In November of 2018, I started a third attempt at switching to Android as my primary mobile operating system. This time, the experiment lasted 9 months. But I switched back to iOS nevertheless.

In 2013 and 2014, I conducted two experiments whether I could switch to Android as my primary operating system. The conclusion after my last experiment was, again, that Android was not fit for my use cases and usage patterns, and so I stuck with iOS. The first experiment lasted under a week, the second lasted 18 days.

In November of 2018, with a lot of exciting things happening around a new Firefox app for Android which is in preview right now, and some other factors, I thought it was time to reevaluate whether Android was now fit for my daily usage for a longer period of time.

Finding the right device 🔗

The first step was to find the right device for this experiment. I had been going with Google Nexus 4 and 5 respectively the two previous attempts, but wanted to start with something less pricey this time. Moreover, there were new initiatives out like Android One which promised a stock-Android experience across a wider range of devices. So, I went for a Nokia 6.1, which Amazon even had on sale that week.

Unfortunately, that low price immediately showed. Even though this phone was already running the then new version 9 of Android, code-named Pie, the device itself was a bit underpowered and felt sluggish with almost everything I did. It took over a second to open any panel in Settings, for example, and I found the TalkBack gestures to be somewhat lagging even when only sliding my finger across the screen. Others might not even notice, but I felt that this just wasn’t right.

I then swapped that out for an Nokia 8, which almost feels like an iPod Touch and was very comfortable to hold. Unfortunately, that one didn’t have Android One yet. Moreover, the fingerprint sensor integrated into the front of the device was uncomfortable with the finger having to be put on it sideways. It required to hold one’s hand at an odd angle. Also, it could not be repurposed for TalkBack gestures like even that of the Nokia 6.1 could.

I swapped that one out for a Nokia 8 Sirocco, the then flagship device, which Amazon happened to have on sale two weeks later. Yes, I did take advantage of Amazon’s generous return policies those few weeks. 😉 And that one, even though it still ran Android 8.1 Orio, was a very good fit for me. It was fast, felt sturdy to the touch, and had everything I needed. This included Google Assistant which I could activate with my voice. One thing all of the previous devices did not have, including this one, was the capability of WiFi calling. That is when your mobile carrier allows calls from and to your cell phone to be routed through the WiFi you happen to be connected to. In my apartment, which has notoriously low cellular reception, this was actually something I had quite enjoyed while using my iPhone.

And then the Sirocco got Android Pie, and they disabled a whole range of features. Google Assistant was very crippled in my mother language German, and after talking to support, they could not even give me an estimate if and when that would improve. My wife, who had got a Google Pixel 3 for Christmas that year, showed me that for her on Pie, everything was working. So, this was an artificial crippling of the functionality imposed by the makers of the Nokia phones, or the Android One infrastructure.

It so happened that shortly after Christmas, in January 2019, prices for the Google Pixel 3 began to drop. I put my Sirocco on Twitter for sale, and a friend gave me some good money for it. She didn’t need Google Assistant or WiFi calling and was perfectly happy with the device as was. I got the Pixel 3 and continued my journey. This is the device I still have today. It even had WiFi calling on from the start, giving me back that feature I had missed with all of the Nokia phones.

Improved app quality 🔗

The first impression I got was that a lot of the apps I used regularly like Twitter, Facebook, WhatsApp, 1Password, Spotify etc. had improved quality and accessibility. In general it can be said that all the bigger players treat Android accessibility with a similar mode of dedication as they do their iOS apps. The limiting factor is often more the inferiority of the APIs, which Android still has despite advancements with each version, than actual lack of commitment.

Finding replacements for some of my iOS apps still proved a bit problematic. While I found Envision AI to give me recognition of pictures in various apps, I found it often lacked the precision and detail I had come to like about Microsoft’s offering. That, however, was, and I believe is to this day, not available on Android. Also, navigation is still a mess, requiring several apps to be used in combination, sometimes even parallel, to get a similar experience to what BlindSquare offers on iOS.

pain points 🔗

The pain points started in March after the monthly security update. This had something in its backpack which made TalkBack very unstable. It restarted frequently, giving usage hiccups. After that problem hadn’t been solved with either the April or May updates, and people on the Eyes-Free mailing list suggested that the Android 10 betas, which had by then started rolling out, would not suffer from this problem, I actually took the leap and installed the beta on my Pixel.

Another problem arose in April or May with one update to the Chrome browser and web view. This suddenly introduced huge sluggishness when scrolling, constantly losing position and other really painful bugs. And even though I could set Firefox as my default browser, this doesn’t permeate every aspect of Android. There are still enough places where the app insists on bringing up a Chrome WebView or branch out to the Chrome browser regardless of that setting. That bug wasn’t solved with the three following updates to Chrome or the WebView. I believe in August or September, it was finally solved. And yes, I was one of the people who reported this to Google, and the replies I got made me feel quite lost. They were, indirectly, indicative that actually nobody at Google had noticed these problems yet, even though they were not only in the release, but beta and Canary channels as well. So they had been in those channels for some months before even hitting my Pixel in release form. And they were very easy to spot. Like: Within 30 seconds of using it.

Other pain points amounted to annoyances over time. One of them is the fact that Android is still only allowing TalkBack to use one finger for gestures. Two and more finger gestures are passed to the system as if they were executed without TalkBack on, meaning tapping with two fingers would register as an ordinary single finger tap. This limits the customization options of TalkBack and other such assistive technologies. They cannot use a rotor like VoiceOver, for example. The result is a pain point I raised in the past, and that is Google’s multi menus and 90 degree double swipe gestures. Google offer two ways of showing the local and global context menus as well as others such as the one for Edit and Custom Actions. But these both have their problems. The overlay mode requires to move the finger in a circular motion across the screen to find the right item. If I don’t place the finger at the center of the screen, I suddenly end up with too little room to maneuver for certain menu items if the menu has many actions or options. The other mode is showing the items in a vertical list. That takes longer to come up than the overlay, is accompanied by extra announcements because this is a full new view that is showing, and is generally more time consuming. If only using it once or twice for testing, this does not account for much. But if using it a hundred times per day or more, which is not unlikely at all if you’re a heavy user, this sums up in inefficiency big time.

This limitation is also why there is no built-in Braille screen input. There are apps out there that allow for Braille screen input in English, and some in even more languages. But all of these require that TalkBack be turned off while you type in Braille on the phone screen. This makes the experience cumbersome, if not unreliable. There is always the chance of you touching something while TalkBack is turned off that may have an undesired effect.

In summary 🔗

It took a much longer time to register, and an even longer time to become an annoyance this time, because I really tried and I really wanted this experiment to succeed. But in the end, these combined pain points, the feeling that the system could become so unreliable due to a security update, or such glaring bugs like in the Chrome WebView not getting fixed for which was in sum almost half a year, made me switch back to my iPhone X in late August. I had kept the iPhone even though I had tried to sell it twice, but both sales didn’t pan out. I guess fate was telling me something there. 😉

And as if to confirm that my decision to after all stick with iOS was right, iOS 13 came out with tons of new accessibility features. I enjoy photographing every now and then. On the new iPhone 11, VoiceOver’s image recognition even works while you are taking photos. So, you can make sure you at least get in the shot approximately what you are aiming for as a blind person. This is huge!

Aside from that, there are new features in the WatchOS operating system especially for Apple Watch series 4 and 5 my wife and I need in relation to her specific health issues. So we actually both switched back to iOS from Android because we need the Watch integration.

And this is where I’ll stay for the time being. I am using Android and that said Pixel for work on Firefox for Android, both the current releases and the one that is in preview, but my main operating system is and will remain iOS. Others may come to different conclusions, depending on their requirements and usage patterns. All in all, Android is a much safer bet than it was a few years ago. Especially if you don’t require Braille screen input, Android may be an option as an alternative. But for me, the sum of limitations and experiences has shown that it is not the right fit.