Why Lollipop is Such a Pain to Get Working (And, our rollout this week)
Hello everyone, this is a busy week for us.
We’re going to be rolling out a lot of stuff… including our first Lollipop builds to backers. They’re being uploaded right now, and will be available over the next day or two. Another Kickstarter update will follow. Unlike KitKat, which we are also expanding access to this week, these will be backer-only for now.
We want to set the expectations right, so please note that these initial Lollipop builds are purely a technology preview, and not ready for general use. Device support will expand over the next month – so don’t be disappointed as we focused support on only a couple of configurations… we’re hoping to ship many more this month.
Why release a preview build? So you can see that we are moving forward with Lollipop.
While we are improving our Lollipop builds, we are going to continue to push and promote KitKat, Android on Core processors, and Android for 2-in-1's... while also refreshing our Wiki support documents with additional tips, tricks, and workarounds.
While we are improving our Lollipop builds, we are going to continue to push and promote KitKat, while also refreshing our Wiki support documents with additional tips, tricks, and workarounds. Our forums and the KitKat public rollout will arrive this week.
Many are asking why Lollipop has been such a pain. This is not a startup or Kickstarter-level-funding issue… it’s affecting every Android platform developer; giant and tiny.
The Technical Details
Both HTC and Motorola Mobility have recently come out saying the same thing - getting Lollipop to work has been much tougher than expected. No Intel-based Android device - other than Google’s own Nexus Player - is running Lollipop just yet. It’s no surprise then that over 99% of Android users (on the ARM or x86 sides of the aisle) are still running KitKat (or earlier).
At the same time, to grow Console OS, we have to push forward with Lollipop. In the future, supporting and sustaining KitKat would mean investing in code that Android/AOSP no longer supports. We’ve realized that it would be a losing game.
Just a mere two weeks ago, we were finally able to get access to all the private drivers and bits of code needed to support the same Intel Atom devices that we had targeted with our original release plans, now with Lollipop-capable drivers. We couldn’t affirm that we had the code… until we had it.
What we’ve found over the past six weeks (including, in particular, the last two weeks), is that Lollipop rewrites a lot of the fundamental code that gets Android up and running. That is why everyone is struggling with it.
Specifically, the AOSP 5.0 kernel lacks formal support for Intel Core processors, and the UEFI fastboot stack knocks down most of our existing bootloader support. Now that we have all the right tools, we can work on tackling these upstream issues.
On the bright side, Google & Co made these platform-level decisions so that major components would not have the rug pulled out from under them down the road. For example, UEFI devices on Android will now act identically to other Android devices - enabling us to push out firmware updates over-the-air with ease. But it will take time for us to rebuild the entire Console OS platform.
Two Big Things Coming Soon
Making good use of our time, behind the scenes, what we've been doing is building one reference platform that we can then branch out and support a broad range of devices.
We will be introducing that new reference device at Mobile World Congress, at Intel’s booth. We can’t share more on that topic just yet, obviously… so hang in there until the first week in March. It will show we haven’t been idle all this time, despite a laggard access to the bits of code we needed.
Opening Up Console OS
Finally, while Lollipop is quite painful, it has opened the door to us taking Console OS open source. The same changes that make Lollipop such a technical challenge, also give us the potential to finally open up Console OS.
We realize there are other Android on x86 solutions out there that are open source. There really are two - One that is not meant for end-users, and the other we have lost confidence in delivering commercial success. We think Console OS can deliver a new balance in this equation, and we look forward to taking it open source over the next few months. We can't do it just yet, but we are working hard on it.
We promise that when we do, we will put your Kickstarter money to work encouraging open-source developers to contribute improvements. Improvements that everyone in the Android world will benefit from.
While Lollipop has been a swift kick in the rear in terms of our development timetable, it has opened the door to us doing what we’ve wanted to do - make it so that anyone can build, and more importantly, contribute, to Console OS. And that is going to be our mandate for the first half of this year. In the mean time, we are going to open up Console OS DR1 with KitKat to everyone for free download, starting over the next week.