Gaël Duval


Leaving Apple and Google: my “eelo odyssey” - Part1: the mobile OS

Edit: support eelo now! The Kickstarter campaign has just opened!

So I came out about my decision to leave Apple and Google. It’s a lifestyle choice to escape the tech giants that make me a product by privatizing my personal data .And I don’t like what Apple is doing now, Apple’s attitude, new iPhone and their price… It’s also an act of freedom for my children and all the people who will care: I want them to have a choice, and also a clear and informed view on how their choices can impact their life and their economical ecosystem as well. That’s what eelo is all about: offering a viable and attractive alternative to users for their digital life.

In this new post I’m going to describe what I was able to do so far on the mobile to get rid of Google and Apple, and what remains to do (spoiler: there’s a lot). In the next part I will explain what how things will need to be adressed on web services and draw a whole picture of the eelo project.

What’s wrong with default AOSP/LineageOS?

Talking about LineageOS, you might think “why do you want to hack something that is already mostly open source and works well?”

The answer is easy: the core of AOSP/LineageOS is usable, and performing well, but it’s not good enough for my needs: the design is not very attractive and there are tons of micro-details that can be showstoppers for a regular user. Also, unless you are a geek, LineageOS is not realistically usable if you don’t want google inside.

The design point

Regarding design, I know that some Android users like it, but I really dislike the default graphical user interface. I find it ugly: icons don’t look good, colors are sad, and I don’t like the launcher ergonomy and behaviour.

So at least we need a new launcher, and better icons. Default notifications don’t look very good either, and I’m not a big fan of the settings part. Compared to the rest of the UI it could be worse, but it’s still quite sad, with a single green color in LineageOS. I’d like something more appealing, and probably better organized.

“Good news”: you can find hundreds of custom launchers and icon themes in the Google Play Store. But either you have to pay for them, or you get free stuff with lots of ads and possibly scams. So not for me.

Bad news, good news

The bad news is that I’m new to Android development and I don’t consider myself a great developer. I can hack things, I can recompile and integrate stuff, but I don’t have enough practise to program a new launcher from scratch without spending weeks on it.

The good news is that I have found a very talented full-stack developer who is interested in the project. We have agreed, as a first collaboration, to release a new launcher, new notification system and new “control center”.

First successes

I’ve choosen to test custom builds of LineageOS/eelo on a LeEco Le2. It’s a nice 5.5″ smartphone with a 1080×1920 pixel screen, 3GB RAM, 32GB storage, finger sensor in the back, and a 4K camera. It costs about 130€. Yes, that’s about $150. Yes.

Also I’m waiting for a Xiaomi Mi 5S. It’s got a smaller screen and I prefer smaller devices for my own usage. And I’ll probably give a try to the LG G6. (Want to suggest a device? tell me!)

After several weeks of work, we now have a new launcher! It still lacks a few features (such as uninstalling an application), but it’s already fully functional. On this video, you can see the “icon group” feature, and swiping between several launcher pages:

eelo BlissLauncher 1 from eelo

On this one you can see the “docking icon” feature:

eelo BlissLaucher docking action

We call it the “BlissLauncher” just because it’s a great launcher. And we also have a first new notification system and a new unlock screen:

Next time will be to have all that integrated by default in a new fresh build. And at the time of finishing this post, I already succeeded to flash a fresh build with the new launcher and the new notification system.

Getting rid of Google stuff completely

Now we have a better launcher for eelo, and I’m working with a great and very professional designer. He contributed a lot to the Mandrake Linux interface icons in the past, when we redefined all the user interface and all icons. Later he also contributed to first releases of Ulteo, when it was still a cloud-operating system project, and not a Citrix-alternative. We’re working together to redesign default application icons, some wallpapers, splashscreens, and also a first real eelo logo. On the long term, we will have to redesign the full user interface.

But what we want is not only something good-looking, attractive and easy to use. We want more privacy! And Google services are not compatible with my idea of privacy.

Therefore, we don’t want Google Services. We don’t want Google play store. And we probably don’t want most of Google apps such as Calendar, Email etc.

Also, we probably don’t want Facebook either and some other so-called “free” services. This will be user’s choice to install them or not. I know that we cannot change the world in one iteration, this will be step by step.

Each of this point will need to be addressed in Eelo. We will need an independent application repository, an independent and secure email provider, an independent online drive, online office services… All that well integrated in eelo. In the user interest first.

First round without Google

The first time I was able to recompile and flash LineageOS, I soon had to install Google Play Store and Google Play Services to install common applications, or I could do pretty nothing.

But there are some alternative stores. For instance, F-Droid is a very successful APK application repository that provides only 100% open source software applications.

There are other alternative app stores for non-open source applications. For instance there is Aptoide. It provides most common applications such as Twitter, Waze etc. But unfortunately when I checked Aptoide APK packages signatures and sizes, I realized that they were not the same as on Google Play Store. I’m not sure to understand well the reasons behind this situation, at least for common applications, so I looked for other alternatives.

I found APKPure to be a great store for free applications. And trust me, a lot of applications are free! Actually, I realized that on my iPhone I had only free applications. And I know many people who are using only free applications. So APKPure is a great way to go if you don’t want to use Google Play Store and don’t need non-free applications. I checked many of their packages, and they are bit-to-bit identical to the ones available on Google Play Store. There are only official packages.

An alternative to APKPure is Yalp. Yalp is an open source application that is acting as a kind of anonymous proxy to Google Play Store, also providing only official APK packages.

So for applications, I’m now using both F-Droid and APKPure. That’s already very confortable, and I successfully tried dozens apps, including the most used apps (Facebook, Messenger, Twitter, Waze, Telegram, Skype, LinkedIn, Spotify…).

But I think we’d need an “eelo store” that would deliver both:

  • official free applications like APKPure
  • open source applications like in F-Droid

All that into a single, appealing and fast application, where users could check easily if an app is open source or not, where users could evaluate the application level of privacy, and where users could be able to report some scam issues. We definitely need to add this to the eelo roadmap.

Lovely Google Services

There is a feature that Google has created to jail users within their environment. That’s called “Google Services”. It’s a non-open source service that you have to install if you want to use Google Play Store, for instance. It’s also used by several applications. It provides services such as:

  • analytics
  • account authentication
  • cloud messaging (notifications)
  • drive
  • geofancing
  • maps API
  • mobile ads
  • games API

Developers of Android applications are not forced to use them, but obviously Google is doing their best effort to make them desirable as much as possible, if not mandatory for certain features.

The good news is that many common applications, the ones that everybody is using everyday, are not using Google Services, or they do not rely a lot on them. Probably a lot of developers don’t like to be jailed in a single ecosystem.

As far as I tried, the most problematic applications in this regard seem to be some games, such as Pokemon Go. This one doesn’t seem to be usable unless you have Google Play services installed.

The good news is that there is a nice project that is providing open-source alternatives to Google Services. It’s called MicroG, and eelo will probably integrate it.

Another “great idea” of Google is their SafetyNet Attestation API. It’s something that Android application developers can use to check if the user’s device is an official device that complies with Google’s environment. It examines the hardware, the software, checks wether the device is rooted or not. In the end this can be used to prevent to application to run if the environment doesn’t comply enough with Google’s rules. Fortunately, there is “Magisk” to circumvent this issue. We will probably need to integrate it by default in eelo as well.

What about web search?

Many parts of a modern operating system can lead to “Privacy indiscretions”. So far, I’ve talked about privacy issues that come from within the system.

But if you search for something on Google, it’s very likely that Google can determine that YOU are looking for something in particular. Even if you are not using a google account in you Chrome browser, they can track your IP for instance.

So we definitely need to provide a default search engine alternative to Google search. Probably that we don’t want Bing or Yahoo either, although it’s better to use various search engines so that each of them doesn’t know exactly everything about your searches and therefore cannot consolidate your private information efficiently. We have a few alternatives:

  • the well-known DuckDuckGo: even though it heavily relies on Google Search results, it offers privacy guarantees that Google doesn’t offer.
  • Qwant is a new search engine that is making big progresses and now has its own index and is offering guarantees on privacy
  • there is also the fully open CommonSearch: project, but it’s not ready yet

So I’m considering offering both DuckDuckGo and Qwant as default search engines for eelo search and web browsers that will ship with eelo, while still offering Google (and others) as an option. It’s true that in some cases, it is still offering the best results.

And also…

There is a long list of Internet services that can track you, send and process your personal data in many ways. For instance, using a Gmail (or similar) email account is a great way for Google to learn a lot about you.

But also, some of you probably know about the very fast Google DNS resolver: and DNS resolvers are used all the time and by many applications. They convert domain names to IP addresses. And I say: DO NOT USE Google DNS resolvers. Each time your smartphone is looking for a domain name, Google knows about it and they can add this information to other information they know about you.

Instead, you can use (or 2620:fe::fe IPv6) which is a fast public DNS resolver operated by a non-profit research institute that does not store your IP. And it be accessed throught a secure protocole (TLS).

Of course, it’s all the web-service ecosystem that we need to address. As I said earlier, eelo will provide a mobile system with better privacy, but also some web services such as an online office suite, some online storage etc. We will aggregate some existing web services, improve them if needed, or build new services if nothing is available.

Still, we will face one dark zone: low-level proprietary hardware drivers on smartphones. They are driving the camera, the GPS, various sensors… Hardware vendors do not provide source code for these drivers. And they are extremely difficult to rewrite unless doing some heavy and resource-consuming reverse-engineering. And of course, some of those “black box” drivers could possibly leak users’ private data.

Future options for eelo to address this issue will be to:

  • partner with FairPhone or similar 100% open hardware projects
  • audit low-level drivers to detect unappropriate behaviors
  • design an eelo phone…

Join the eelo odyssey!

As you can see, eelo is a true odyssey. But I think that, maybe for the first time, all bricks are available to build a new, consistent, attractive, independent and mostly-free digital ecosystem that will be more respectful of users, and respect their privacy. And this could eventually challenge the advertizing model that is probably the source of this such bad and supposedly “free” model.

Again, eelo is a non-profit project, it’s a project in the public interest. Everyone who wants to join, please do!

There are many ways to contribute:

  • say hello! 🙂 having supporters help a lot
  • contribute some ideas, some resources, what you are good at
  • introduce us to people who can help
  • talk about eelo, share eelo news and articles…
  • offer a few mɃ to pay some servers

Also, I’ve started to work on a crowdfunding campaign for eelo, because some resources are needed to bootstrap this project correctly. I’m not sure exactly what this campaign will be able to offer in rewards, but I’m thinking about it. Anybody’s suggestions are welcome!

Keep in touch with eelo: register at landing page. Get in touch with me: gael @ eelo . io

EDIT: read more:

— Gaël (follow me on Twitter / on Mastodon)

PS read the first part 1 of this serie of post: Leaving Apple and Google : my “eelo odyssey” — Introduction

PPS this article was originally posted on my website

More by Gaël Duval

Topics of interest

More Related Stories