Challenge 2: Combining Three Independently-Validated Fields — Possible Solution
====================================================================

Main.hs:

data Address = Address { street :: String, city :: String, zip_ :: String }
  deriving (Show)

validateStreet :: String -> Maybe String
validateStreet s = if not (null s) then Just s else Nothing

validateCity :: String -> Maybe String
validateCity c = if not (null c) then Just c else Nothing

validateZip :: String -> Maybe String
validateZip z = if length z == 5 then Just z else Nothing

mkAddress :: String -> String -> String -> Maybe Address
mkAddress s c z = Address <$> validateStreet s <*> validateCity c <*> validateZip z

main :: IO ()
main = do
  print (mkAddress "123 Main St" "Springfield" "12345")  -- all valid
  print (mkAddress "123 Main St" "Springfield" "123")     -- bad zip

Output:

    Just (Address {street = "123 Main St", city = "Springfield", zip_ = "12345"})
    Nothing

Explanation:
Address <$> validateStreet s applies fmap first, producing a Maybe
holding a PARTIALLY-applied Address constructor (still needing city
and zip). Each subsequent <*> supplies one more independently-
validated field. If ALL THREE validations succeed, the fully-applied
Address ends up wrapped in Just. If ANY ONE of them returns Nothing --
here, the zip code "123" fails its length check -- the whole chain
short-circuits to Nothing immediately, regardless of whether the
other two fields were valid.

WHY THIS WORKS AS AN ANSWER
------------------------------
This builds a genuinely three-field applicative chain (matching the
chapter's own two-field Person example, extended by one field),
testing both the fully-successful case and a case where exactly one
field fails, confirming the whole result becomes Nothing.