The numeric ladder

The ladder

| Type | Size | Range / use | Literal | |---|---|---|---| | int | 32-bit | whole numbers ±~2.1 billion | int n = 42; | | long | 64-bit | big counts, timestamps | long t = 9_876_543_210L; | | double | 64-bit | general decimals, science | double d = 0.5; | | decimal | 128-bit | money, exact base-10 | decimal m = 19.99m; |

Two traps the ladder hides:

1. Integer division truncates. 7 / 2 is 3, not 3.5 — both operands are int, so the result is int. Write 7 / 2.0 (one operand makes the expression double) when you want the fraction. This is the most common numeric bug in beginner code.

2. double is base-2; money is base-10. double x = 0.1 + 0.2; gives 0.30000000000000004. Binary floats can't represent most decimal fractions exactly — fine for measurements, wrong for money. Use decimal (note the m suffix) for currency:

decimal price = 19.99m;
decimal total = price * 3;    // 59.97 exactly

Overflow is silent by default: int.MaxValue + 1 wraps around in unchecked context (the default for constants is checked at compile time, but runtime arithmetic is not). Don't reach for checked yet — know that the boundary exists and that long is the escape hatch when counts can exceed ~2.1 billion.

Conversion rules: small → big is implicit (int → long → double); big → small needs an explicit cast (int), which truncates ((int)3.9 is 3). Casting double → int also throws away, never rounds. Use Math.Round first when you need rounding.