Dave Hudson


c8: Implementing integers

2017-01-09 00:00 - 1 min read

The first few commits started with handling natural numbers. Today brings in integers.

Integers are built on top of the natural numbers, but have a “negative” flag that allows them to express negative quantities too. The code is actually very simple as a consequence of delegating all of the numeric operations to an embedded natural number. This same approach will also be used to introduce rational and real numbers later.

Delegation, not inheritance

The choice of delegation over inheritance is important. While number classes might look like they’re very similar, in practice each type of number has subtly different semantics. As an example the subtract operator for natural numbers cannot be used to express values that would be less than zero, and thus throws an exception if a result would be less than zero. With integers there are no exceptions, and so there’s a subtle difference in the semantics. As more complex number classes are added then the semantics vary more wildly (e.g. rational numbers will be expressed as a ratio of an integer and a natural number).

More journal entries