Some thoughts about what could have gone wrong ...


Robert Taylor

Very Active Member
Joined
Jun 7, 2017
Messages
208
I ran into this not very interesting thread on reddit:
Which was linked to a VERY VERY interesting article here: http://www.lieberbiber.de/2017/06/20/my-ubuntu-for-mobile-devices-post-mortem/

Okay, that is just one devs opinions and there are probably a hundred other perspectives. However, what I find is just HOW MANY bad engineering and design choices were made.

On the one hand, it is indicative of Ubuntu. Ubuntu started off as this utterly fantastic project supported by a wonderful millionaire (hi Mark!) that was going to delta off of each stable Debian release and merge back change. Somewhere along the way it looks like Mark turned over the running of Canonical and Ubuntu to some well meaning but awful leadership, and the Ubuntu phone is the result.

Software is hard. Hardware is MUCH MUCH harder. The pyra project seems to be a stark contrast to the way the Ubuntu Phone project was run. Slow, steady, tight focus, financially conservative, engineering resources very carefully allocated where it needs to be. What is amazing is just how much could have gone wrong, in light of MUCH MUCH better financed and staffed projects totally failing where even mediocre success seemed within grasp.

Keep on going guys. You are doing really well.

Whe evildragon and everyone else is ready, I would love to see some post mortem on the project. Surely a well deserved vacation is in order.

I'm sending beers!

:)
 
Indeed, an interesting article. I tend to think it's much harder for a company that's used to a certain level of performance and income to think their way down to a smaller and less profitable product. The Pyra hardware is quite a unique combination of lots of off-the-shelf parts. Phones and tablets are also made out of off-the-shelf parts unless you're apple or samsung (and even then their stuff is heavily based on existing stuff), but they generally don't tend to try to cram so many ports in, which take up a big chunk of the Pyra mainboard. And they are able to and used to throwing engineers at a problem, while ED has to be more circumspect about recruiting the same.

On the subject of Ubuntu in isolation, it's interesting that this has not only caused the canning of Ubuntu Mobile, but also of Mir and Unity. While I'm glad they're returning to using more debian-derived stuff, it does mean they're sacking off most of their software and hardware development of a good few years. A rather big mea culpa.
 
The problem with GNU and smartphones is that they just don't get along with each other.
GNU/Linux is a system in which freedom is guaranteed by license. You have a system with source code. You can analyze it, you can publish about it, you can modify it as you need, submit a patch or publish the code as fork. The only limitation is virulent nature of license. There are PC components which are firmware-less and there is even free software BIOS implementation like Libreboot or Coreboot.
Smartphones are totally opposite. The hardware is restricted by design. There is a secret agent sitting in a system bus, it's called GSM modem. Most smartphones don't have electronic switches to cut its power, and sometimes even this may not be sufficient (the old lab rule that device with disconnected Vcc and connected to bus will power itself enough to work unstable way and hang the system - maybe it's used intentionally in modems?). The modem uses internal firmware which is not only non-free, but very restricted in terms of its analysis and creating of alternative. The market for smartphones is dependent on paid "apps" which, if a whole GNU toolchain could be supplied, would turn out to be the most expensive 15 lines of Perl code :).
So they made a device for nobody. The target for Ubuntu phones would not be a young technology fanboys, as they pay for these browser wrappers with spyware add-ons (apps) and Android is sufficient for them. Being active anti-net-neutrality (which is a direct result of supporting these "apps") they have no idea about GNU and would not even try it. Mobile Ubuntu is also not for rich techno-fashionists, they choose Apple because they see it in movies. The only niche Canonical could get is power users. For some reason, most power users which use GNU toolchain for automation, are skilled enough to read punched cards by touch :). Definitely they know how computer works internally. Did they think that these geeks won't see the hardware malicious component?
GNU handheld for power users is a niche, I think Pyra may also fit in there.
 
To say that is what phones are, is not historically true.

What makes it even interesting, or possible, is that you can load user defined software on the part of the phone that hasnt changed.
There are all these inroads into phones, and if they fail on account of SDRs and FPGAs not being employed, then i think we will at least get there first, if not to a point
where everything is computer-like.
 
Up until the project was cancelled, most people silently assumed to be able to run the same applications as on the desktop (Firefox, SIP clients etc.) and manage the phone with apt-get, and that’s where the whole marketing became simply misleading.
Well, I would have assumed the same thing. If I ever had considered buying myself a smartphone or tablet in years past, I would have considered Mobile Ubuntu, not knowing it was such a crap (well, probably I would have found out and not bought it, but as it is I just didn't care).
So they actually rode out on a quest to replace every system component with something self written that is incompatible to everything that came before, that they never got to work properly and that nobody wanted. Great, we have that already, it's called Android, except the does not work part. If it was EDs intention to do a similar project, I would stay as far away from here as I can.
Mobile Ubuntu looks like a giant disorganzied software project that crapped out some not so special standard phones on the way to run it on, while Pyra/Pandora is mostly hardware design using an existing software ecosystem. I'm not sure how to compare those.
Software is hard. Hardware is MUCH MUCH harder.
Strange, what I get out out of this story is that it's actually the other way around.
edit:
But I don't think it's that easy to say one thing is harder. It's basically as hard as you make it to be.

I guess the hardware analogy to Ubuntu Mobile would be if EDs approach would be like "I want a handheld PC/gaming device, let's start by designing our own CPU with it's own instruction set first!"
 
Last edited:
Back
Top