Superpowers of Browser's Web API

Hello React Summit US! I wish you a warm welcome to today's talk. Today we're going to talk about something called Web API. But before we do that, we will need to take a small step back and go through some JavaScript basics.

If we see this example, and I ask you the question, what the order of the logs would be in the console, probably most of you would say 1, 3, 2. And that is correct. But how do we know that? How do we know how setTimeout behaves?

As we know, JavaScript is a single thread language. Let's visualize this a bit. There is a call stack, we execute the first function which is synchronous, and then we log the 1 as a value immediately. But then the second one is asynchronous, and it needs to go somewhere else, not to block that single thread. And it goes to the queue once the timer is finished. In the meantime, the call stack continues with executing the functions, then event loop checks if the call stack is empty, and pushes back the callback with the value of 2. But the main question is, where does the setTimeOut wait while we continue with the execution of our code?

If we Google this function, the first result we get is this one. So there is MDN documentation where we can see some stuff about setTimeOut function. And here we can see that it's under a tab WebAPIs. If we click that, we can see that basically WebAPIs is a collection, a list of APIs and interfaces which we may use right out of the box when developing web applications. Basically, it's a functionality provided by a runtime environments. In this case, our browsers. And if we scroll a bit, we can see that there are a lot of APIs, a really, really long list of APIs. And here we can maybe observe that there are some quite commonly known and used APIs, like WebRTC, session storage, local storage, file upload API, a DOM API, and so on. But looking through this list, I was wondering, what are the some of the exotic and not so commonly used APIs that we may use which can give us our web application certain superpowers in certain use cases? So basically, we're here to find out that today. And like we established, the answer to this question about set timeout is Web API, which is basically a collection of built-in interfaces that allow developers to interact with our pages and provide some fancy functionality.

I'm going to quickly introduce myself. When I was in high school, I always wanted to study psychology. Now I work my dream job at Wega IT based in Noišac, Serbia, where I work as a development which means that I teach people soft skills, how to communicate with the clients, how to deal with stress, how to do presentations like this, why not, and so on. My name is Nikola Mitrović and these are my Web APIs highlights for you for today. OK, now we go to our first example for today. Let's say we have a page like this and there is nothing there but the background, but once we scroll a bit down we can start observing that there is something in the bottom left corner. It's a small astronaut, which when fully visible says, hello world. Once we scroll a bit up and the astronaut is not fully visible, the message goes away. We scroll again a bit down and again, once the astronaut is fully visible again he says hello world.

So how did we manage to do this? There is an API called IntersectionObserverAPI which basically figures out if the element is visible on a page or not. If we would build a hook in React for this, we would probably call it useVisible and that hook would accept two parameters.

It would accept the reference of the object that we are trying to observe and we have certain configuration options. One of those options is root. If we pass null, then we assume that we are observing the element according to the whole document. Basically it means we can have smaller containers with smaller scrollable areas and we can observe some elements there, but if we pass null we look according to the whole document. We can have a margin around that root, so we can catch that intersection a little bit earlier if needed. And we have a certain threshold, meaning that if we pass this value we need 100% of visibility for the element, so we can adjust this accordingly. Then we would have a state called isVisible, and then we instantiate the IntersectionObserver object. In the callback we get the Entry object, which has a property called isIntersecting. This property holds the information about the visibility of an element, and then we can set that to our state. So now what's left to be done is just to call observe method on that reference and return that state. And that would be it.

The example with the astronaut was nice and cute, but you're probably now wondering what are some real-life use cases where we can use this API. We could lazy load images, right? So we load the initial content for the page and everything below the fold, what is not needed initially, we can defer and lazily load. And once we scroll to a certain section, then we start downloading images and other assets. We could create infinite scroll lists. For example, we don't need library for this. We could have a list out of 10 elements. We could have a separator at the end of that list. And once we scroll back and that separator is visible, then we trigger the next request. We could defer animations. So we don't need to run animations if something is not... If the certain element is not visible on that page. And by doing so, we can save some computer power. So we don't need to run some JavaScript if the element is not visible in the viewport. One thing to always consider when using some of these APIs is browser support. If we look at the major vendors, we can see that this API is fully supported. So great news, we can safely use this in all browsers. OK.

Now we go to another example for today. Now, I'm going to do something with my personal preferences. I'm going to say that, first note that my laptop is not on the charger. And then, if it's not on a battery, I'm going to say, if my laptop is inactive for one minute, it will turn off. Also, I have my phone connected next to me, which has the same setting. And we're gonna have the same application. So note by the URL that we have the same application and the same code running behind these two examples.

Now, I'm going to restart the timer for the both of applications at the same time, hopefully. Okay, and here I'm going to click this toggle, but for the mobile version, I'm not. You may notice at the very beginning, I didn't talk much about myself, but now we have a minute to spare and it's my time to shine. Well, I like to travel, and when I do, Italy is one of my favorite destinations to go to. The architecture, the people, the wine, pasta, it's all amazing for me. So, it's inspired by that. Recently, when I get home, I like to prepare food for myself. So, I go to some cooking recipes website. My phone is next on the counter next to me and I'm preparing that food. But after a minute, something happens. And we're just going to wait for a few seconds. Okay. After one minute, my phone turns off. And now, with sticky fingers, I have to unlock my phone again. Note that the timer went to the zero seconds here, but went to one second on my phone. It's simply because my phone turned off. And we have lost the connection with my phone. Once I touched it again, my phone turned on. But again, we need to unlock it and use the phone again properly. Here it didn't. Because we have the most awesome feature, which says, basically, hey, I'm cooking. Do not turn off my screen. So how did we manage to do this? There is an API called Screen Wake Lock API, which prevents the device from locking.

Although our personal preferences said something different. And it can only work if the document is visible on our page.

If we do this in React, we will make a hook called Use Wake Lock and it looks something like this. We pass that toggle Boolean, which is turned on or off, right? And if it's turned on, then we call a request Wake Lock Sentinel function. That function, I will call on a navigator, we have Wake Lock object and a request method with the parameter screen. And that function call is a promise. It's a promise because it can fail in certain use cases. For example, if our battery is way too low and now the situation is a little bit trickier than in the previous example, because we need to put this in a reference. We need to do that because if we put it in a state, then we will lose the previous state and this function returns a certain value. We need to clear that value exactly that value in memory afterwards when we release the lock, otherwise it won't work. That's why we need to put it in a reference. Then we return if we have the lock or not. And when we release, once we toggle back off, then we need to call the release method on that very reference and we need to clear that reference. Those are all necessary steps to be done, otherwise the lock will remain in memory.

Again, some of the use cases for this, reading an e-book, presenting to an audience, following a recipe, map navigation, or scanning a QR or barcode. Again, if we look at browser support for this API. Now we can see that the situation, again, is a little bit different. Chromium based browsers support this feature, Firefox doesn't. And this sign means that it's in experimental phase for Safari. So they have some work in place but not fully. One more thing to notice about this API is that there's a certain device support. So some of the devices support this feature, some not. For example, I have an iPhone and whatever I did, I couldn't make this feature working although I use Chrome. I think this has to do with the fact that the Chrome is just the skin for the iPhone and it's basically Safari under the hood. It uses WebKit engine to run the code, but that's something we have to deal with when using this API.

And moving on to our last example for today. Here we have an example of svemirko.rs which translated to English means spaceman. Svemirko is a little bit futuristic app, a few thousand years in the future, where in the backend, we're communicating with the NASA's API which retrieves us the list of all habitable planets.

Now we can search through that list and choose where we want to go on vacation. But in the future as well, we have some offline connectivity issues. If we go offline and if I tell you that hovering over this explore button will trigger the request, you would probably expect that request happens in the network. But if we are offline, that network request would fail, of course. We may notice that there is nothing going on in the network. But in the Applications tab under Background Sync, there is something else going on. If we click, we will see the loader. Basically, our request hasn't gone yet. And here in the Background Sync, we can see that there is something called RegisterEvent. RegisterSync. Once we go back online, then we have this PepsSync event. Our request went on. And if we wait a bit, once that request is finished, then we will have the third event, which is called SyncCompleted. Now we have the list of our planets. And we can search and decide where we want to go to the vacation.

How did we manage to do this? There is an API called Background Sync API, which basically provides a way to do the fire task it to a service worker, which we can run in the background. If we build this example in React, we have that link. Once we hover, there is on a navigator, inside of a service worker, there is a ready object, which is basically the registration for the background sync. And after that, we need to call sync and register method and pass a tag of a task. We do that because we might have multiple tasks running in the background while we are offline. And inside of our service worker, what's left to be done is just to compare the name of that tag. And once we get that event, then we can continue with our request. Of course, some of the use cases are offline UX and maybe we want to send an email, but we went offline that very moment, we can recover for that and run that task in the background which will be continued once we go back online. Again, checking the browser support for this API. Not so well at the moment, but we'll wait a bit for this because this is quite an awesome feature.

In the previous example, if you have wondered how did we manage to pass the data from the service workers to our component. Well there is an API called broadcast channel API, which is basically a channel for the communication between same contexts, which means, if we have the same windows, tabs, frames or iframes, which are on the same origin, then we can communicate between those origins.

And it goes both ways. In our previous example, in our service worker, we need to connect to a channel. We create a new broadcast channel and connect to it. And we, through the post message, we can send certain information to our component. In our component, we subscribe to that very same channel and on a message event, we will get, as a parameter, we will get the data that we send from a service worker.

Again, some of the useful use cases are to detect user action from another origin. Or if we have multiple tabs open for the same application, if we log out in one, we want to log out user in all tabs. Browser support in this case says all cool, all good. We can safely use this API.

Unfortunately, because we don't have too much time, we have to stop here for today, but I'm going to give you some bonus points and I hopefully, inspired you enough to go and search some of these APIs on your own. But here are some more of my highlights. There is a beacon API, which basically is used when you want to send a request, but you don't care about the response. Useful for analytics. There is a web speech API. Page visibility API, if you want to track the visibility of the whole document and certain elements. There is a network API which gives you your network system information. There is a web share API which can share information, share files between different devices. There is a background fetch API. Basically if you want to download a huge document but then you go offline, you can continue with the download once you're back online again. And of course one of the probably most exciting and new ones are web authentication API, which is basically a way for the passwordless experiences. We won't need passwords anymore. In the future we could just authenticate with our Touch ID with our face or with our USB key or some other devices. So really really exciting to see how this API will turn out. And at the end some of the takeaways. We have seen that some of these APIs have no browser support at this very moment. Some of that support is in the experimental phase. We have to watch for the device support in some cases and there is a certain TypeScript support that is lacking. But why should we use this API and adopt this mindset? Because it's standardized code, it's standardized code, it's standardized code. I probably cannot stress enough how important this is, especially in today's Web development where we have a new framework, new library each week. This is like a single source of truth.

This is standardized code, we can use it in all browsers. So it's really, really important. Most of these APIs are really easy to learn and by adopting this mindset and using some of the built-in functionality in our browsers, and not using the libraries too much, we can have the, we can increase performance as a side effect. Of course, I'm not saying that we shouldn't use libraries and shouldn't reinvent the wheel in certain use cases, but in some use cases, we can probably build some features on our own with the help of these APIs.

And that brings me to my point, that with great power comes great responsibility. So I felt like this is my responsibility to say this to you. I wanted to talk a bit about our planet. If you're having a 404 facial expression, like not found at the very moment, like what, what does the, our code has to do with our planet? Well, it kind of does. We build our React application. We run a build. There is an output for that build. We deploy that, our HTML, CSS and JavaScript files to some cloud providers, which are distributed across the world. They are in the data centers. Those data centers have a lot of computers, are basically essentially a lot of computers next to each other, and they're all plugged into electricity. In fact, so much that the internet is consuming 21% of all electricity in the whole world, which is really staggering information for me, if you ask me.

But luckily for us, there is a very cool application called Website Carbon Calculator, which can check out the carbon footprint, our website traffic, energy intensity, and data transfers. So I would highly advise you to use this application and go and check how green you are. This is super important because we have only one planet. This planet is our home and if we're not kind to it, one day is going to be too late. So what we can do is all of us can take action, join forces, and in that sense we can provide big, big results all together.

Thank you so much for coming today and for hearing me out. If you want slides to recap and to go over them again, this is the QR code. You can scan it. You can write me. This is my email. You can go to my GitHub and revisit all examples that I've built. Thank you.