Arrow notation specifies a class of data using a combination of other classes of data. In this sense it is analogous to the "listof" notation in some programming languages. Symbolically, the arrow notation is represented by an arrow symbol.

The specification of a data class involves the creation of so-called direct data definitions, actual aspects of the data as opposed to a name. More concretely, the arrow notation specifies the types of arguments that define the data and the kind of value it produces

An arrow notation serves several useful functions. In some programming languages, like Perl, the use of the block notation to describe an item can be difficult to read. Arrow notation can make the description crisper and more easily interpretable. For example, a block notation reference description could appear as:

• @{array_ref}[5]{Name}

The alternative arrow notation, conveying the same information, would be:

• $array_ref5]

Arrow notation is also useful for large numbers. An up-arrow notation system for large numbers was devised by Donald Knuth in 1976. The number of up arrows corresponds to a certain value. As an example, in Knuth's system a googol, which is 10 to the power of 100 is represented as 10 10 2.