Is Nexus 7 the first hardware of choice for Ubuntu developers?
I would love for them to start with Microsoft Surface (with keyboard as standard in some models it would kinda be the logical choice) but I understand that Nexus is a) earlier b) cheaper c) you have to start with something.
The transformer tablets also suit this, and also suits your responders desire for an unlocked bootloader.
The reason for the Nexus 7 focus is simple. It is, by far, the most popular while being very powerful tablet with market penetration.
I still think Canonical has this misguided idea that average Joe will go through the process of unlocking their Nexus 7 to install Ubuntu, but it has a reasoning behind it.
I would much rather see Ubuntu on my Transformer Infinity, since it has hdmi out and a keyboard dock. I could easily plug it into a monitor and have a desktop like experience that way.
> I would much rather see Ubuntu on my Transformer Infinity
I've loaded up Ubuntu on my original Transformer.
Granted, it has a Tegra 2 CPU, which to put it mildly is not very powerful, but Nvidia has put out the required source and modules to put up a Linux distro which takes advantage of it.
Maybe Ubuntu just has gotten "bloated" over the years and everyone has failed to notice because all PCs are overpowered these days, but when I booted it up on the Transformer, a machine which runs Android flawlessly, smoothly and responsive, I was severely disappointed.
Everything. Was. Dead. Dead. Slow. Prepare to wait multiple, multiple seconds to open that terminal. And please don't talk aobut browser-performance. Ouch.
Maybe it will run better on the newer models, but trying it out in real life was a little off-putting. On the other side: If Canonical is putting in effort to get it running well on the Nexus 7, it should translate to an overall performance-improvement across the line.
> Maybe Ubuntu just has gotten "bloated" over the years and everyone has failed to notice because all PCs are overpowered these days
When my terminal after starting unity is pid 12000, you can guess bloat has occured. Even on my arch boot my terminal after auto login through the x virtual console is pid ~500 still.
Also, Unity uses at least 300Mb of memory, and composting. Compostinggggg.
Most of the inefficiencies in a modern Arch distro, for example, are build around plug and play mostly I think. Udev, systemd, dhcp, etc are all designed to gently allow arbitrary hardware arrangements without interrupting the experience, but cost performance to determine what is available and constantly babysit the environment.
The bright side of this push to get things running well on the nexus 7 is that these inefficiencies will have to be dealt with. I'm running this build the Nexus7 now and, yeah its slow, but as the goal for 13.04 is getting the performance and battery life acceptable, these developments 'should' leak back and help the whole Ubuntu experience. It might just be the itch required to get development focused on performance again.
I led the team that shipped this image, and I'm pretty proud of what we did. A few thoughts about our goals for this release.
The entire point of Ubuntu core on the Nexus 7 is to highlight our desktop's performance and resource issues. We know we're way too fat in terms of memory consumption, CPU usage, disk footprint, etc. and now we have a convenient developer platform that folks can use to help us optimize our core in preparation for a future world where mobile dominates. (nb, I've been calling it Ubuntu Pilates)
The great thing about Ubuntu on the Nexus 7 is that it finally provides a convenient, cheap, ARM platform where all the standard Linux tools work. Believe me, I've hacked on pandas, rpis, etc. and for what we're trying to do here, the Nexus 7 is so much easier to develop on.
And if I may insert some editorial, I often see the HN crowd complain about Apple's developer policies, working around strange bugs in their black box APIs, etc. This is your chance to help build out an open platform. I'm not saying our APIs are better (in fact, they tend to be less well thought out than Apple's), but at least you have a chance to help improve things in the platform, rather than accepting whatever the platform gives you.
In any case, the summary here is that for now, we've got a tight focus on improving our core OS footprint so don't expect that our current UI experience is great (it's not) or that it's a usable replacement for Android (it's not, unless you hook up a USB keyboard/mouse in which case it's just a super cheap, silent terminal).
Every bit that we improve the core OS on the Nexus 7 flows back into the rest of our platform so our desktop and our server gets leaner and faster. To make it painfully obvious, this will help all your Amazon EC2 instances. :)
We'd love to have any help. And stay tuned for more to come.
thanks, /ac
I'm a little surprised that their solution for installation is to replace Android. I would think dual-booting would be easier to get developers to commit their own devices for testing.
Is dual-booting not possible/really difficult on the Nexus 7?
There's no real boot framework on mobile devices the way there is for PCs. Android boxes have traditionally had the ability to boot into one of two kernel+initrd environments: "recovery" and "system", and that's it. The hardware OEMs make this framework work via whatever tricks they want to play, and then ship.
Obviously a second stage bootloader could be written, but it's going to be highly hardware-specific -- one for the OMAP4-based Galaxy Nexus wouldn't work with the Tegra3-based Nexus 7, etc...
Could you at least write a rebooter that would rewrite /boot and then have two "partitions"? You could go into Ubuntu and choose "reboot into Android" and it would rewrite /boot and then restart the device?
I wonder how the bootloader for the Touchpad works. It had the functionality I describe though I have no idea how it was implemented.
Yes, you could do that. On all the devices I've played with the kernel+initrd image is a standard format and exposed as a single partition. You'd just dd your new one over it and reboot. But this would be fragile, to say the least.
Being an Ubuntu user, and a Nexus 7 owner, I really can't see this particular implementation being of much use – the strengths of the Nexus 7 are things like Google Now, location based stuff, it's speed and ease of use etc etc.
Ubuntu is great for doing computery things, like coding, or setting up servers etc. On a tablet, I'm not coding (too fiddly to type $[]@% etc), and not using it as a server, but I am using it as a map, as a book, as an email machine etc.
To be honest, I'd rather run Android on a laptop, rather than Ubuntu on a tablet.
Weirdly enough I'm the other way around. I have an android 2.3 device and find the UI frustrating. A mobile version of Unity with different lenses and searches would be far better, though I would want access to android apps.
I would find this useful assuming I could connect a keyboard and mouse to it. Assuming Android and Ubuntu could be dual booted quickly enough and I had a decent stand for tablet I could forgo a laptop entirely.
It looks like this is for testing key components on a tablet platform, not something that will be promoted to end users (yet). For that, it makes sense to aim for a device that's already out there and has sold well.
Or, as the post puts it: "To be clear, this is not going to be a tablet Unity interface ... we can ensure pieces such as the kernel, power management and other related areas are working effectively on a tablet device."
If the Nexus 10 is sticking with Tegra apus, the Wayne series won't be out until next year, so if they target an xmas release on the Nexus 10 then they are still stuck on the same Tegra 3 (probably overclocked).
I would love for them to start with Microsoft Surface (with keyboard as standard in some models it would kinda be the logical choice) but I understand that Nexus is a) earlier b) cheaper c) you have to start with something.
Right?