The Shape of Functional Programming

The Shapeof

FunctionalProgramming

Saturday, October 5, 13

@fogushttp://www.fogus.me

http://github.com/documentcloud/underscore-contrib


http://www.fogus.me

http://www.fogus.me

http://www.fogus.me

http://www.fogus.me

@fogus

“FunctionalJavaScript”Lemonad

codd.js


Aprogram

written in anobject-oriented

style


Functional programming

for the MEOW!



Pattern 1 - Map(ish)var array = [0,1,2,3,4,5];var transformed = []; for (var i = 0; i < array.length; i++) { transformed.push(array[i] + 100);} transformed;//=> [100, 101, 102, 103, 104, 105]


Pattern 1 - Map(ish)

// Underscore _.map(array, function(n) { return n + 100;}); // Lemonad L.map(L.add(100), array);


Pattern 2 - Filter(ish)var array = [0,1,2,3,4,5];var keepers = []; for (var i = 0; i < array.length; i++) { if (array[i] % 2 === 0) { keepers.push(array[i]); }} keepers;//=> [101, 102, 103, 104, 105]


Pattern 2 - Filter(ish)// Underscore _.filter(array, function(n) { if (n % 2 === 0) { return true; } else { return false; }}); // Lemonad L.filter(L.isEven, array);


Pattern 3 - Reduce(ish)

var array = [0,1,2,3,4,5];var sum = 0; for (var i = 0; i < array.length; i++) { sum += array[i];} sum;//=> 15


Pattern 3 - Reduce(ish)

// Underscore _.reduce(array, function(acc, n) { return acc + n;}, 0); // Lemonad L.reduce(L.uncurry(L.add), 0, array);


APL

life←{↑1 ⍵∨.∧3 4=+/,¯1 0 1∘.⊖¯1 0 1∘.⌽⊂⍵}

0 1 0 0 1 0 0 1 0

0 1 0

0 1 0

0 1 0


Fun.js at a glance

• Simple data

• Many small functions

• Larger abstractions built via composition

• OO to supplement (no dogma)


Simple data


“


#


42


?



[ ]Saturday, October 5, 13

#


#

[1, 2, 3, 4, 5]



{ }Saturday, October 5, 13

#“


#“

{x: 0, y: 1}


{x: 0, y: 1}

#



*Saturday, October 5, 13

Small functions



λSaturday, October 5, 13



Disclaimer


#

Procedure


#

Procedure


function logNth(array, index) { console.log(array[index]);}

logNth([‘a’, ‘b’, ‘c’], 1);// (console) b

#


Purity


Ostrich Purity


Ostrich Purity


#

Function

Kinda importantSaturday, October 5, 13

function nth(array, index) { return array[index];}

nth([‘a’, ‘b’, ‘c’], 1);//=> ‘b’

#





id name age

555-55-5555 Moe 45

777-77-7777 Curly 47

0

1


var table = [

{id: ‘555-55-5555’, name: ‘Moe’, age: 45},

{id: ‘777-77-7777’, name: ‘Curly’, age: 47}

];


id name age

555-55-5555 Moe 45

777-77-7777 Curly 47

0

1


#“


function cell(table, col, row){ return table[row][col];}

cell(table, ‘name’, 0);//=> ‘Moe’

cell(table, ‘age’, 0);//=> 45

#“


Larger abstractions built via composition


LClojure, Haskell, ML, Joy, _


Currying, partial application, fixation, oh

my!


#“


#“

L.fix


#

L.fix(cell, L._, ‘name’, L._);

‘name’


#

L.fix(cell, L._, ‘name’, L._);

var getName = L.fix(cell, L._, ‘name’, L._);

getName(table, 0);//=> ‘Moe’

getName(table, 1);//=> ‘Curly’

‘name’


#“

L.rot


#

L.rot(cell);

“


#“

L.partial


L.partial(L.rot(cell), ‘name’);

#‘name’


L.partial(L.rot(cell), ‘name’);

var getName = L.partial(L.rot(cell), ‘name’);

getName(0, table);//=> ‘Moe’

getName(1, table);//=> ‘Curly’

#‘name’


L.rcurry

“

Codd.col(table, ‘name’);//=> [‘Moe’, ‘Curly’]


“

“

L.rcurry2(Codd.col);


“

L.rcurry2(Codd.col);

var valuesFor = L.rcurry2(Codd.col);var allNames = valuesFor(‘name’);

allNames(table);//=> [‘Moe’, ‘Curly’]


Compositions and Pipelines


How to countall of the values

in a column?


“function countValues(table, tag) { return L.len(Codd.col(table, tag));}

countValues(table, ‘name’);//=> 2

#


L.comp

return L.len(Codd.col(table, tag));

First thisThen thisSaturday, October 5, 13

“Codd.col

#L.len


“Codd.col

#L.len


“

var countValues = L.comp(L.len, Codd.col);

countValues(table, ‘name’);//=> 2

#


L.pipeline

L.pipeline(table, Codd.col(‘name’), L.len);//=> 2

First this Then thisSaturday, October 5, 13

codd.js (demo)Relational Algebra


Thanks

• The NationJS team

• O’Reilly

• Alan Dipert

• Paul Khuong

• The Underscore community

• The Clojure community


The Shape of Functional Programming

Technology