset": 255.37222222222222222
				},
				"rankine": {
					"name": "rankine",
					"otherNames": ["degree rankine", "degrees rankine"],
					"symbol": "°R",
					"otherSymbols": ["Ra"],
					"type": "customary",
					"systems": ["nonStandard"],
					"rare": true,
					"multiplier": 0.555555555555555555555
				},
				"romer": {
					"name": "romer",
					"otherNames": ["degree romer", "degrees romer"],
					"symbol": "°Rø",
					"otherSymbols": ["°R"],
					"type": "customary",
					"systems": ["nonStandard"],
					"rare": true,
					"multiplier": 1.9047619047619,
					"offset": 258.8642857,
					"notes": "Technically Rømer"
				},
				"newton": {
					"name": "newton",
					"otherNames": ["degree newton", "degrees newton"],
					"symbol": "°N",
					"type": "customary",
					"systems": ["nonStandard"],
					"rare": true,
					"multiplier": 3.0303030303,
					"offset": 273.15
				},
				"delisle": {
					"name": "delisle",
					"otherNames": ["degree delisle", "degrees delisle"],
					"symbol": "°D",
					"otherSymbols": ["°De"],
					"type": "customary",
					"systems": ["nonStandard"],
					"rare": true,
					"multiplier": 0.666666666666666666,
					"offset": 339.8166666666666
				},
				"reaumur": {
					"name": "reaumur",
					"otherNames": ["degree reaumur", "degrees reaumur"],
					"symbol": "°Ré",
					"otherSymbols": ["°Re", "°r"],
					"type": "customary",
					"systems": ["nonStandard"],
					"rare": true,
					"multiplier": 1.25,
					"offset": 273.15,
					"notes": "Technically Réaumur"
				},
				"planckTemperature": {
					"name": "Planck temperature",
					"symbol": "T\u209A",
					"type": "customary",
					"systems": ["planck"],
					"multiplier": 1.416833e32,
					"uncertainty": 0.000085e32
				},
				"naturalTemperature": {
					"name": "electron volt of temperature",
					"plural": "electron volts of temperature",
					"symbol": "E",
					"otherSymbols": ["E", "eV", "eV of temperature"],
					"type": "customary",
					"systems": ["natural"],
					"multiplier": 1.16e4,
					"uncertainty": 0.01e4
				}
			}
		},

There were couple other things that were incorrect. Centesimal minute of arc's multiplier should be 1.57e-5 not 0.157e-3=1.57e-4. Degree's multiplier could be better rounded to 17.45329252e-3.

I see in your to-do that you would want to include multiple multipliers, might I suggest a numerator and denominator for the multipliers as well. With this we could gain back some accuracy we lost in the above temperature conversions. You might also want to include a keyword in the multipliers for pi. That way the operating system can use a better value of it and we wouldn't have to store a computed approximation using it. Accuracy would be on the user not the json.

Kelvin doesn't have a degree symbol.