Has the AOSP project failed consumers?

When Google, along with a consortium of other companies, announced the open-source operating system we call Android way back in 2007, the world was paying attention. The iPhone had launched the same year, and the entire mobile space was wary of the rush of excitement around the admittedly revolutionary device. AOSP (Android Open Source Project) was born, and within a few years Android swallowed up market share with phones of all shapes and sizes from manufacturers all over the globe. Android eventually found its way into TVs, fridges, washing machines, cars, and the in-flight entertainment system of the Boeing 787 Dreamliner.

So you'd be right to wonder, how on earth could Android (and by extension, AOSP) have failed? Well, that depends on who you are, what benefits you, and how you view the increasingly closed ecosystems that form the backbone of our digital lives.

Android is based on its open source predecessor

Before we get into the hot takes, let's review a bit of history. Android itself is an open-source operating system developed by a consortium known as the Open Handset Alliance, though most of the heavy lifting has historically been done by project leader Google. The core of the OS is known as the AOSP, which includes libraries, documentation, and source code required to both create custom versions of Android and to get Android running on new devices. With the novel examples we mentioned above, i.e. the 787 Dreamliner, it's likely a custom version of the AOSP that is running, not the 'Android' interface that you are familiar with on your phone.

The 'Android' running on your phone is likely developed by Google, as a closed source modified version of the AOSP generally just known as Android. This is likely what you've used on most phones over the last few decades, and for clarity we'll refer to Google's familiar version of the AOSP as just 'Android' from now on. At launch, several versions of Android were still open-source. This didn't last long, however. The last version of Google's Android to be fully open-source was Android 4.4 KitKat, released in October 2013.

Google has slowly stepped back from the open-source roots of the AOSP

When the AOSP was founded, Google was contributing heavily to the development of open-source apps to be bundled alongside the free AOSP. This made sense at the time - by contributing development effort and resources to the AOSP, Google succeeded in growing the market share of Android/AOSP-based devices massively over the following years. But as the years rolled on, Google slowly replaced the open-source AOSP versions of common apps like Calendar and its messaging app with its own closed-source versions, ceasing development on the open-source version in the process. As the primary developer of many of these apps, this effectively killed (or massively stunted) development of open-source AOSP versions of these apps.

The effect of this has been to slowly tighten Google's grip on Android as a whole, as anyone wanting to fork the AOSP from scratch to build their own operating system would need to create their own versions of all of these apps.

Enter, Google Mobile Services

Now Google does offer access to its services, and its versions of proprietary apps, through a scheme known as Google Mobile Services. Google Mobile Services is the current iteration of Google Play Services, introduced in 2012. The intention behind GMS was to allow Google to take control of distribution of some elements of Android that relate directly to its services, pushing core updates to its versions of apps like Gmail, Drive and YouTube without the need for OS updates. This went some way towards solving some of Google's frustration with Android at the time, which was poor and inconsistent developer support to keep up to date with the latest versions of its apps that are core to the Android platform.

So what's wrong with Android?

History lesson over

Android's open-source roots were essential for adoption and enabled easy developer support for porting and integrating with a range of devices. This is what has really allowed Android to dominate the world. AOSP has allowed hardware manufacturers to have a simple and consistent platform to build on, for everything from entertainment services to game consoles. In this way, the AOSP has undoubtedly been a very good thing for all consumers. It cuts development costs for OEMs, making it easier for manufacturers to easily get their software up and running on custom hardware. This has supported a wide range of important innovations.

Where the AOSP has failed is for consumer devices. It's not the fault of the AOSP itself, moreso Google and other manufacturers building on it. The introduction of the AOSP offered a promise of multiple, open-source, inter-compatible versions of the AOSP, each capable of running the same apps on significantly different software, each maintained by different entities. In an ideal world, consumers could benefit from a range of companies all supporting their own implementations of the AOSP, with their own versions of apps (ideally even, open-source versions of apps) with interoperability between manufacturers. Now, the consumer market (at least for most western devices) is dominated by Google's version of Android, and modifications have been consolidated to a few major companies.

This domination didn't come overnight and, in fairness, was largely a result of Google proactively trying to fix issues with Android. OEMs have struggled to maintain consumer devices with new features, version updates, and security fixes. Android has also sometimes suffered with a reputation for bloatware and inconsistent user experiences. So we do understand the challenges.

Google has often tried to solve problems by offering first-party devices

Google has often tried to solve problems with the wider Android ecosystem by competing directly with other OEMs, pushing them to respond. For example, Google recently committed to 7 years of software updates for its Pixel 8, and with its commitments to very clean, timely, and consistent updates to the Android experience with its Nexus (and Pixel) lineups. The downside of this though, whether intentional or not on Google's part, is that the control of the consumer Android experience as a whole has increasingly been guided by Google.

OEMs building consumer devices are reluctant to make wholesale changes to Android, and even more reluctant to support them in the long term. This, combined with Google's own first-party devices, has made it difficult for competing versions of the Android platform to exist with significant variations. Ultimately, the requirements for security and feature updates, as well as long term support, require manufacturers to keep up in order to compete with the iPhone. This has led to pressure on consumer OEMs to stick close to Google's version of Android to make that possible. In the medium to long term, the rolling of more and more services into Google's own proprietary GMS license has also decreased the viability of forking the AOSP.

There have been other significant and sometimes sensible blockers in play. The SafetyNet API killed a lot of custom ROMs for users, allowing apps (and Google Pay, for example) to disable themselves if a device doesn't meet standards or pass certain tests.

A lost opportunity for consumers

Whether as a fork of Android, or the AOSP, the slow slide of Android-control back into Google's hands has been a missed opportunity for consumers. Google utilized the open-source bootstrapping of AOSP and Android to seize a wide market share, and it's undoubtedly been a collosal success, as they have built an ecosystem of apps where so many others have failed. But modern Android is increasingly becoming 'Google's OS,' with only surface level changes made by even large manufacturers. The aspirations of AOSP were to enable a platform for a range of diverse, novel operating systems from a variety of manufacturers with common app support, easy device forking, and shared standards. The promise of Android as an open-source platform for consumers has essentially fallen into Google's control under the pressure of iOS' relentless grind for improvement, security updates, and version defragmentation. That said, AOSP has benefited OEMs, and Google, massively.

I like modern Android. It's great, and a worthy competitor to iOS. But I can't help but feel the allure of a missed opportunity. Maybe I'm dreaming too big, but I wonder what modern Android experiences might have been like if manufacturers (and Google) had found more common ground in allowing custom versions of the AOSP to flourish. Perhaps the issues around version fragmentation and access to Google's services could've been solved in a way that allowed Google to walk behind its flock, instead of dominating from the front of the pack.