# Number and number literals

[Syntax and values](README.md)

Numbers describe amounts and positions. A **literal** is a value you write directly, such as `42`. Pliro stores numbers with limited precision, so some fractions are approximate. The [rounding explanation](../concepts/floating-point.md) shows why `0.1 + 0.2` can surprise you.

Number literals use ASCII decimal digits with an optional fractional part:

```text
0
42
3.14
```

The decimal point requires digits on both sides. A sign is a separate unary
operator:

```text
-12
+3.5
```

The following forms are not supported:

```text
.5       # use 0.5
1.       # use 1.0
1e6      # no exponent syntax
1_000    # no digit separators
0xff     # no hexadecimal literals
0b1010   # no binary literals
```

At runtime, every `Number` is a finite IEEE 754 binary64 value. NaN and infinity
are not successful values.

Numbers use binary [floating-point](../concepts/floating-point.md). Familiar floating-point rounding can occur.
Results must stay finite. Number display uses a short round-trippable decimal
form and does not add `.0` to whole values.

Indexes, collection sizes, random bounds, and GPIO pins impose additional
whole-number and range requirements.

## Example

```pliro
# language: en
say 3.5 + 2
say(-8 / 2)
```

## Related entries

[Syntax and values](README.md) · [Built-ins](../builtins/README.md) · [Programming guides](../guides/README.md)