Giving Names to Numeric Literals [VU1]

Numeric Literals That May Be Self-Documenting in Limited Circumstances

Replacing numeric literals with well-named constants is a good practice that helps eliminate doubt in the meaning of numeric literals. However, there are a few circumstances where direct usage of numeric literals may be self-documenting enough to be acceptable:

• Using 0 as the starting index for loop iteration or another initial value for a variable.

• Using 0 to check if something's empty (or similar greater/less than comparisons).

• Using 1 to check if something's populated with a "single" value (or if a "count" of something is specifically single).

• Using 1 in increment/decrement operations.

• Using 2 to double or half something.

When indexing into a container/collection, it's still often better to have a constant over a magic number (even if just accessing something like the first element in a container/collection). Such a constant can communicate the intent of the element you’re trying to access.

Even if a situation falls underneath one of the "acceptable" cases above, it still might be better to use a named constant. For example, consider the following:

The meaning of the + 1 might be recognizable to many C programmers, but even then, the intended meaning is uncertain. The following introduces an additional named constant to clarify the intent:

So even if using a numeric literal might be acceptable in some cases, think very deeply if there is a name you could assign to a literal that would communicate problem domain understanding.