Haskelite language

This page describes the subset of Haskell accepted by the Haskelite interpreter.

Expressions

Integers, characters, strings (i.e. lists of characters), variables, arithmetic operations (+, -, *, div, mod), comparisons (==, <=, etc.), tuples, lists, enumerations (e.g. [1..10]), lambda-expressions, let and where local bindings, case expressions.

Let and where bindings can only define single variables, not patterns.

Data type definitions

Haskell-98 like data and type declarations; no record syntax, existentials or GADTs; no deriving (because there are no typeclasses); no higher-kinded types.

Data types have structural equality and order (similar to derived Eq and Ord); note that comparisions can fail at runtime (e.g. when attempting to compare functions).

Function definitions

You can define functions using multiple equations with pattern matching over integers, characters, lists and tuples, boolean guards, and recursive definitions.

Standard Prelude

even, odd :: Int -> Bool
max, min :: a -> a -> a        -- too polymorphic; may fail at runtime!
compare :: a -> a -> Ordering  
fst :: (a,b) -> a
snd :: (a,b) -> b
(&&), (||) :: Bool -> Bool -> Bool
head :: [a] -> a
tail :: [a] -> [a]
(++) :: [a] -> [a] -> [a]
(!!) :: [a] -> Int -> a 
length :: [a] -> Int
reverse :: [a] -> [a]          -- defined using an accumulator
init :: [a] -> [a]
last :: [a] -> a  
sum, product :: [Int] -> Int
and, or :: [Bool] -> Bool
take, drop :: Int -> [a] -> [a]
maximum, minimum :: [a] -> a   -- too polymorphic!
concat :: [[a]] -> [a]
repeat :: a -> [a]
replicate :: Int -> a -> [a]  
cycle :: [a] -> [a]
iterate :: (a -> a) -> a -> [a]
map :: (a -> b) -> [a] -> [b]
filter :: (a -> Bool) -> [a] -> [a]
foldr :: (a -> b -> b) -> b -> [a] -> b
foldl :: (a -> b -> a) -> a -> [b] -> a
foldl' :: (a -> b -> a) -> a -> [b] -> a   -- strict version of foldl 
(.) :: (b -> c) -> (a -> b) -> a -> c
($), ($!) :: (a -> b) -> a -> b            -- $! is the strict application
zip :: [a] -> [b] -> [(a,b)]
zipWith :: (a -> b -> c) -> [a] -> [b] -> [c]
takeWhile, dropWhile :: (a -> Bool) -> [a] -> [a]
any, all :: (a -> Bool) -> [a] -> Bool
chr :: Int -> Char
ord :: Char -> Int
isAlpha, isDigit, isAlphaNum, isUpper, isLower :: Char -> Bool
show :: Int -> String   -- only for integers

Evaluation

You can force evaluation to weak normal form using bang patterns (!) in variable patterns. Alternatively, to force evaluation to full normal form you can use the following built-in function:

force :: a -> a     -- evaluate to full normal form  

This behaves like identity function, except that it evaluates the argument fully.

Note that force will loose sharing in normal forms, e.g. a pair (x,x) with two identical (shared) values will be expanded into a pair with copies of the normal form of x.

Not implemented

• modules;
• type classes;
• record syntax;
• the numeric tower (Float, Integer, Double, Racional, etc.);
• list comprehensions;
• operator sections;
• declations of new operators;
• monadic operators and do-notation.