What Are The Specific Meanings Of The Terms: Functions, Methods, Procedures, and Subroutines?

Question

I'm wondering what are the specific differences in the terminology we use for grouping related parts of code. I've sometimes seen the terms used interchangeably: many OO languages even use the keyword "function" to define a method. (Why?)

If you wanted to be precise, what are the specific meanings of each? Or is it just whatever each language chooses to call it?

Answer 1

• subprogram, subroutine
  "subroutine" probably comes from assembly language. Some processors include instructions to support subroutines as a way of organizing code and reusing common sections of code. For example, the 6502 processor had JSR (jump to subroutine) and RTS (return from subroutine) instructions. I remember it also being used a lot in structured programming, in which a program is a hierarchy of units of code which were sometimes called subroutines or subprograms. IMO, these are the most generic terms for some unit of code to which control is temporarily transferred for the purpose of completing a given task.

• function, procedure
  These are often used interchangeably, but in some languages there's a distinction. In Pascal, a function is a subprogram that returns a value, while a procedure is a subprogram that doesn't. In C and related languages, every subprogram has a return type (even if it's void), so there's no distinction.

• method, member function
  These are two names for the same thing -- essentially a function that's associated with a given class or object.

• operator
  Every language has a set of built-in operators. In some languages, such as C++, operators are functions that can be overridden (i.e. replaced) and/or overloaded (i.e. defined for new types).

• anonymous function
  This refers to a function without a name. Anonymous functions are essentially blocks of code that can be assigned to variables or passed as parameters for subsequent use, for example as a completion routine.

• closure, lambda expression, block
  A closure is a chunk of code that's bound to a set of variables. I think of closures as anonymous functions plus context.

Answer 2

It depends on the languages, what you can do with them

• do you need a distinction
• is there a distinction in the syntax, or is it just convenience, when talking about code

AFAIK, in Basic, if you call something, which does not return anything, but performs a job with side effects, like printing, write to file or change a global variable, it was called a procedure.

If it returns something, it was called a function.

In an OOP-Language like Java I rarely met the term 'function', but 'method', and it was teached that functions are things which are global and can be called by everyone.

In the functional and OOP-language scala, a method isn't called function, but you can pass such a thing to another method, and then it becomes a function.

A subroutine is a routine, which is called by another routine, to my understanding.

In summary, I don't think there is a concise definition over all languages and more so paradigms. You have to consider the specific culture, you're in, if you use these terms - maybe you should introduce your definition first.

Answer 3

I'd say depending on context, all those words could mean the same thing or wildly different things, for example what Caleb has written (but there is more, i.e. languages where operators are first-class functions etc.). So, it's hard to give a specific definition, and wise to not assume too much unless the context (programming language) is known.