JavaScript uses one number type for integers and floating-point values. That makes arithmetic convenient for element dimensions, animations, dates, financial data, and array loops, but precision can produce surprising results. section 2.0 section 2.9
Every number is stored as a 64-bit IEEE 754 floating-point value. Only integers from -9,007,199,254,740,991 through 9,007,199,254,740,991 are guaranteed to remain precise; check boundaries with Number.isSafeInteger(). section 2.2.1
Use the number type with these practical constraints:
- Expect decimal imprecision because binary representation cannot exactly store every decimal; for example, 0.1 + 0.2 !== 0.3. section 2.9
- Use parentheses to clarify precedence, strict equality (===) to avoid coercion, and BigInt for integers beyond the safe number range; do not mix number and BigInt values directly. section 2.8 section 2.9
- Use Math.round(), Math.floor(), and Math.ceil() for rounding, toFixed() for formatted decimals, and Math.random() for pseudo-random values in [0, 1). section 2.9
Safe integer boundary
Number.MAX_SAFE_INTEGER + 1 and Number.MAX_SAFE_INTEGER + 2 both produce 9007199254740992. The values are representable, but JavaScript can no longer reliably distinguish nearby integers. section 2.2.1
Listing 2.1 Working with unsafe integers
console.log(`The maximum safe integer is ${Number.MAX_SAFE_INTEGER}`);
const unsafeNum1 = Number.MAX_SAFE_INTEGER + 1;
const unsafeNum2 = Number.MAX_SAFE_INTEGER + 2;
console.log(`The maximum safe integer plus 1 is ${unsafeNum1}`);
console.log(`The maximum safe integer plus 2 is ${unsafeNum2}`);
Run the listing to observe how adding one and adding two beyond the safe boundary produce the same displayed value. section 2.2.1
