Conversion Methodology

Select evidence for the stated quantity

BIPM’s SI definitions and decimal prefixes establish SI units and their multiples. NIST SP 811 Appendix B.8 contains both exact and approximate factors; the entire table is not an exact-factor list.

Specific definitions matter: NIST footnote 22 gives the exact avoirdupois pound of 0.45359237 kg. NIST Handbook 44 Appendix C supplies named customary relationships, including the US liquid gallon of 231 cubic inches. NIST’s revised land-unit table identifies international-foot definitions. A source label or HTTPS link alone does not establish that a particular claim is supported.

Check the length definitions, mass conventions and squared and land units for the unit named by the original measurement. Survey-foot source quantities require their own definition; the area calculator uses the international foot.

Calculate with the appropriate relationship

Linear categories convert through their common base unit. The canonical unit data stores a factor per unit; the implementation calculates:

result = value × from.toBase / to.toBase

Temperature uses offsets and scale changes through Celsius. For example, (68 − 32) × 5/9 = 20 °C, matching NIST’s temperature formulas. A temperature interval and an absolute temperature are different quantities.

The fuel economy calculation routes positive finite inputs through L/100km. Economy-to-economy comparisons such as US MPG to Imperial MPG are proportional; consumption-to-economy comparisons such as L/100km to MPG are reciprocal. The US constant is derived as 100 × 3.785411784 L/US gallon ÷ 1.609344 km/international mile. Conversion does not establish a vehicle’s measured or rated consumption.

The data definitions separate MB = 1000000 bytes from MiB = 1048576 bytes, and a byte is 8 bits. These relationships are documented in NIST’s public binary-prefix reference.

Worked calculations from shared routines

Keep recipe portions separate from physical density

King Arthur’s ingredient chart supplies original author portions: 120 g per cup of all-purpose flour, 198 g per cup of granulated sugar and 113 g per half-cup of butter. The source does not specify its cup volume in mL. The cooking measurement tools scale those portions as recipe estimates, using the stated assumption 1 cup = 16 tablespoons = 48 teaspoons.

An author’s grams-per-cup portion is not an independently verified physical density. mL-to-mass conversion requires separate density evidence. The supported water reference is 0.99820675 g/mL for air-free water at 20 °C, from NISTIR 7383 (2019), GLP 10 Table 2. Under that condition, 100 mL gives 99.820675 g before display rounding. This is a condition-specific density reference, not an exact universal unit definition.

Rice has no verified portion or density in the current ingredient data, so its cross-quantity results are unavailable. Pure mass-to-mass or volume-to-volume conversions do not depend on the selected ingredient. The unit-system reference names the physical unit conventions; it does not supply missing ingredient evidence.

Distinguish calculation and display precision

Stored factors and calculations use JavaScript floating-point numbers. An exact definition may have a rounded decimal representation, and an approximate source value remains approximate. Extra displayed digits do not establish measurement accuracy; preserve the uncertainty and precision of the original measurement.

• The main unit widget defaults to four decimal places, selectable from 1–10, and trims trailing zeros from fixed-point results. Nonzero magnitudes at least 10¹⁰ or below 10⁻⁶ use scientific notation with exactly the selected number of decimal places (at most six) in the significand, without trimming.
• Generated pair conversion tables allow at most two decimal places for magnitudes at least 1000, four for magnitudes at least 1, and eight below 1. A sufficiently small nonzero value can therefore display as zero.
• Ingredient results and generated baking-chart entries use at most two decimal places. Unit-system reference entries use at most six decimal places. Category pillar reference tables use at most ten significant digits.
• Key formula cards display factors to six significant digits. A pair’s formula, explanation, source and examples are authored content records, displayed separately from canonical unit factors. Example card numbers do not pass through the widget formatter.

A displayed formula is not guaranteed to export every internal digit. Calculate from the named definition or applicable source value before rounding the final answer. The mass estimate guide demonstrates the error introduced by rounded shortcuts.

Check the scope of a reference chart

Physical conversion tables differ from clothing-size correspondences. The baking-weight chart is generated from the ingredient portion model; the metric/imperial chart is generated from named unit definitions. Other chart records contain static entries. A chart’s source label is not evidence that every brand correspondence has passed a manufacturer audit. Check the chart’s stated scope and the original manufacturer’s applicable sizing information.

Understand publication dates and automated checks

Standalone pair routes are selected from an explicit publication manifest. Published records require reviewed status and a valid review date; future review dates are rejected against the build/reference UTC date. Advertisement eligibility is a separate flag requiring publication. Published pair content’s “Last reviewed” date must match its manifest record.

The content audit checks known units, required nonempty fields, HTTPS source URL syntax and at least two finite examples. It compares authored examples with production calculations, allowing 0.1% relative difference for nonzero results and an absolute tolerance of 10⁻¹² at zero. It also checks specified false metric-unit boilerplate and rejects review-date fields on unpublished skeleton content.

These checks do not retrieve external sources or verify every prose claim or FAQ. A review date records the dated content review; it does not establish a scheduled re-review, a named external reviewer or an automatic source audit. This methodology was inspected against the implementation on 2026-10-06.

Report a specific discrepancy

Use the contact page with the page URL, entered amount, original unit convention, expected result and primary evidence. Identify whether the discrepancy concerns a factor, assumption, authored example or rounded display so it can be reproduced.