Brochure site for PFA
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

While this should go without saying, my putting this on github does not mean you are allowed to wholesale copy the content on my website. Hey, have a blast digging around in my code but copying my content is a no-no. Thanks!

All content (c) PFA INC 2017

Hi! This is a super basic brochure for

Why put my consulting microsite on github? Well, just so you can see how I code small things, I suppose. Also, it's a good way to walk through a few examples.

The site is super basic: one web font; no big fat js frameworks; a very tiny bit of css.

All up the gzipped JS is 50k and the CSS is 1.4k. Which, for a site like this, is how it should be.

Landing page

I like a conversational tone and I felt the fade in worked nicely with that. These are simple CSS transitions and inside the home page javascript object I wait on the transitionend event (vendor prefixed because it is still new-ish.)

$, () => {

Not much to see there. In fact that part of the landing page is simple enough that it is the only stateful part of the site that isn't managed through Redux, though perhaps it could be.

Dropdown menu

This gets a little more involved because I wanted the dropdown menut to work across a variety of inputs: touchscreen, focus/find-as-you-type, and mouseover/click. Hey, that's a lot of state to manage, sounds like a job for unidirectional data.

The thing is, I wanted to keep this site super lightweight. It's a frickin' brochure page, it should load in less than a second. React is great, but if I just need state management I can get away with using Redux on its own.

The actions:

import { setProblems, setIndex, focus, blur, match, go, up, down } from './actions';

problems come from the ul on the home page:

  h2.invisible YES! We have
      input type="text" placeholder="a . . ."
        li data-value="javascript-project" a javascript project
        li data-value="elixir-project" an elixir project
        li data-value="rails-project" a rails project
        li data-value="mentoring" a team in need of guidance
        li data-value="project-rescue" #omghelp!

These are easy enough to grab in the ProblemMenu constructor

const . . .
          $lis        = $menu.find('li'),
          problems    = $lis.toArray().map((li) => {
            const $li = $(li);
            return [ $'value'), $li.text() ];
. . .


The pattern I used with these lightweight view objects was like so:

this.setState(store.getState(), false);
store.subscribe(() => this.setState(store.getState()));

And setState is easy enough:

setState({ autocomplete, menu, input }, shouldRender = true) {
  this.state = {
    isFocused: input.isFocused,
    selectedItem: menu.index,
    inputValue: input.value

  if(shouldRender) {

These views kinda mix up the React notion of props and state but it just doesn't matter for a project this small. If that was important, well, I'd be using React instead of rolling my own ad hoc thing for the grand total of three views that exist on the site.

As you can see at the bottom, a call to render is made when the state changes. That method looks like this:

render() {
  const { isFocused, selectedItem, activeItems, inputValue } = this.state;
  this.$lis.removeClass('selected hidden');

  if(isFocused) {
    this.$lis.toArray().forEach((li, i) => {
      if(activeItems.indexOf(i) === -1) {

    if(selectedItem !== null) {
  } else {


This isn't as pure as a react render method that just returns a bunch of nodes. We mutate the DOM in place and all sorts of other things that would be a no-no if I was worried about other components touching the same part of the page.

Anyway, that's about it. You can see more at problem-menu.js. For touch devices, I instead load the even simpler touch-problem-menu.js


That Easter Egg 🥚

If you let a page sit still without clicking on it for four minutes or so you will eventually see my trusty hound, Henry pop his head up from the bottom left corner.

Heny in all his glory

Lo-Fi Henry in all his glory

I shoved this off on Redux as well. It's a very simple canvas animation with a cubic easing function thrown in to make it look less mechanical.

There are really good libraries out there for managing animations with Redux and React but, again, they were overkill for the purpose.

Most of the logic is in two actions, startEgg and eggTick.

  pixelsPerTick: randSpeed(),
  objectWidth: this.image.width,
  objectHeight: this.image.height

We pass the bounding box, our (average) speed, and our object (henry.png) width.

Insde the startEgg action:

export function startEgg({ pixelsPerTick, boundX, boundY, objectWidth, objectHeight, maxLaps }) {
  const startY          = boundY + objectHeight,
        endY            = boundY - objectHeight,
        totalDistance   = endY - startY,
        totalIterations = Math.ceil(toPos(totalDistance / pixelsPerTick));

First we grab Henry's starting position and his ending position. We're only grabbing the Y axis because he's groundhogging.

Then, once we can figure out his total travelling distance, we can figure out how many times we'll have to "tick" the animation to get him from point A to point B.

  const { timer } = getState("egg");

  if(timer) {

Here we clear any countdown timers. A timer id could be in the state either from the initial countdown, or from a pause between Henry's "laps". More on that when we get to the eggTick action.

    type: EGG_START,
    maxLaps: maxLaps || 1,
    laps: 0,
    x: 0,
    y: startY,
    iteration: 0,
    moveUp: true,


So now we have all of our starting values. The maximum number of times Henry will pop his head up before we reset the values (laps). And of course, we know that he's going to start off by moving up.

export function eggTick() {
  const {
   . . . // destructure the egg state
  } = getState("egg");

  if(doCancel || x === undefined) {

We duck right out of here if we've been sent a cancel signal or our position is undefined (which would be the case with a resetAll).

Next, we have to figure out our direction and increment the iteration counter.

  let nextIteration, nextStart, nextDistance, nextUp, newLaps;

  if(iteration < totalIterations) {
    nextIteration = iteration + 1;
    nextUp        = moveUp;
    newLaps       = laps;
  } else {
    nextIteration = 0;
    nextUp        = !moveUp;
    newLaps       = laps + 1;

  if(nextUp) {
    nextStart     = startY;
    nextDistance  = endY - startY;
  } else {
    nextStart     = endY;
    nextDistance  = startY - endY;

Next, we check to see if we're done with our current run. Take a little breather for the CPU (because bitmap animations are slow), and then start the process again.

  if(newLaps > maxLaps) {
    delayEgg(2000); // nobody wants to hear their fan just because a dog is looking at them.
  } else {

Otherwise we set up the animation loop and dispatch the coordinates of our friendly little companion pup.

    if('requestAnimationFrame' in window) {

      type: EGG_TICK,
      iteration: nextIteration,
      totalDistance: nextDistance,
      moveUp: nextUp,
      laps: newLaps,
      x: x,
      y: easeOutCubic(nextIteration, nextStart, nextDistance, totalIterations)


So there you have it. The code for this stuff:

Konami code

Oh yeah! You can also summon Henry by invoking the hallowed Konami Code: UP, UP, DOWN, DOWN, LEFT, RIGHT, LEFT, RIGHT, B, A.

That was about 30 lines of code, too.

First, for easy comparison I colapsed all of the key codes into a string.

export const KONAMI  = "38384040373937396665";

Then I rigged up the listener:

$(window).on('keydown', (e) => {
  if(e.which === KEY_ESC) {
  } else {
    konami(e.which, { objectWidth: this.image.width, objectHeight: this.image.height });

Had to grab the object bounds to pass in on the call because it's the only thing I couldn't just grab or infer from inside the action, since when we've received all our keystrokes we'll have to actually start the animation.

The action:

export function konami(key, { objectWidth, objectHeight }) {
  const { konami } = getState("egg"),
        nextKonami = (konami || "") + key.toString();

  if(nextKonami === KONAMI) {
      pixelsPerTick: randSpeed(),
      boundX: window.innerWidth,
      boundY: window.innerHeight,
      maxLaps: 2,
      konami: "",

First, there's the easy case that we have the whole code. This is the success case and we start the animation. Otherwise, we check to see if we're on the right track:

  } else if(KONAMI.match(new RegExp(`^${nextKonami}`))) {
      type: KONAMI_CODE,
      konami: nextKonami

In that case, we pass the next konami state and wait to see if we find a match.

Finally, if the current buffer of key codes doesn't match, we reset our buffer:

} else {
      type: KONAMI_CODE,
      konami: ""

Easy as pie. The tests for that:

describe('Konami Code', () => {
  const keys = [ 38,38,40,40,37,39,37,39,66,65 ],
        args = { objectWidth: 320, objectHeight: 320 };


  it('should start the egg', () => {
    keys.forEach(k => konami(k, args));

  it('should reset on a missed code', () => {
    let i = 0;
    for(;i<keys.length/2;i++) {
      konami(keys[i], args);
    konami(99, args);