Question

Why would one use a functional language in an otherwise Imperative project?

Answer 1

If your project is truly imperative, you probably don’t want a purely functional language. But you probably still want a language with functional features; functional style addresses low-level code structure in the same way that object-oriented style addresses high-level structure. Both allow you to package certain common patterns in a language-supported way.

In a primarily imperative project, functional style is useful at the expression and statement level, allowing you to abstract common loops and sequences:

For example, take this common pattern:

newlist = []
for x in oldlist:
    y = dosomething(x)
    newlist.append(y)

That’s map:

 newlist = map(dosomething, oldlist)

Or this:

total = 1
for n in numbers:
    total = total * n

Becomes fold (also known as reduce):

total = fold(*, 1, numbers)

Imperative style does not address this low-level duplication all that well–hence the “I wish I had a nickel for every time I typed for(int i = 0; ...)“. Even in OO languages without functional features, code inside methods doesn’t differ much from similar non-OO languages.

Some IDEs for address this by providing code snippets. This addresses the lack of abstraction power in the wrong way. The way to handle a repeated pattern is not to encourage cut-and-paste with little holes for variable names, but to abstract the pattern into a reusable unit.

Note: I addressed embedding functional code in an imperative project. A top-to-bottom project in functional style will look different. Here are some links taken from similar Stack Overflow questions:

• http://www.25hoursaday.com/weblog/2008/06/16/FunctionalProgrammingInC30HowMapReduceFilterCanRockYourWorld.aspx
• http://www.joelonsoftware.com/items/2006/08/01.html