NOTE: Before you read this, I want to clear up some confusion. This post shares an idea that is designed purely for some intellectual fun and discussion. I am not proposing we actually do this, nor advocating for this. So, don’t read too much into these words…
The Ubuntu phone is evolving step by step. The team has worked their socks off to build a convergent user interface, toolkit, and full SDK. The phone exposes an exciting new concept, scopes, that while intriguing in their current form, after some refinement (which the team are already working on) could redefine how we use devices and access content. It is all the play for.
There is one major stumbling block though: apps.
While scopes offer a way of getting access to content quickly, they don’t completely replace apps. There will always be certain apps that people are going to want. The common examples are Skype, WhatsApp, Uber, Google Maps, Fruit Ninja, and Temple Run.
Now this is a bit of a problem. The way new platforms usually solve this is by spending hundreds of thousands of dollars to pay those companies to create and support a port. This isn’t really an option for the Ubuntu phone (there is much more than just the phone being funded by Canonical).
So, it seems to me that the opportunity of the Ubuntu phone is a sleek and sexy user interface that converges and puts content first, but the stumbling block is the lack of apps, and the lack of apps may well have a dramatic impact on adoption.
So, i have an idea to share based on a discussion last night with a friend.
Why don’t we rebase the phone off Android?
OK, bear with me…
In other words, the Ubuntu phone would be an Android phone but instead of the normal user interface it would be a UI that looks and feels like the Ubuntu phone. It would have the messaging menu, scopes, and other pieces, and select Android API calls could be mapped to the different parts of the Unity UI such as the messaging menu and online account support.
The project could even operate like how we build Ubuntu today. Every six months upstream Android would be synced into Launchpad where a patchset would live on
patches.ubuntu.com and applied to the codebase (in much the same way we do with Debian today).
This would mean that Ubuntu would continue to be an Open Source project, based on a codebase easily supported by hardware manufacturers (thus easier to ship), it would run all Android apps without requiring a cludgy porting/translation layer running on Ubuntu, it would look and feel like an Ubuntu phone, it would still expose scopes as a first-class user interface, the Ubuntu SDK would still be the main ecosystem play, Ubuntu apps would still stand out as more elegant and engaging apps, and it would reduce the amount of engineering required (I assume).
Now, the question is how this would impact a single convergent Operating System across desktop, phone, tablet, and TV. If Unity is essentially a UI that runs on top of Android and exposes a set of services, the convergence story should work well too, after all…it is all Linux. It may need different desktop, phone, tablet, and TV kernels, but I think we would need different kernels anyway.
So where does this put Debian and Ubuntu packages? Well, good question. I don’t know. The other unknown of course would be the impact of such a move on our flavors and derivatives, but then again I suspect the march towards snappy is going to put us in a similar situation if flavors/derivatives choose to stick with the Debian packaging system.
Of course, I am saying all this as who really only understands a small part of the picture, but this just strikes me as a logical step forward. I know there has been a reluctance to support Android apps on Ubuntu as it devalues the Ubuntu app ecosystem and people would just use Android apps, but I honestly think some kind of middle-ground is needed to get into the game, otherwise I worry we won’t even make it to the subs bench no matter how awesome our technology is.
Just a thought, would love to hear what everyone thinks, including if what I am suggesting is total nonsense. 🙂
Again, remember, this is just an idea I am throwing out for the fun of the discussion; I am not suggesting we actually do this.