Android P is in preview and Kubernetes becomes the first project to graduate the CNCF.
Google released an initial developer preview of the next version of Android. While it’s probably missing some significant features they can announce at Google I/O in May; there are still some interesting improvements already in here.
Several of the improvements seem to be aimed at improving the hardware and software integration. It’s always easy to make fun of Android for not offering as integrated an experience as iOS until you remember that iOS runs on a handful of different devices1 while Android has to run on hundreds if not thousands of different models. Some examples of this better integration are support for display cutout2, a better platform-wide fingerprint API, and support for dual cameras.
On the fun to play with front, there is the new indoor positioning using triangulation with wifi access points and an improved neural network API for better on-device machine learning.
That said, I think that possibly the most crucial part of the preview isn’t so much what it adds as what it says it will remove under the goal of modernising Android. There are several calls to ensuring developers will update the SDK they target as Google Play will stop accepting apps (and updates to those apps) that are targeted at older SDKs and actively warn users when they install something that was built for really old SDKs3.
In addition, they’re going to close off access to private Android APIs, aka non-SDK interfaces, which are in use now. This will start with APIs that have a public alternative while seeing which APIs will need public versions before replacing those. Still, it won’t surprise me if this means that quite a few older apps that aren’t maintained will stop working in Android P or newer. The reasoning is obvious, restricting access to public APIs means that Google can work on preserving API stability there without needing to worry about keeping the underlying code the same.
As most of the changes in this release are focused on hardware support and underlying the platform, it does make me curious to see what else is in store.
Kubernetes Graduates From CNCF
Two years ago the CNCF was created as a committee to help guide Kubernetes. Since then it has taken many more projects on board and I mention it frequently on this site. However, now its very first project has been deemed good enough to graduate.
To be honest, in most ways this doesn’t make any difference for existing users of the project. The main advantage is that it shows off the maturity of the project and thereby is likely to convince some people that it is good enough to be used in their organisation.
Personally, what I find interesting here isn’t so much the graduation itself, but how the process works. There are definite criteria that need to be met for each status within the CNCF, and a committee decides if these are appropriately met.