Project Eweda

A (probably im-)practical functional library for Javascript programmers. Eweda lamb!

This turned out to be more of an academic exercise than something practical. But there is a parallel library that is more practical in Ramda.

Goals

Using this library should feel as much like using Javascript as possible. Of course it's functional Javascript, but we're not introducting lambda expressions in strings, we're not borrowing CONsed lists, we're not porting over all of the Clojure functions.

Our basic data structures will be normal Javascript objects, and our usual collections will be Javascript arrays. We will not try to reach the point where all the functions have only zero, one, or two arguments. We will certainly try to keep some of the normal features of Javascript that seem to be unusual in functional languages, including variable length function signatures and functions as objects with properties.

Functional programming is in good part about immutable objects and side-effect free functions. We will stick to that as much as feasible, but will not be dogmatic about it.

As much as we can, we would like the implementation to be both clean and elegant. But the API is king: we will sacrifice a great deal of implementation elegance for even a slightly cleaner API.

Documentation

Annoated source code documentation generated by Docco is at https://rawgithub.com/CrossEye/eweda/master/docs/eweda.html. So far, that's the only documentation available.

The Name

Ok, so we like bad puns. That's all. It's a short name, not already taken. It could as easily have been ramda, but then we wouldn't be able to say " eweda lamb! " For non-English speakers, lambs are baby sheep, ewes are female sheep, and rams are male sheep. So perhaps eweda is a grown-up lambda... but probably not.

Structure

Replaceable Core

Eweda is designed in good part around manipulating lists of objects. But the implementation of lists is meant to be replaceable. By default, eweda uses a simple implementation using arrays, but the exposed object is a function that accepts a different implementation of lists and returns a new version of the API built around those lists. In test/test.bootstrapping.js, we explore a detailed example built around Scheme/LISP-style lists, as described in buzzdecafe's gist.

We can construct a new implementation of the API like this:

var cons = function(x, y) {return function(pick) {return pick(x, y);};};
var car = function(fn) {return fn(function(x, y) { return x; });};
var cdr = function(fn) {return fn(function(x, y) { return y; });};
var atom = function(x) {return (x !== null) && (x !== undefined) && !x.pair;};
var isEmpty = function(list) {return list == null;},

// These core properties need to be supplied.  Others can also be supplied to override the defaults.
var ramda = eweda({
    EMPTY: null,
    isEmpty: isEmpty,
    prepend: cons,
    head: car,
    tail: cdr,
    isAtom: atom
});

We can make the functions here global by calling

ramda.installTo(this)

in a global context, and then with these functions defined:

var square = function(x) {return x * x;};
var even = function(n) {return (n % 2) == 0;};

We can now call

map(square, filter(even, range(0, 15)));

to get something equivalent to:

(0 4 16 36 64 100 144 196)

which, under the hood, is the same as

cons(0, cons(4, ... cons(144, cons(196, null)) ... ))

Automatic Currying

The functions included should automatically allow for partial application without an explicit call to lPartial. Many of these operate on lists. A single list parameter should probably come last, which might conflict with the design of other libraries that have strong functional components (I'm looking at you Underscore!)

The idea is that, if foldl has this signature:

var foldl = function(fn, accum, arr) { /* ... */}

and we have this simple function:

var add = function(a, b) {return a + b;};

then, instead of having to manually call lPartial like this:

 var sum = lPartial(foldl, add, 0);
 var total = sum([1, 2, 3, 4]);

we could just do this:

 var sum = foldl(add, 0);
 var total = sum([1, 2, 3, 4]);

Functions included

We want to include the basic functions that will help a Javscript programmer work with objects and arrays. We will try to use the most common names for these, possibly using multiple aliases for those that are most debated.

Core

isEmpty
prepend
append
merge
head
tail
size

Functions

compose (fog)
pipe (sequence) (i.e., like compose but in reverse order)
flip
partial (applyLeft)
rPartial (applyRight)
memoize
once
wrap

Lists

map
foldl (reduce)
foldl1
foldr (reduceRight)
foldr1
filter
reject
take
skip (drop)
find
all (every)
any (some)
contains
uniq
pluck
flatten
zip
zipWith
xprod (i.e. cartesian product)
xprodWith (i.e. cartesian product with function)
reverse
range

Objects

tap (K)
eq
prop
func
props
identity (I)
alwaysTrue
alwaysFalse
alwaysZero
maybe
keys
values

Logic

and
or
not
andFn
orFn
notFn

Arithmetic

add
multiply
subtract
divide
sum
product

CrossEye / eweda