16

There are a lot of interesting snippets to be found online. This post could be found under this (awesome) Stack Overflow question. The author writes the following:

discount ∷ Floating α ⇒ α → α → α → α
discount τ df x = x * (1 + df) ** (-τ)

Are those fancy arrows and dots just a way to make the online page look nicer, or is there an actual Haskell extension (or whatever, I don't quite know the terminology) which would compile something like that? I should note that the usual -> is used in the code just as well.

I have a strong feeling it's not the first time I see things like that.

Zhiltsoff Igor
  • 1,812
  • 8
  • 24
  • 7
    Not specific to Haskell, but this could also be achieved by using a font with ligatures and an editor that supports it (but it will just look this way, not actually use the characters as source) – Andy Feb 08 '21 at 16:12

2 Answers2

26

There's a GHC extension called UnicodeSyntax that allows some Unicode alternatives for certain syntax. However, in general, Haskell source code is written in Unicode, so non-ASCII characters can be used in plain Haskell source code for identifiers and operators, even without any extension.

In the code snippet you include in your question, the author is using both facilities. They are using UnicodeSyntax to allow the Unicode characters ∷, ⇒ and → in place of the built-in ::, => and -> syntax, but they are using the regular Haskell Unicode support to write α and τ for identifiers.

The following program is valid without any extension:

discount :: Floating α => α -> α -> α -> α
discount τ df x = x * (1 + df) ** (-τ)
K. A. Buhr
  • 45,621
  • 3
  • 45
  • 71
  • 6
    Yes, greek letters or infix symbols like `a⚡b = a*2 + b` work without any extension. I personally use `-XUnicodeSyntax` exclusively for being able to write `∀` instead of `forall`. OTOH, the `→` and `⇒` symbols only make type signatures _less_ readable than with `->` and `=>`, in my opinion. – leftaroundabout Feb 08 '21 at 10:20
18

From the GHC docs:

The language extension UnicodeSyntax enables Unicode characters to be used to stand for certain ASCII character sequences.

chi
  • 111,837
  • 3
  • 133
  • 218