OK, so: every day, billions of people all over the world use their phones and computers to work, play, and connect with each other. If you’re (presently) typically-abled, it’s very easy to assume everyone has the same experience with technology you do. But that’s an illusion; over one billion people (16% of the the world pop.!) live with some form of disability, and those disabilities affect whether or not someone has the same access to technology as you do.

Plus, it’s really expensive if you don’t make your software accessible and then, like, get sued.

At Stark, it was my job to help answer this question we’d given ourselves — how do we ensure the people making software used every day by billions, make software that is usable for everyone?

In the pursuit of the answer, I did stuff like:

• contribute to design and development across a bunch of product touchpoints: web, desktop, plugins… you name it, I put my lil’ fingers in it

• build our core accessibility engine that quickly finds and suggests solutions to issues in design files. (This engine powered 1000s of accessibility audits a day at companies like NASA, Google, and Microsoft. Did I say those names already?)

• co-lead the development of Stark for Mac. Stark for Mac was built with 100% free-range SwiftUI at a time when there were few other apps doing that, which meant that if you found a bug, or tried to do something slightly unusual you were exactly what we’ll all be when we die: alone.
  
  That got darker than I meant for it to be, but you get the idea.
  
  Stark for Mac’s goal was to make it easy for designers to find accessibility issues in their designs before they launched them. Eventually we realized that experience needed to also be available on the web, which is where we eventually ported it (and I also helped build.)

I built UNSLEPT with a buddy of mine, Tregg. (Who, by the way, happens to be one of my favorite people, a great dad, product designer and person in general.)

While “What if there was a book club, except instead of books… it was metalcore?” sounds like the sort of question someone asks in a sketchy college dorm room after smoking ungodly amounts of weed, it did form the premise of UNSLEPT. (Yeah we spelled that in all caps, baby. It was edgier that way. Also, sorry, I can’t promise no weed was involved.)

Sign up, give us your phone number, and once a week we’d send you a new album to listen to. Maybe you love it. Maybe you hate it. Maybe you want our address so you can firebomb us: too bad! Our HQ was in the most secure and impenetrable place on earth: the bunker where they keep the Epstein files. Instead, listen, then take whatever emotions the music filled you with out on your fellow listeners in Discord.

UNSLEPT is currently on haitus… but who knows? Maybe it’ll be back one day, reborn as a music-based AI-powered dating app. Just kidding. We’re not that high.

Yet.

In its early days, I used to be fairly involved in the Sketch plugin development community. I would use Sketch pretty frequently to design websites and apps, and it was super convenient to be able to just… write a plugin to make whatever the current design task I was working on easier.

Back then, Sketch didn’t have autolayout (the ability to automatically position and size layers) like it does now. You had to manually position and resize rectangles like some kind of caveman. Meanwhile, in the Land of Code, systems for laying out elements and enforcing consistent sizing and spacing were super common, but — at the time — missing from most design tools. So it came pretty naturally to me to want the same sort of system and tools when designing. (Fun fact: internally, Fluid literally reused Apple’s Auto Layout framework to calculate where to place and how to size design layers, because I am lazy efficient.)

Turns out, Cat Noone (another one of my Favorite People) had the same idea. Cat tweeted about a similar concept; I replied “bet?” and the rest is history. We launched Fluid on Product Hunt to 700+ upvotes and were ranked #2 of all product launches that day, with 2.4k stars on Github. Shortly thereafter Obama gave us the Nobel prize. I’m lying about that last part — that happened in my dreams — but the rest is true.

These days, suspiciously, both Sketch and Figma have autolayout-like features baked-in. Now, I’m not saying we gave them the idea, but I’m also not not saying that, you know?

But I am saying, like, maybe, just maybe, Cat and I deserve royalties.

Growing up I would read a lot. (This is also an area where I was probably somewhat insufficiently supervised. Not that I was reading anything, like, particularly indecent or something. Well. Actually, there was that time I tried to read Flowers for Algernon and later Twilight (not linking you to that one, you know what it is), both before my precious little mind, or body, was ready.)

My reading started like all gateway drugs do: with libraries. My mom would take my sisters and I to the local library and let us run wild (bless you, mom). Eventually my parents relaxed this habit and weren’t always available to take us, and I also started wanting to read things without having to go through the living, breathing communist censor that was my parents.

This meant my reading habit went digital. I’d just gotten an iPod Touch (which I won in an online giveaway, by the way. This was at a time when God was Real, Nigerian princes actually did want to give you $50k, hot singles were not just in your area, ma’am, but in your backyard, and there was enough money flowing through the internet to fund companies that gave things away for free) and did all of my reading on it.

Steve Bezos killed reading for me. In 2009, Amazon purchased Stanza, the app I used to do all of my reading. A few years later, after letting Stanza die a slow, painful death of insufficient maintenance, Amazon finally took Stanza behind the App Store and put it out of its misery.

Goddamit, Steve.

Luckily… I had been programming a fair bit by that point. I’d never built or designed anything as complex as full-scale mobile app… but, because I was then young and delusional (unlike now, where I’m older and even more delusional), I thought: why shouldn’t I try to just… replace it?

And so I taught myself Objective-C, learned what Xcode was (and that you shouldn’t spell it XCode, or else Strangers on The Internet would yell at you. Strangers on the internet yell a lot, it would turn out), copywriting, mobile design, developed opinions about tabs and spaces, and most importantly, a crippling sense of social anxiety. All of the things a young virgin needs for a successful start in life.

In 2015, after worrying my parents with how much time I was spending on my laptop, I released Hyphen. Hyphen ended up getting a decent amount of buzz and positive reviews in several publications. And while it did earn me a non-zero amount of money, it also taught me something even more important: that if you make something people like, sometimes they’ll say nice things. Most of the time they won’t say anything. But if, god forbid, you write a bug, or forget that one feature someone needs? People you didn’t know existed will make themselves known, and they will not say nice things about it. And that you’ll develop a weird sort of relationship to this dynamic… you’ll start to actually like thinking about features, and designing stuff to solve peoples’ problems, even if they’re not grateful for it, because sometimes they will be? Maybe? And this is definitely healthy, right?

Anyway I learned nothing from that experience and kept doing it.

(Hyphen, for the record, is no longer actively maintained. But who knows… maybe that’ll change, some day :))

Zemel is an outgrowth of my work on Hyphen. Hyphen needed to be able to parse lots of XML fast, which is exactly what Zemel does. Zemel features a number of stuff that cuts down on the time and memory you have to spend to parse lots of XML. Unlike traditional XML parsing libraries, Zemel lets you do things like quit parsing early, parse in chunks, and extract information from an element as soon as it is parsed, rather than waiting until the entire document has been parsed.