How do I decide what a function should return?

Start with the caller’s next step: what does it need from this function to continue?

1. If it needs one result, return that value.
2. If it needs a few results that belong together, consider returning them together.
3. If it only needs the function to do something, and has no result to use, void may fit.

In C#, a method declares a return type or void; a non-void method returns a value, while a void method can perform work without returning one. C# also supports returning related values as a tuple, including named elements. [S1] These are available choices, not rules that decide the best design for every method.

Use one return value when one answer is the natural result of the operation. For example, suppose a method computes a subtotal from a price and quantity:

static decimal CalculateSubtotal(decimal price, int quantity)
{
    return price * quantity;
}

The caller can use the result directly:

decimal subtotal = CalculateSubtotal(12.50m, 3);

The hypothetical values here are just an illustration. The method’s stated purpose and its returned value line up: it calculates a subtotal, and the caller receives that subtotal. If a caller also needs some other distinct piece of information, don’t squeeze it into the subtotal or make it guess what the result means. Decide whether that information is genuinely part of the same result.

A tuple can return multiple related values, and named tuple elements can make the parts easier to identify at the call site. [S1] Use that option when the values describe one coherent result and the caller is likely to need them together. For example, if a hypothetical calculation reports both a subtotal and the quantity used:

static (decimal Subtotal, int Quantity) CalculateSummary(
    decimal price, int quantity)
{
    return (price * quantity, quantity);
}

var summary = CalculateSummary(12.50m, 3);
Console.WriteLine(summary.Subtotal);
Console.WriteLine(summary.Quantity);

Here the values are named rather than left as unexplained positions. But first ask whether Quantity is useful output or merely an input the caller already has. If the caller already knows it and does not need it returned, a single subtotal may be simpler. The documentation establishes that C# supports tuple returns; it does not establish that tuples are always preferable, or prescribe which values belong together. [S1]

Use void when the method’s job is an action and the caller does not need a result from it. For example, a method that displays a supplied message could be:

static void ShowMessage(string message)
{
    Console.WriteLine(message);
}

The caller asks it to perform the action; there is no computed message value to assign. C# void methods can perform work without returning a value, and a bare return can stop their execution. [S1]

For a save operation, decide what the caller needs to know afterward. If it needs a result, choose and define that result rather than returning an arbitrary value. If it only needs the action performed, void can be appropriate. The documentation describes the language’s return options, but does not specify what a particular save method should report or how saving should work. A small decision checklist helps keep that choice tied to your caller’s needs.