Submissions from the public were invited, so I replied and was invited, with “Clark Kent” (a UK iOS Web-Apps developer who wishes to be anonymous) and Stuart Langridge to brief the CMA in more detail. We were asked to attend for 2 hours, so spoke for one hour, and then took questions.
CMA are happy for our presentations to be published (they’re eager to be seen to be engaging with developers) but asked that the questions be kept private. Suffice it to say that the questioning went over the allocated time, and showed a level of technical understanding of the issues that surpasses many engineers. (The lawyers and economists knew, for example, that iOS browsers are all compelled to use WebKit under the hood, so Chrome/ Firefox/ Edge on iOS share only a name and a UI with their fully-fledged real versions on all other Operating Systems).
Unfortunately, Safari’s bugs made development much more expensive by breaking Indexed DB; breaking scrolling etc. Also, Apple refuses to support Web Bluetooth, so his customers had to purchase expensive iOS-compatible printers for printing receipts, at 11 times the cost of Bluetooth printers. iOS apps can integrate with Bluetooth, however.
He can’t ask his customers to jettison their iPads and buy Android tablets. Neither can he ask them to use an alternative browser, because there isn’t any meaningful choice on iOS. Clark simply wants to be able to make “amazing Web 2.0 and Ajax apps that look exactly and behave exactly like apps on the iPhone”, as Jobs promised. He summed up:
Stuart examined whether Apple’s assertion that lack of real browser diversity is actually better for consumers’ privacy and security.
Here’s my captioned video. Below is the marked-up transcript with links. Note: this submission is my personal opinion and does not represent the opinion of any of my past or present clients or employers.
Transcript and Links
Hi, I’m Bruce Lawson, and I’m here to talk about Progressive Web Apps and iOS. I’ve been mucking around on the web since about 2002, mostly as a consultant working on expanding people’s websites to be truly worldwide, so they’re performant and work with constrained devices and networks in emerging markets. I’m also involved with making websites that are accessible for people with disabilities. And I was on the committee with the British Standards Institution writing BS8878, the British standard for commissioning accessible websites that last year got a rolled into the ISO. And before that, I was deputy for the chief technical officer at Opera software that makes a popular web browser used by about 250 million people worldwide, predominantly in emerging markets.
Most of my customers are in the UK, USA, or Israel, and they’re largely focused on UK and US markets. In the month of June this year, stats counter showed that 30% of all US web traffic was on iOS, 28% windows, 22% Android. And to drill that into just mobile browsers, 57% were iOS and 42% were Android.
In the UK, 30% of people use Windows to browse the web, 27% iOS, 25% Android. And that jells pretty well with stats from the GDS, the government digital services. Their head of front-end, Matt Hobbs, every month tweets some stats. He said that in June, iOS accounted for 44% of traffic, Android 28%. (Although he does note that because you have to opt in to be tracked, it tends to over-represent the mobile browsers). But we can see that in the UK, 50% of traffic is on iOS and 49% is Android.
So, we obviously have to be concerned with iOS as well as Android, Windows, et cetera, and all the different browsers that are available on every non-iOS operating system. And so, I always recommend to my customers that they use the web. After all, that’s what it’s for.
The principle of universality allows the web to work no matter what hardware, software, network connection or language you use, and to handle information of all types and qualities. This principle guides Web technology design
And that was said by Sir Tim Berners-Lee, and his parents were both from Birmingham, as am I. So he’s practically a relative. And with my work in web standards, those principles still hold true for the web standards we make today. So, I wanted to use the web and I wanted to use what’s called a Progressive Web App, which is basically a website plus. It’s a website and it’s web pages, each of which has a URL. It’s built with web technology, but in modern browsers, conforming browsers, it can look and feel more app-like.
In fact, the UK government itself recommends these. Our government writes,
advances in browser technology have made it possible to do things on the mobile web that you could previously only do with native apps. Services that take advantage of these modern browser enhancements are often called mobile web apps or Progressive Web Apps. The user experience of PWAs and a handset is almost identical to a native app, but unlike native apps, PWAs have a single code base that works on the web, like any normal website. A benefit of PWAs over native apps for the government is that there is no need to maintain lots of different operating system versions of the same app. This makes them relatively inexpensive to develop compared to native apps.
And just like government, business likes to reduce costs, and therefore can reduce costs to the end user. Our government goes on to say,
users can choose to install PWAs on any device that supports technology. If a user chooses to install a PWA, they typically take up less space than the equivalent native app.
And this is very important in emerging markets where lower priced phones have lower specifications and less storage space. The government continues,
PWAs cost less than native apps since they need to be only developed once for different platforms. They are also forward compatible with future upgrades to operating systems
And then it lists browsers that do support PWAs, and it lists Apple, Safari and iOS, but this is only partially true. Apple themselves make this claim in their February 2021 submission to the Australian investigation into monopolies. Apple said of PWAs,
Web browsers are used not only as a distribution portal, but also as platforms themselves, hosting progressive web applications that eliminate the need to download a developer’s app through the app store or other means at all
Therefore, Apple is suggesting that PWAs are a viable alternative to the app store. Apple continues,
PWAs are increasingly available for and through mobile based browsers and devices, including on iOs [my emphasis]. PWAs are apps that are built using common web technology like HTML 5, but have the look feel and functionality of a native app
But, as we will see this, isn’t true. PWAs do not have the same functionality as a native app.
Let’s recap about what a PWA actually is. It’s a bit of an umbrella term, but fundamentally it’s a website, but it can install to the home screen of your mobile device, just like a native app. And it can have its own icon defined by the web developer, launch full screen, portrait or landscape, depending up on what the developer chooses, but it lives on the web.
And this is a really important. It’s important for several reasons. Firstly, because it lives on the web, you’re not downloading everything in one go. A native app is like downloading a whole website all at once. A PWA, you download just the shell, and as you navigate around it, it will get any content it doesn’t already have stored from the web server. Another advantage is that means that you have instant publication. The moment you press the button and it goes on your web server, the next person who opens your app on their device will see the new version. Great for security updates and also great for time-sensitive material.
Another great thing is, of course, is that because there’s no app store, no gatekeeper. Smaller shops don’t have to pay the 15 to 30% tax to Google and Apple. And also, for industries that are not compatible with the app stores, such as adult entertainment, gambling, et cetera, it’s a mechanism of giving an app because they can’t use the app store. And PWAs can work offline.
Let’s briefly look at one. I’m here speaking in a personal capacity. I don’t represent any of my clients or employers, so I can’t show you one of their apps, but I took 15 minutes to convert my WordPress blog into a PWA.
Okay, here, I’m using Firefox on Android. Android, unlike iOS, allows not only different badged browsers, but it actually allows that full browser on, so Firefox on Android is using Firefox’s own Gecko engine. Here, Firefox and Android are collaborating and they can tell when a user goes to my website that it is a PWA, and it offers the opportunity to the visitor to add it to the home screen. You can see it says “you can easily add this website to your device’s home screen to have instant access and browse faster with an app-like experience”. As a developer, I didn’t make this and the user doesn’t have to know anything because the operating system and the browser just show this. If the user chooses, they can add to home screen, they can click on the icon, and it will add automatically or else they can drag it themselves to where they want it.
And here, as you can see, it’s now on my Android home screen. There’s a little Firefox icon at the bottom of my logo to tell the user it’s a PWA. Chrome doesn’t do this, but it does also detect that it’s a PWA and offer to add it to the home screen. And as you can see from the screenshot on the right, when they tickle that icon into life, up pops my website full screen, without an address bar, looking like a native app. If however, a user goes to my website in a browser that doesn’t support PWAs, such as UC Web here, they don’t get offered the add to homescreen thing, they just see my website and navigate it as normal. So people with capable browsers get an app like experience, everybody else just gets the website and nobody loses out.
On iOS it’s subtly, but importantly, different.
So on iOS, if you go to my website, you’re not told anything about this being installable. You have to know to click that icon in the bottom bar that I’ve circled in red. And once you’ve done that, then you get the screen on the right. And you just have to know as a user to scroll down, whereupon you’ll see add to home screen there with no instructions or explanation of what’s going to happen. Anyway, once you click that, then you confirm, and finally it’s on the homescreen. And when you press it, it launches full screen without an address bar. But it still, even though it’s much harder to discover this, it still can’t do everything that a native app could do.
This is Echo Pharmacy. It’s a service that I’ve been using during the pandemic. It’s nothing to do with my employers or clients. I’m just a customer. I use it to get my repeat prescriptions, and I used it and it worked really well. And on the confirmation screen, after I’d ordered some more medication, it said, “Hey, why not use our app?” And I thought, well, why would I need an app, because this website has worked perfectly well for me. So I asked on Twitter, I said, “Why does it need me to download the app?” The website allows me to place and track my meds, why clog up my device, because it’s a download, right? Is it for push notifications, because on iOS websites and PWAs cannot send push notifications. And for a pharmacy like this, it’s important for them to say, “It’s 8:00 PM, have you taken this tablet?” Or, “We notice in four days time, you’re going to run out of this medication. Please come back and order.”
And I got a reply from the CTO of Echo Pharmacy saying,
Hey Bruce, you’re right about push notifications, which for some people are preferable to email or SMS reminders. Beyond that our app and website have feature parity,so it’s really just about what you prefer.
In other words, these people at Echo Pharmacy, not only have they got a really great website, but they also have to build an app for iOS just because they want to send push notifications. And, perhaps ironically, given Apple’s insistence that they do all of this for security and privacy, is that if I did choose to install this app, I would also be giving it permission to access my health and fitness data, my contact info, my identifiers sensitive info, financial info, user content, user data and diagnostics. Whereas, if I had push notifications and I were using a PWA, I’d be leaking none of this data.
So, we can see that despite Apple’s claims, I cannot recommend a PWA as being an equal experience an iOS simply here because of push notifications. But it’s not just hurting current business, it’s also holding back future business. A lot of my clients are looking to expand into new markets. The UN has said by the end of this century, 50% of the population of the world will live in these 10 countries. And as you can see, only one of them is a traditional market for a UK business, the US. And it’s obvious why people want to do this. Whereas, the population of the “west”is not increasing, it’s increasing a lot in new markets.
“The Southeast Asian internet economy is expected to reach $200 billion by 2025“, for example, but all countries are not equal in the technological stakes. Here, for example, is the percentage of monthly income that it takes to buy one gigabyte of mobile data. In the UK, for a gigabyte, (not actually that much [data]), it costs me 0.16% of my monthly income. In India, it’s a little bit more at 0.05%. In the states, 0.1%. In Rwanda, where one of my customers is doing great business, it’s 0.5% of monthly income. In Yemen, where I was born, it’s very nearly 16% of monthly income. So, big downloads cost a lot of money, particularly for something like a medication tracking system, which I might use only once every quarter. It’s a heck of a lot to download at these prices.
VisualCapitalist.com say worldwide, the highest average costs for one gigabyte of data is 30,000% more than the cheapest average price. And it’s not just about data costs either. It’s about data speed. Here, for example, in the UK, the average or the median UK speed is 28.51 megabits per second, in the States, it’s 55. In Hong Kong, it’s 112. In Rwanda, it’s 0.81. In Cambodia, it’s 1.29. In India, it’s 4. In Indonesia, it’s 1.88.
So, a big download not only costs potential customers more, but it takes a lot longer. Facebook recognized this when they released their Facebook Lite app. Facebook said
downloading a typical app with 20 megabyte APK can take more than 30 minutes on a 2G network, and the download is likely to fail before completion due to the flaky nature of the network
Twitter released something called Twitter Lite, which is now just everybody’s Twitter. The Lite’s a PWA, and Twitter said
Twitter Lite is network resilient. To reach every person on the planet, we need to reach people on slow and unreliable networks. Twitter Lite is interactive in under five seconds over 3G on most devices. Most of the world is using 2G or 3G networks. A fast initial experience is essential
So, if we could offer PWAs, we would have a competitive advantage over other people, have much faster installation and a much cheaper installation. Penny McLachlan, who works as a product manager at Google said,
A typical Android app size [a native app, in other words], is about 11 megabytes. Web APKs, TWAs, [this is technical speak for a Progressive Web App], are usually in the range of seven to 900 kilobytes
So, only about 10% of the size of an Android app. So, we can see that PWAs is a competitive advantage for people trying to do business in the wider world, because at post-Brexit, global Britain, et cetera.
It’s no coincidence that the majority of the early PWAs that we saw coming out in 2016, 2017 came out of Asia and Africa, and that’s because when people who make websites and web apps use the same devices and the same networks as their consumers, they absolutely understand why PWAs are great.
But, we can’t make a PWA because we have to care about Safari, which is more than 50% of our traffic in our home market. So, what can I advise my clients to do? Well, a lot of them have settled on something called React Native. They can’t afford to have one developer team for Android, one developer team for iOS, and one developer team for web.
If you’re a massive organization such as Google or Amazon, of course you can afford that, but developers ain’t cheap. And three teams is prohibitively expensive for a smaller organization. React Native, which is open source (but stewarded by Facebook), promises that you write once and it works across platforms, where you’re writing in a language called React, and it compiles down to native applications. So you write in React, you press a button, and it squirts out a native Android app and a native iOS app. And it’s pretty good, I guess.
There’s pros and cons, of course. The pros we write just once. Great. The cons. Well, it still emits a native app, so we still have to submit to app stores, and the timing of that is at the whim of the whoever’s moderating the app store, et cetera. It’s still a big download because it’s a native app.
It’s a more complex built process. There’s not a great deal of build to the web. React Native requires lots of moving parts, internally. It’s harder for us to test the functionality because although we write it once, we have to test it on iOS and we had to test it on Android, and Android testing isn’t as simple as iOS because there’s a massive difference between a top-of-the-range Google Pixel device, or the expensive Samsung devices and the kind of no name cheap Android devices you can get in street markets in Bangkok and Jakarta and Mumbai.
And a problem for me, and for one of my clients, is accessibility. Accessibility is making sure that your app or your site can be used by people with disabilities. Apple and Google have done great jobs making iOS and Android accessible, but React Native has some quirks under the hood, which make it much trickier.
One of the reasons for this is that React Native is quite a new technology and the web is a mature technology, and accessibility on the web is pretty much a solved problem. It isn’t solved if developers don’t do the right thing, but Chrome is 98.5% compliant with the W3C guidelines for accessibility. Microsoft Edge is a 100%. Firefox is 94%. Safari is 97%.
But because we’re writing in React and it’s doing magic tricks under the hood, there are holes in its accessibility. Facebook themselves say,
we found that React Native APIs provide strong support for accessibility. However, we also found many core components do not yet fully utilize platform accessibility APIs. Support is missing for some platform specific features.
They say it’s true, that support is missing. But for example, if you have an external keyboard plugged into an Android device, perhaps because you have Parkinson’s, or (like me) you have multiple sclerosis and you prefer to use a keyboard than the tiny little software keyboards. But if you have an external keyboard, you cannot use that keyboard to focus in to enter any information in any input field that’s made with React Native [on Android]. This of, course, is a huge problem.
Facebook has done some effort into writing up all of their accessibility problems. Here, for example, we can see this is a GitHub board with all the issues and I scroll, scroll, scroll, scroll, scroll through them. There is apparently a fix for the inability to focus into any input fields with Android and a keyboard, but we are at the mercy of Facebook’s release schedules here.
So, what could I advise my clients to do? I tend to advise them, go PWA and let’s hope that the CMA make Apple do the right thing. What I would really like to do is say to them, “Okay, just put this notice on your web app. Say: ‘we’d like to send you push notifications, if that’s okay by you. Please download Microsoft Edge, Mozilla Firefox, or Google Chrome on iOS’.”
However, if they do download Edge for iOS, Firefox for iOS, or Chrome for iOS, that’s no help at all because Apple prohibits those browsers using their own more capable engines and they have to use the Safari engine under the hood. So, consumers get no choice at all.