IdleComp

A composable way to perform non-blocking computations in JavaScript.

IdleComp is based on @philipwalton’s Idle Until Urgent article and provide you a simple and composable way to run code on your web application without blocking user input. The only thing you have to do is follow one simple rule.

Write small simple functions and compose they together.

Why?

Traditionally, web browsers javascript runs on the same single thread as other page’s tasks like painting and parsing. Meaning your code will either take long enough to blocks those tasks or be sort to, hopefully, not interfering on user experience. While we do have other alternatives like workers, they usually are very limited and can’t manipulete the DOM tree.

IdleComp takes advantege of page’s idle times to executing sort tasks, providing an easy interface that allows for composition.

Composition

A insteristing property of IdleComp is that doing this:

IdleComp
  .of(5)
  .map(increment)
  .map(double)

Is equivalent to this:

IdleComp
  .of(5)
  .map(x => double(increment(x)))

This mean that calling subsequents maps is exactly like composing functions!

Usage

To install:

npm install --save idle-comp
#or
yarn add idle-comp

To create a new IdleComp object, you should use IdleComp.of and pass in your initial value.

import IdleComp from 'idle-comp'

const idleFive = IdleComp.of(5)

Now idleFive is an Object with to methods: map and returns.

map is how we’re going to do our idle computations.

idleFive
  .map(five => five * 2)
  .map(ten => ten - 1)
  .map(console.log) // 9

As we’re computing only when the browser is idle, this also means we’re delegating your computation to some time in the future - Asynchronous.

But sometimes you will need the value rigth away, even if blocking. This is when returns kicks in.

returns will execute all pending computation and returns the final result.

const idleNine = idleFive
  .map(five => five * 2)
  .map(ten => ten - 1)

//Right now idleNine isn't resolved yet, let's force all computations synchronously

console.log(idleNine.returns()) // 9

Example

First, lets define a dragons array.

const dragons = [
    { age: 2, name: 'Halph' },
    { age: 5, name: 'Pottus' },
    { age: 3, name: 'Traus' },
    { age: 1, name: 'Nelf' },
    { age: 4, name: 'Gart' },
    { age: 7, name: 'Mange' },
    { age: 6, name: 'Zalu' }
]

Now a logging helper

// Just log x and then return it
const log = x => {
  console.log(x)
  return x
}

Now lets sort the dragons by age in descending order, get the last and shout it’s name

const idleName = IdleComp
  .of(dragons) // Bring our dragons to the idle realm
  .map(dragons => dragons.sort((dA, dB) => dB.age - dA.age)) // sort them by age
  .map(dragons => dragons[6]) // get the last
  .map(log) // log out or dragon and return it
  .map(lastDragon => lastDragon.name)
  .map(name => name.toUpperCase())

console.log('First me')
console.log('Than me')

const name = idleName.returns() // Forces all remaning idles to run synchronously

console.log(name)

idleName
  .map(name => name[0]) // resumes the chain
  .map(firstLetter => firstLetter + 'ICE')
  .map(log)

console.log('end of file')

When this example is ran, we got the printings

First me
Than me
{ age: 1, name: 'Nelf' }
NELF
end of file
NICE

As you see, the fisrt two console.logs are executed first and the executation of the first .map(log) (as well as the entire map chain) is deferred until we explicit request the value with .returns().

As we resume the mapping chain, we defer the rest of the execution to either the next .returns() or the next iddle slice of time, whatever comes first.

Roadmap

Those are features that are on our backlog.