An experience with Ubuntu Touch development | Hacker News Reader

An experience with Ubuntu Touch development

zxstudio.org

99 pointscmsimike59 comments

I'll answer any basic questions anyone has about using the Ubuntu phone - been using it for over 18 months now. Never have actually owned an Android/iOS phone and hope to keep it that way. It does everything I want it to. Not 'beautifully', but it does it. It's cheap, open and supports free software.

Ideally I'm still looking for some alternative device that matches or surpasses the N900 in terms of functionality and openness. But I'm of course not holding my breath. With UT I get SSH and terminal access, but it is slightly hamstrung in terms of what can be installed from repos, unless I forego the OTA updates.

A quick ideal-world scenario: a phone/mobile PC store on the highstreet where you can browse great hardware running Red Hat, Debian and the BSDs. All dirt simple to use with the ability to pop the hood (terminal, repos) whenever you needed a little more tooling. Devices perhaps in two parts, one pocket-size and one bag/large-pocket size, slotting together for storage and charging.

UT is still working on convergence, and has made great strides from what I can tell, but it's still some way off being a useful reality for any more than a handful of devs.

what games do you like?

Started off playing through a few of the recommended games, but to be honest that isn't what I do with a phone. I mostly read or communicate. Literally no games on there at the moment.

To answer the question though, I did try tuxracer, obligatory, and pandolove (iirc) and a few of the abstract physics-type ones. Nothing stood out as massively replayable.

Was the lemmings clone (pingus?) ever updated with touch support?

I really couldn't tell you, other than I haven't seen it in the Touch app store. And I would have, trust me.

How often do you use the terminal and/or SSH, and what for?

I kept hearing for a long time how great $this or $that android ssh-client or terminal was, but they were all terrible IME. I am quite picky though, and spend >1/2 my day in terminal. All I want is a good terminal session, with ssh, on my phone.

I actually use it for personal more than work stuff. Connecting to the pi@home to restart kodi, set up a torrent, check IMs/IRC.

Changed a shortcut in tmux to get around the quirks with not having a keyboard (it offers a menu for Esc and common Ctrl+ shortcuts).

I won't tell you it's perfect, sure you're not expecting that. Serviceable though. Setting it up in 'convergence mode' could actually work for you, perhaps with a tablet instead of phone, if you're so terminal-based. Reminder: the phone cost ~£100.

What has been some hurdles you've run into using this phone?

- No email notifications - a blessing and a curse. The Dekko client is/was third party, but being the best available has become the de facto default (not a bad name for a band...).

- Inability to make use of the package repos. Mentioned this above - I only want some CLI apps and still may do it, I just have to find out if when I enable r+w mode that I can subsequently disable it when mutt/irssi/newsbeuter etc. are installed. Enabling it disables OTA updates, you see, which have proven massively useful. The browser has gone from awkward to extremely usable via the updates, convergence support has been added, 100s/1000s of bugs fixed - they are a must-have on a phone like this.

- Being Ubuntu, they had "agreements" in place with a few third-party providers that I quickly disabled on setting up. The always-on stuff might be convenient for some, but it's not for me.

- My particular model of phone, the BQ Aquaris 4.5, struggles with location data, usually placing me at the ISPs location for Wikipedia nearest article suggestions, etc. It does work for the uNav mapping app just fine, however.

Upsides: I can (and have) reported bugs and chatted to devs on IRC and the central mailing list.

Downsides: There aren't many devs knocking around!

I think they need to support more flagship phones, or popular ones like the MotoG of Nexus phones. That takes away risk, you can buy a good Android Phone and play with Ubuntu Touch. The Meizu's are not so appealing to me.

Currently the OnePlus3 port is under development [0], I think it would be great if that starts to work. The OP3 also has 6 GB of RAM, making it very well suited for Convergence (their name for getting a desktop interface when attaching mouse/keyboard/monitor).

[0] https://devices.ubports.com/#/op3

They seem to have the same problem Jolla has – a phone OS without devices. Every once in a while a smaller hardware manufacturer (BQ, PuzzlePhone, FairPhone, Turing Phone, etc) makes some promises or actual releases a model with some support, but those companies are struggling themselves for any real market share – or mind-share. I guess the margins are too low to grow fast enough and get leverage against the big manufacturers and the default Android choice.

I really hope one of these more open-source friendly companies gets a foothold in the phone space, even if it happens by courting markets in countries like Russia and China who are politically less than thrilled about handing over too much information to American companies like Google and Apple.

For the moment, they should if possible be selling phones that dual boot Android and Ubuntu. As you say, that would make it zero risk for the consumer, either that they'll buy an Ubuntu phone and end up with a phone that doesn't do everything they need to do, or that they'll have to buy an Android phone and spend an unknown number of hours or days fiddling around to sort out a Custom ROM. Android compatibility is unfortunately necessary, two examples for me are for transport apps and parking apps. I personally would be happy to buy a phone like that, but a pure Ubuntu phone will not do it for me.

I think if they did this, they could get a decent percentage of tech-orientated people to sign up. It may be they could make a mass market bid in two to three years if at that stage they are the only company offering cheap smartphone convergence, but for the next few years they need to target the tech/OSS crowd where they already have name recognition and goodwill.

I hope this project takes off. Would love to have a completely open source mobile phone or tablet.

I think the key is low entry barrier for developers. Developing an SDK/IDE is too much work, and personally I feel locked in and rather use the command line instead. A fast emulator or a KVM image would also be nice.

You can buy one now. Not open source hardware, and there's probably closed binaries in there somewhere, but you can buy a BQ Aquaris M10 with Ubuntu Touch installed. The thing itself has a command line and in theory you can compile your own programs and install packages with apt, though you are heavily discouraged from doing so.

(the software is not very good yet, and it's not really usable, but I'm biding my time)

I hope there is a niche where Ubuntu Touch can fit in. I have the Meizu Pro 5 with Ubuntu Touch, and it suits me just fine. The reason I chose this phone is that it comes as close to a free software smartphone as I can get now.

A huge part of wanting something other than Android is that I don't feel comfortable with the current app-driven smartphone world. Everything seems financed either via analysis of what I do (targeted advertising) or subscriptions with vendor lock-in. Even the most popular apps require access to data they shouldn't need.

I also dislike the way a lot of services used by sometimes even the majority of people (i.e., WhatsApp) are exclusive to pre-approved smartphone operating systems (even desktop operating systems are excluded without tethering!). The delicate counterweight we have on the desktop (between major commercial OSes and free software OSes) is missing on the smartphone; it boils down to a duopoly of Apple and Google, and a whole bunch of service-silos, and there is no threat of users migrating to an OS with more freedom if you push them too far (as I believe there is on the desktop).

You do you do great UI design with the command line?

I've had a marvelous experience building HTML5 UIs in my command-line vim.

I bet you had lots of productive hours save/browser refresh fun cycles making something like Material design with that workflow.

What about the users, have they had a marvelous experience using that UI?

There exist tools outside of your apparent zone of comfort that let us do whatever we damn please from the command line. I think you'd be surprised how much user facing code was written using Emacs/Vim.

I was already coding in the 80's, back then it was called VI.

So I know pretty well how the UNIX System V experience for GUI programming looks like.

And I used Windows already back in the 90s, so I know exactly what the PC experience looks like with all its BSODs and token rings.

At least it already had RAD GUI tools that cared about usability while most UNIX terminals that I was still using in 1994 were green phosphor ones.

Well, speak for yourself. About the same time, our local university had very impressive Irix X-terminals with a rock solid GUI.

It might be that you have only had a small glimpse of the things beyond PCs back then. And that seems to still have an impact on your perception of things on the "Unix" side ;)

You mean like this in 1991?

https://www.youtube.com/watch?v=UGhfB-NICzg

I would have rather used the GUI tooling of a NeXT, which eschews the UNIX traditions even though it builds partially on top of them, than the Solaris way of programming GUIs, if given the option.

Just because I could only access green phosphor terminals, doesn't mean I wasn't aware of other, more expensive, options.

How do you develop for the web?

You have a picture in your head. You write down the XML in a text editor. Then you load the app, and the picture comes alive.

So productive, versus using a graphical design tool with real time feedback.

Why does it bother you if someone prefers a different tool to get the job done? It doesn't hurt or hinder you in any way....

Because then Bret Victor comes along, does a presentation about development workflows what used to be already partially possible with Xerox PARC systems and everyone is amazed why our computers aren't like that.

Bret Victor The Future of Programming: https://www.youtube.com/watch?v=8pTEmbeENF4

Another good example is the NeXT vs Sun duel, regarding the RAD tooling of NeXTStep vs the traditional UNIX development (1991).

https://www.youtube.com/watch?v=UGhfB-NICzg

My experience with graphical tools is that they are good when doing something quick, but when you need to do something unordinary you'll have to dive into the code, and these tools usually don't produce nice clean code.

I also think it's a matter of personality or personal taste. I'm very good at thinking in abstract.

I think the opposite, having used Delphi, VB, Netbeans Matisse, MS Blend among many others tools, which allow one to go very far without doing change code/compile/reload/swift a few pixels/loop.

Also you see this with game editors used by most game studios. No one is designing levels by hand nowadays.

But lets just agree to disagree. :)

I feel like the ease in which developers can make applications for a new system plays a pretty vital role in that system's success.

For example, the Sega Saturn (from my understanding) was ridiculously hard to program for, and consequently a lot of the games for it were objectively worse than their PlayStation counterparts (with some exceptions).

My point here is that if Ubuntu wants to compete, I think it's incredibly smart to make it approachable for devs to just jump right in; I know the entire experience isn't polished right now as evidenced by this article, but it doesn't look insurmountable, and I'm optimistic.

The Saturn had 8 processors! Two main processors, two sound processors, two graphics processors, a processor for controlling the bus and interconnect, and a processor dedicated to the cd-rom.

That's what I mean!

We're only now getting to a point where people really know how to take full advantage of multi(core/processor) systems and are given nice frameworks in which to do so; in the mid-90's I'm pretty sure that that was a really difficult task!

"For example, the Sega Saturn (from my understanding) was ridiculously hard to program for, and consequently a lot of the games for it were objectively worse than their PlayStation counterparts"

And then, interestingly, Sony repeated Sega's mistake with the PS3 which arguably lead to a lot of the success of the Xbox 360 in non-Japanese markets.

In the current generation, both consoles made the right choices in terms of ease of development, but Microsoft still got themselves off to a rocky start with the whole digital-downloads-only decision that they later backed off from. To be honest, I think their decision there was ultimately the one that is going to be standard in the future, they just pushed too hard, too fast when the market wasn't ready for such a big change.

What I really want to know is what the adoption rate of Ubuntu Touch is among the hacker/hobbyist/diehard Linux crowd. Firefox OS has lost its official support on smartphone, no idea where Sailfish OS is at, so Canonical should be doing everything they can to make their platform the alternative to the big 2. Are they?

Im happy to mess around with ubuntu on a laptop because if a stuff it up completely its not hard to get a windows imgae and start again. This is not the case for mobiles or tables at the moment.

It's not that hard to flash a Custom ROM on most Android phones. I mean, it is a major pain in the butt, as you usually lose your warranty, you don't usually get the original ROM as you already hinted, and dual-booting is only possible on a handful of devices, but if you for example already have CyanogenMod flashed anyways, then the first two of those problems are already out of the way.

It seems quite low I think. Part of the problem is that Android is just so ubiquitous. As the article mentions "Developing for Ubuntu Touch right now is not easy, mainly because of lack of documentation and examples". Android has a million examples, libraries, first and third party documentation, a load people at Google improving the tools all the time (though this churn can also be a bad thing of course...) Canonical has less of this and the development plan seems focused on Web apps.

Another problem is that there are no Ubuntu phones currently available to buy new. The Meizu mentioned in the article is no longer sold, and installing on semi-supported hardware is flakey. For example, you can buy a Meizu phone that has Android and reflash it with Ubuntu Touch, but then future OTA updates don't work and you have to reflash each time. Devices sold on ebay might not be the so-called "international version" and therefore have a locked bootloader, you don't know until you buy one.

The way updates are handled and the community support seem much more organised than with Firefox OS though. That was a complete disaster, with the first gen ZTE phones receiving no updates at all and even the Flame (a developer device) being pretty much abandoned right away. At least Canonical keep pushing OTAs for the official BQ and Meizu devices, and builds for Nexus devices are still produced so you can flash it on semi-common hardware if you're feeling brave.

I know it's easy to say and very difficult to do, but really what Canonical need is a phone that has some long-term availability. You often hear Alan Pope on podcasts use the excuse that it is not Canonical's problem that devices are not being sold, it is up to the manufacturers. But if there are no devices to buy then how is UT ever going to gain any real traction?

You can actually flash a Meizu and continue to get OTA updates now.

Not really meaningful for anyone wanting to make a living of selling his/her application developer skills.

Windows Phone with its meager market size has more than all 4+ places summed together.

That wasn't the question:

> What I really want to know is what the adoption rate of Ubuntu Touch is among the hacker/hobbyist/diehard Linux crowd.

It's not at all about selling one's services or gaining income from "apps". It's about gauging if there's enough (community) momentum that at least for this group of people, UT could turn out to be an interesting alternative.

Without developers willing to learn to code for the platform the community won't have any applications worth using.

I understand it's a circular problem- an app ecosystem cannot grow without developers, and devs want to invest only in ecosystems that have users. Or do they? I feel the thing about these alt-OSes is that without mainstream appeal, they need to go hard for the hacker/hobbyist/diehard minority, who are motivated by things other than the possibility of attracting mainstream customers. Ubuntu, as a Linux distro, already has that appeal- the trick is to expand that to mobile.

I am willing to bet it is almost zero. Android does the job for most people that crave modding-ability/customisation, so Ubuntu Touch is really only interesting for a very small niche.

Among the people i know who really care about open source and security, the max is running Android without the Google Services, but i have never seen Ubuntu Touch in the wild.

The problem is, even most apps from Google's competitors now won't run without Google Play Services. Android is becoming increasingly useless as an open source OS.

The diehard crowd and maybe with people that require better security might use it.

For everybody else, they OS is there only to support the applications they want to run. I'm a Linux fan, but there's no way I would be happy with it for my phone OS.

Google cache mirror http://webcache.googleusercontent.com/search?q=cache:qComzDs...

How does Ubuntu Touch compare performance wise (battery and speed) and developer friendlyness with that other open OS Tizen, which uses Wayland instead of Mir for display operations? Does Ubuntu Touch have an Android compatibility layer?

As Rasterman (Carsten Haitzler of Enlightenment fame) is involved and Samsung intends to run it on every device I think it might be more efficient. It also has a socalled 'Ultra Power Saving Mode' built in.

The Tizen Samsung Z2 has been released for 70 euro in India last week and in Africa soon Tizen might be destined to take a significant portion of the mobile space. The development tools seem open enough, you could even run UE4 html5 games (that goes for both Ubuntu One and Tizen I assume) https://wiki.tizen.org/wiki/Game_development

Tizen already went through so many reboots on their SDK stack that I wouldn't bother with it.

It started with Meego adoption, then they brought the BADA OS SDK C++ (think Symbian C++ like experience), followed by re-writing everything in C, and offering a new C++ API on top of it when developers rebelled to go back to C.

Also regarding the HTML5 stack, there are lots of dead mobile OSes that tried that route already.

I doubt anyone in the markets where Tizen was released is bothered to target it instead of Android, unless Samsung is paying them.

I would love to know if there are any developers out there writing for Ubuntu Touch. I got a Aquaris BQ M10 tablet and it has virtually no apps available. What comes preinstalled is anaemic and barely works. And the promised 'desktop' feature is DOA.

It's gathering dust until some more developers come and make stuff for it!

there are some apps for the tablet and phones https://uappexplorer.com/apps?sort=-points

Thanks! That links broken for me though.

Could it be your browser / https / an extension? It works for me in Firefox and Chrome but not ie.

Oh yeah it works in Chrome but not on my mobile browser (I just got a blank screen).

A JS issue perhaps? Seems it shows crap all without it turned on.

Ubuntu Touch and the idea of a device capable of fulfilling two modalities (active typing, passive reading) is wonderful.

Unfortunately most hardware and firmware is as awful as software. I pity Intel network driver writers and Android ROM writers...

Companies spend millions working around the imperfections of software logic and hardware creations. In a bubble Ubuntu philosophy is wonderful. In reality, it seems, we are collectively far from realizing the brilliance of our own creations in simple terms.

Reply on news.ycombinator.com