Referential transparency

Referential transparency is an oft-touted property of (pure) functional languages, which makes it easier to reason about the behavior of programs. I don't think there is any formal definition, but it usually means that an expression always evaluates to the same result in any context. Side effects like (uncontrolled) imperative update break this desirable property. C and ML are languages with constructs that are not referentially transparent.

As an example, consider the following program in Standard ML:

 print "h"; print "a"; print "h"; print "a"

which prints "haha". In an attempt to factor out the repetition, we write

let val ha = (print "h"; print "a")
in  ha; ha end

but now the laugh is on us, because "ha" is only printed once. The reason is that print's side effect is realized when ha gets bound, so we should have written

let fun ha () = (print "h"; print "a")
in  ha (); ha () end

Haskell's monadic I/O system distinguishes between values and actions like the print procedure above. So we do indeed have that

putChar 'h' >> putChar 'a' >> putChar 'h' >> putChar 'a'

is equivalent to

let ha = putChar 'h' >> putChar 'a'
in  ha >> ha

---

This example is taken from

• Philip Wadler. How to declare an imperative. ACM Computing Surveys, 29(3):240--263, September 1997. [1]

---

Some definitions of referential transparency are summarized in a USENET post by Tom DeBoni [2]