Function types

…

Within a list of parameters or results, the names (IdentifierList) must either all be present or all be absent. If present, each name stands for one item (parameter or result) of the specified type and all non-blank names in the signature must be unique.

There are several things to unpack here. Let’s start with the most intuitive part: The names of function parameters and results must be unique.

func(a int, a string) // Invalid: can't have two parameters with the same name
func(a int) (a int)   // Invalid: can't have a parameter and result with the same name

Working our way backwards through the paragraph, we next come to the idea of a blank parameter or result. What’s that?

Remember: parameter and result names don’t affect the function type at all. They serve merely as documentation. If you don’t feel the need to provide a name for the sake of documentation, you may omit the name by using the blank identifier:

func(a int, _ string)                  // The second argument is ignored by the function, but must be provided by the caller
func(a int) (_ int, err error)         // The first result has no name, and is not instantiated when the function is called.
func(_ int, _ string) (_ int, _ error) // You may use the blank identifier for all parameters and results

But that last example is a bit overly verbose. So the we come finally to the idea that the names may be entirely omitted. But they must all be omitted in that case. The following are all valid representations of the same function type.

func(a int, b string) (c int, e error)
func(a int, _ string) (_ int, err error)
func(_ int, _ string) (_ int, _ error)
func(int, string) (int, error)

Quotes from The Go Programming Language Specification Version of December 15, 2022