CODEWITHSUNDEEP
haskell
Paul Nathan
Paul Nathan

Reputation: 40319

How can I ensure a function is fully applied?

In the following code, the where value= clause returns a partially applied function.

How can I ensure it gets fully applied?

------------------------------
future_value :: Float -> Float -> Float -> Float
future_value present interest periods =
  (present * (( 1 + interest) ** periods))

------------------------------
-- Given an initial amount
-- Given a yearly fee, as well as a yearly interest...
-- Calculates return over a number of years, given a yearly fee.
return_over_time :: Float -> Float -> Float -> Float -> Float
return_over_time present interest num_years fee =
  if num_years == 1 then (future_value  present interest 1.0) - fee
  else future_value value
       where value =  return_over_time present interest (num_years - 1) fee

Upvotes: 2

Views: 155

Answers (2)

Landei
Landei

Reputation: 54584

If it makes no sense that a function is ever partially applied, you can change it to take a tuple:

future_value :: (Float, Float, Float) -> Float
future_value (present, interest, periods) =
   (present * (( 1 + interest) ** periods))

Using this definition error messages get much more "localized", the compiler will complain about passing a (Float, Float) instead of (Float,Float,Float) or so, and not about something fancy like a missing Show instance for (->). However this is just a little trick for beginners, as for most functions currying makes a lot of sense, and sooner or later you get used to it and learn to fix errors resulting from missing arguments quickly.

Upvotes: 1

Tikhon Jelvis
Tikhon Jelvis

Reputation: 68152

The issue isn't with value but with future_value. future_value takes three arguments but you only give it one (value), so future_value value has the type Float -> Float - > Float. value, on the other hand, is just a Float because return_over_time is fully applied in the where clause.

You can make sure future_value is fully applied by passing in two more floats for interest and periods.

Incidentally, is there any particular reason you are using Float instead of Double?

Upvotes: 4

Related Questions