Build Things That Don’t Scale at All

So my talk is about building things that don't scale at all. So how does this story start? It starts with you wanting to use an app that does something a very specific way. So what do you do? You bring out your phone. You launch the app store. You start searching for the app. Let's say it's a to-do app because we always need new to-do apps all the time.

So what do you do? You scroll. You see all these great apps, five star reviews, tens of thousands. How do you choose? I don't know. You just pick one, the one that looks the best, I guess. You install it, you launch it, and you're greeted with these beautiful system dialogues. I don't want to be tracked. I don't need notifications from a to-do app. Camera.

So, well, eventually you're done, and you start with this beautiful onboarding flow that some marketing guy said it's nice to have and be as long as possible because in the end it's going to probably convert better when you hit the paywall. And you're like, Jesus, seriously. And then there's GDPR. So maybe it's not that bad, but maybe it's just an app that doesn't do what you want exactly your way.

So you, being a developer, what do you think? You know what? I'm going to build my own to-do app. And so then you start thinking, oh gosh, but then I need to deal with distribution. So the app stores, and that's always a headache, such as authentication is also a headache. And then I need to support iOS and Android. But not only that, also phones and tablets. And why not? Even dark mode and multiple themes, whatever.

So at that point you're like, oh, God, I'm investing so much into this thing. I might as well build something that scales actually. And so you need to start thinking into analytics, maybe because you're investing that much time, you need to start thinking about monetization. And if you want your app to be visible, you need to think about SEO. And then maybe you expand to other markets and localize your app. And so you need to start thinking about competition.

And then why not? Let's just build a pipeline for my app. Well, I might as well add over the air updates. And you can go on as much as you want. You can dig deeper and deeper. And it never ends.

So this is me. I'm a full stack developer. I was recently hired by Zatu, and I'm working as a principal engineer. So usually I do care a lot about everything that I said, and code quality, and testing, and everything. But what this talk is about is that you don't always need to follow these best practices and care about every single minute detail. But it's fine and it's fun to actually build something that doesn't scale at all.

So I have a few examples of things that I've built this way that I think are interesting. So the first example is this app that I call Ciuttino. What this does, it just takes screenshots and automatically- I think there are some Italians here. It's a joke in Italian. It takes screenshots and it automatically crops them. Why did I build this? Because Twitter didn't let me share images. For whatever reason. If you share a link to an image, it will share the tweet and not the image.

So I built this incredibly complex algorithm that all it does, it takes these two points, look at the color, iterates towards the center, and tries to figure out what the crop should be, and it gives you handles so that you can fine-tune it.

And the other thing that this does is it streamlines sharing. So I configured it so that only when you hit the little checkmark, it shares the picture with one contact on WhatsApp. That's it.

And this then features one target, so it's Android phone. I mean, Android phone, my phone, because I'm one user. That's me. One locale. It wasn't actually in Italian, but I didn't want to choose between the American flag and Union Jack, so let's pretend it was in Italian. There's no authentication whatsoever, of course. And distribution was simply installing the app, plugging my phone via USB to my computer. And when I needed to get updates, I would plug it again. This took me all of one day, and I've been using it for years. Now I stopped it because this person that I was sharing all these pictures with just told me to stop. And also, BlueSky actually fixed it. So in BlueSky, you can actually share pictures. So that's great.

Second example, Kaminora. Kaminora is a private chat app. So my daughter asked me for a chat app to chat with her friend. This was a couple of years ago. So we didn't want the girls to have phones, because there were eight. And we didn't want them to use just some random regular chat app where they could find, meet whatever random stranger. So I decided, well, why not? Let's just build an app for them.

So this is a library that's called React Native Gifted Chat that does everything for you. What it doesn't do is push notifications. But this is an expo app, so I could use expo push notifications. So I didn't even need to deal with Firebase or Apple push services. And the other thing that this app does is it has a little timer. Because these were not the girls' phones, so we didn't have any parental control.

We would just hand them our iPads. And so this keeps track of how long you've used the app. And after a half an hour, you need to give a password that mommy and daddy know. Or it would just reset itself automatically the day after. So this app uses Expo Go.

In case you're not familiar with Expo Go, it's like a launcher, let's say, for Expo apps. So you only write JavaScript. There's no native code involved at all. And this thing will contact a server, get the JavaScript bundle, and run it and launch it for you.

So this thing got me thinking, hm. So this thing actually connects to a server and gets a bundle every time. So why don't I put this server on a server that I host on Headset for five years a month and just have it work in that way? So every time the app launches, it gets the bundle. So it auto-updates all the time. It's probably not the safest, but hey, it works. It's been working for years. And actually, my daughter came back to me yesterday and told me, oh, it doesn't work anymore. And I fixed it. And now it works.

So this is also one target iPad, two users, the two Go's, one locale because they chat in German all the time. Authentication was me calling the other parents and telling them, hey, type this into the app and it will log in this way forever and you don't have to care anymore. And to install it, Expo Go lets you share QR codes. I send them a QR code over WhatsApp. They install it. Happy story. And this took me a weekend. So I didn't have to think of all those issues. And yeah, that makes you very fast.

Last example. Uncool. It's a Google calendar synchronizer. So the story here is that we have many Google calendars shared with my wife and it's always a mess. And so I wanted a way to log in with Google, get your calendars. You select which ones. In case you've developed React Native apps, this is like the default template. So I didn't spend that much time into the layout. But then you select which calendars you want to synchronize. And it just gives you a very, like this is what I wanted. This is the view that I wanted today, tomorrow, coming up. And I want to add some more stuff.

This is using Expo Development Builds, which is the new, I mean, no longer new, but it's the default way of developing React Native apps when you have some native code. So Google sign-in requires you to have native code. So you're kind of forced into having this Expo Development Builds. So what this is, is the Expo Go thing that I showed you earlier within your app. And the nice thing about this is that it still connects to a server, but it also lets you have a launcher, which Expo Go didn't let you. And it worked with push notifications because pressing the push notification would launch the app, but you couldn't launch the app directly. You would need to go into Expo Go and it was a little annoying. This took me so far a weekend, but it's still kind of a working project. So I hope to get it done with another like week of work, but let's see.

Wish me luck. So final takeaways. It's okay sometimes to think very small because the smaller the scope of your app is, the more likely you are to actually ship it, even if shipping means installing it on one phone. You can have fun with libraries, try out new things. I suggest not to try too many at a time, because otherwise you go back to having a big scope and then you're likely to hit some bumps and stoppers, and then you're not actually going to build your app.

Final last piece of advice. If you can at all, stay away from the app store for as long as possible, because it's always a black hole of time investment, whatever process, whatever technology you're using, it's always painful. So if you can, just stay away.

And that was it. This is my website. The slides are there. So thank you very much.