Skip to content

Monovalue Types

A “monovalue type” is a type which can only hold a single value. This means that knowing the type is equivalent to knowing the value, and vice versa. We can convert back and forth between representing it as a type, and as a value, depending on our needs.

We named this concept because it occurs again and again in Au, and the name makes it easier to refer to. The “monovalue” name is based on std::monostate, which has these properties. However, “monostate” could have been confused with the monostate pattern, which exposes actual, changeable, global state to its users. “Monovalue” also emphasizes value semantics, which is a core property of these types.

Identifying monovalue types

A type T is a monovalue type when it fulfills these conditions.

  1. T can be instantiated.
  2. Every instance of T behaves identically to every other instance, in every way, and this behavior does not depend on any program state.
  3. Instances of T support some set of operations with other types.

These properties mean we can freely convert a monovalue object between its “type” and “value” representations. This is a core feature of monovalue types.

The second property also distinguishes monovalue types from the monostate pattern mentioned above.

The third property means there has to be something you can do with the instances.

Examples of monovalue types

Here are some canonical examples in Au.

Type Instance Example Operations
Zero ZERO Comparing to any Quantity
Magnitude<> ONE
  • Equality comparison with other Magnitudes
  • get_value<T>(ONE)
Magnitude<Pi> PI
  • Equality comparison with other Magnitudes
  • get_value<T>(PI)
Radians (and other units) Radians{} (no special pre-formed instance) Arithmetic with other units, such as Radians{} / Meters{}

Switching between types and values

To get the value of a monovalue type T, instantiate the type using T{}.

To get the type of a monovalue type value t, pass it to decltype(t). However, if t is const (including constexpr), you’ll need to use std::decay_t<decltype(t)>.

More details on when to use std::decay_t

It’s common to provide constexpr instances of monovalue types, like the following.

static constexpr auto ZERO = Zero{};

In this case, decltype(ZERO) would be const Zero, not simply Zero. If we tried comparing this to Zero in a type trait, it could fail.

Using std::decay_t is a concise way to avoid this problem. However, it only arises for const or constexpr instances — and only when comparing types for exact equality — so most users won’t need to worry about this most of the time.

This diagram summarizes how to go back and forth using {} and decltype().

flowchart LR

subgraph "#quot;Realm of Types#quot;"
  Type
end

subgraph "#quot;Realm of Instances#quot;"
  value
end

Type -->|"auto value =<br>Type{}"| value -->|"using Type =<br>decltype(value)"| Type