Google today made the second developer preview of Android O available to developers. This second preview builds on the first version released in March and is stable enough for public beta testing. In addition to offering the new bits to developers, Google detailed Android O's many features.
Android O focuses on what Google calls fluid experiences and vitals. Fluid experiences include the new picture-in-picture tool, notifications dots, autofill, adaptive icons, smart text selection, and TensorFlow, while vitals include security, system optimizations, background limits, and a new dashboard console for developers.
The PIP tool lets people manage two tasks at the same time be creating a movable window for one app, such as YouTube. Google says the developer tools included in Android O make it possible for most apps and screen sizes to support it.
Notification dots are mandatory. Google says developers will have to support them in Android O and beyond. These let developers surface activity in their app in a wide variety of ways, while also giving end users a nearly endless way to configure blinking lights, vibrations, sounds, and popups. The good news is Google believes notification dots can be added to most apps with almost zero effort.
The Autofill function relies on a new API is meant to help solve the login issue. At its simplest, Autofill brings Chrome's Autofill feature to Android. End users will need to opt-in and save their logins within Chrome separately. Once signed up, Android O will post usernames and passwords to most third-party apps from Chrome. Google claims it will be relatively simple for developers to add this to their apps.
Smart Text Selection should help improve end user experiences within messaging and search apps. Google's David Burke said phone numbers, addresses, and business names and/or proper nouns are the types of text most often selected for copy/paste functions. WIth Smart Text Selection, tapping a phone number or address will automatically select the entire number, rather than call up the text-selection tool.
Last on the fluid experiences side of the coin we find TensorFlow. This is the core behind Google's machine learning. Google is making TensorFlow Lite available to developers, which it calls a fast and lightweight for embedded use cases. The company is working on a new Neural Network API that will work hand-in-hand with TensorFlow Lite in order to take advantage of hardware acceleration.
Google is taking security more seriously with Android O. A new tool called Play Protect will automatically scan apps to ensure they are safe to use on device. Google says it already performs some 50 billion app scans per day within the Play Store itself. Play Protect works separately on handsets.
The system optimizations are meant to help apps run faster and smoother. Google says it made extensive changes to the runtime, such as concurrent compacting garbage collection, code locality, and more.
Background Limits should help improve battery life Android O will more proactively shut down applications that are soaking up too much power in the background. It also looks at wifi and location use, and will look to free up memory when possible.
Developers should be happy to learn that the Play Console includes six new dashboards to help them pinpoint crash rates, ANR rates, frozen frames, slow rendering, excessive wakeups, and stuck wake locks. The dashboard will provide data on how these issues are impacting end users and guidance to help developers resolve the problem.
Google warns that it is early days for Android O. This second developer preview is not necessarily meant for primary devices. It is only compatible with more recent Google hardware, including the Pixel, Pixel XL, Nexus 6P, Nexus 5X, Pixel C, and Nexus Player. The beta build can be downloaded and installed over the air.