In Depth: Next-gen mobile OSes: what's in store?

[sincity]

It's hard to pinpoint the date exactly, but sometime in the past two years, Linux became the dominant OS for smartphones, and in the last couple of months, reports are starting to show that it's doing the same on tablets. All this is thanks to Android.

However, Android doesn't feel like real Linux. Originally, the kernel forked off from Linux (though the two are now being merged), and the way that closed-source components have been bundled with it leaves a sour taste in many open source advocates' mouths.

The Linux world hasn't just watched as the tech titans fought it out. Various companies and organisations have been hard at work designing new mobile operating systems built on Linux. Some of the OSes are based on new ideas and concepts while others are hackable, open versions of existing systems.

It's almost certain that some of the systems we'll look at here will quietly die off without ever making much of an impact, but with the right combination of technical skills, design flair and business acumen, one of these could take the world by storm and usher in a new era of mobile Linux.

• Mozilla on Firefox OS: 'what we're doing has a very good chance of working'

The downside of the mobile marketplace is that it's much harder to install new operating systems on devices than it is on desktops. There are fewer standards, and newer systems typically only support a few devices. The old Nexus 7 is one of the most popular, and our test system for this article.

At the time of writing, only one of the systems is currently available in the shops (Firefox OS), though in the fast-paced world of mobile phones, more may be available by the time you read this.

Cyanogenmod

It's Android, but now as we know it.

CyanogenMod, is built using the Android Open Source Project, and it's a free OS that's very similar (from a users point of view) to Google's version. We set out to convert our trusty Nexus 7 into a completely open source device to see how this compared to the normal blend of open source and proprietary.

Installation proved painless; we just had to make sure we had the appropriate tools installed (the android-tools-fastboot package in Debian and Ubuntu based systems), then we followed the instructions for our device from www.cyanogenmod.org. The result looked pretty similar to how our device looked when it arrived fresh from Mountain View. Even the wallpaper had the familiar Android 4 swish. There were a few more options that allowed us to customise the device more to our liking, but in six months of using the Nexus, this writer has come to quite like the Android default.

While the OS may be similar, none of the Google apps come with CyanogenMod. What's more, there's no Google Play, so we couldn't install them. Actually, this was a choice we made. It is perfectly possible to install Google Play on CyanogenMod, but doing so was against the spirit of creating an open source environment on our Nexus 7.

First, we wanted a decent web browser. Without this, frankly, a tablet is almost useless to us. The stock browser that comes with CyanogenMod is ugly as sin, and a bit unpleasant to use. On Android, we'd used Chrome, but in keeping with the openness of this challenge, we went for Firefox. The installation file is available from https://ftp.mozilla.org/pub/mozilla.org/mobile/releases/latest/android/multi. With this, we had a first-class mobile browser on our tablet. If nothing else worked, this could just about be acceptable as a computing device.

Fortunately, though, we weren't done here. We don't have to abandon app stores just because we can't have Google Play. F-Droid is a market place of free software for Android and CyanogenMod. Just download the installation file from their website and you'll get a wealth of open source software to run on your device. It's not as large as Google's store, so if you drastically need a thousand different ways to view pictures of kittens, you may be disappointed, but it does have apps for most needs (Android users take note here - F-Droid is an excellent place to find apps that aren't full of spammy adverts even if you're not ready to go down the fully open source route).

Everyone uses their tablet differently, but for us, access to cloud-based files and calendar are important. We've already switched from Google to OwnCloud for these, and we could access them through a browser. However, we found aCal and OwnCloud on F-Droid that enabled us to do these things.

OSMAnd stepped in as mapping software, and offers offline map browsing that Google has removed from its own product. The stock email and music apps worked fine, and we were pretty much set up for all our productive uses. The less productive uses of our tablet, though, looked a bit bleak. There are some games on F-Droid, though the range is far from overwhelming.

In installing Firefox, we also got access to the Firefox Marketplace where you can install apps, including games. This opens up more options, and hopefully, as Firefox OS becomes more popular, more and more apps will appear in the marketplace.

Unlike F-Droid, though, not everything in here is open source. At this point, we allowed ourselves to bend the rule. You can view this as pragmatism or weakness depending on your point of view. Over the years, we've purchased many of our favourite Android games through the Humble Bundle, and these can be installed on CyanogenMod in exactly the same way as Android.

While they're not open source, they are DRM-free, and in purchasing them, we supported the EFF, so that's good enough for us. At this point, our freedom-respecting tablet was fully set up, and every bit as useful as it had been under Google's Android.

Ubuntu Touch

Bringing a controversial desktop to phones.

Since its inception, people have been saying that Unity is designed for touchscreens, so it hardly came as a surprise when Canonical announced that they were launching a touch-screen specific version of Android. Of course, not everyone was expecting a phone-specific version though.

The Ubuntu Edge indiegogo campaign may have come and failed, but Canonical is still hoping to launch a phone in the first quarter of next year, and has an impressive array of partners on board to help them. On first boot, you'll either be relieved or disappointed to see that the tablet doesn't have the same Unity desktop that Ubuntu is famous (or infamous) for.

The Touch user interface, however, obviously takes it's inspiration from the desktop environment. At first boot, it also quickly becomes apparent that this is an early developer release. There are plenty of foibles on show, unimplemented features and other annoyances.

The power consumption, for example, is so bad that the Nexus 7 looses charge while still plugged into the mains and under moderate use. This will quickly get annoying if you use your tablet away from power sources. Of course, it does keep your hands toasty warm while you use it.

To give you an example of the current state of development, we tried to play our music on it. The first problem lay in getting the music onto the tablet. As yet, Ubuntu Touch doesn't support any USB data protocol, so we couldn't plug it in and transfer things across from the computer. Next, we tried to download them from our OwnCloud backup server, but the JavaScript in Owncloud crashed the browser. In the end, we were able to do it by using the Android development tools to open a shell on the tablet (adb shell), and install the openssh-server package, then SFTP our data across (the username/password is phablet/phablet).

With this achieved, we then found that the music player couldn't actually play the MP3 files. With all that said, there's no point in us complaining that an early release isn't complete yet. Of course it isn't. Instead, we will look at what we can glean about the OS from it.

Ubuntu Touch is centred around Scopes. These work in more or less the same way that they do on the Unity desktop. Their purpose is to give the user a view not just on their own tablet, but the entire digital universe. The Scopes that come as standard allow you to search commercial music and video collections, as well as viewing your apps and a home screen. These are fine to give you a feel for who the system will work (and, incidentally, look gorgeous), but they're nowhere near good enough for a final product.

On the phone, these are far more important than on a desktop. Currently, it feels like Scopes only exist as an opportunity to sell content to users (music and films), rather than to help the user do something. We hope this will be rectified before the first release.

It's the feel of the user interface that really makes Ubuntu Touch different. Much of the control comes through swiping from off the edge of the screen onto the screen. Android 4 users will know this from the way to launch Google Now. Swipe onto the screen from the top, and you'll reveal the Settings menu, from the right will bring out the Launcher, the bottom will produce the Current Apps menu, and use the right to switch between currently running apps. The last swipe provides a really satisfying method of window management that almost feels like flicking through the pages of a book.

We did, however, find these gestures awkward until we liberated our tablet from its case. That pretty much brings us to the end of Ubuntu Touch. Perhaps you might expect us to talk about the app store, but at this point there isn't one. You can install extra software by logging in via your computer using adb shell and then using the familiar apt-get.

Firefox OS

A mobile phone OS built from web technologies.

Mozilla aren't an obvious organisation to enter the mobile phone industry - at least not as an OS developer - though this might be a wise move by the browser maker. In the past few years, there's been a big switch to WebKit based browsers (like Chrome and Safari). Though it's losing desktop market share, Firefox is still a major player.

However, in the mobile world, it barely registers as a contender. It is available for Android, but it has so little market share that it's hard to count. If the fashionable option that desktop computing is dead really is true, then Mozilla, too, is in serious trouble unless it can reverse its fortunes in the mobile market.

Even if (as seems likely) desktop computing does keep going, it's clear to us that mobile computing is growing and will soon account for a large proportion of web traffic. Many people view Firefox OS as an interesting side project by Mozilla, but we think it's more crucial to their retaining a dominant position in the browser wars. From a technical perspective, the Firefox OS is interesting because everything in the user interface, from the Marketplace to the phone dialer, is written in web technologies, such as HTML5 with CSS and JavaScript. This means it all fits together with the web in one seamless experience.

The fact that it uses web technologies doesn't mean you have to be online to make the most of it, though. The system includes various hooks that allow you to download apps to use offline. Through various bits of magic like WebGL, developers can also take advantage of graphics acceleration and other hardware features.

The operating system's interface isn't particularly revolutionary, but it is simple to use. This may be partially due to the fact that Mozilla has decided to initially target lowpower devices in developing countries.

Plasma Active

KDE for mobile devices.

Most mobile devices, from iPhones to Androids are designed primarily for consuming content. They're great for watching videos, surfing the web or listening to music. However, they're not so good at creating content.

This isn't necessarily a problem because most of us aren't looking for a device to write an essay, or do some programming on as that's what our main computers are for. The mobile nirvana is a device that can bridge the two areas. That is, be simple enough to use easily with the touchscreen interface, yet complex powerful enough to do real work on.

The KDE community along with BasysKom and open-six are trying to perform just such a feat with Plasma Active. This is a version of the popular KDE desktop built specially for mobile devices. Since it's taken from KDE, all the popular KDE apps are available, including the Calligra Suite.

Even from a quick glance, the KDE heritage is apparent, though that's mainly from the theme used by the GUI and the icon set. There's no K menu to launch applications, and no Window List. Instead, the main action happens when you pull down the top bar (which is similar to how you access notifications in Android). Here you'll find both a list of open apps, and a searchable grid of installed applications.

Bizarrely for a product of the KDE community, this is quite similar to Gnome 3's Dash. The pull-down to expose more complex features works well, and we feel it offers the best window management of any of the mobile OSes we've used.

The killer feature of Plasma Active, though, is Activities. These, like their KDE counterpart, are a little like souped-up virtual desktops where you can combine widgets and apps to make separate desktop environments that you can easily flip through. To switch Activities, you simply have to pull a tab out of the right-hand side of the screen, then select the one you want from a list. It's a movement that feels natural on a touch interface.

While Activities do have their proponents on the desktop, we've found that few KDE users find them particularly useful. Many people are just confused by what they do, and we're yet to see a distribution that comes with a useful set of Activities running by default. Could the mobile arena be where they shine?

Uniquely among the mobile operating systems we've tested, Plasma Active has distributions much like desktop and server versions do. Getting the interface just right will be a task for these distributions, but it remains to be seen if any of these will put in the effort to make a good mobile product.

Like with Ubuntu Touch's Scopes, getting the right activities available by default will be crucial to Plasma Active's success. It's unlikely that many users will spend time configuring unless they can see that they'll be useful. Also like Ubuntu, this is an area that KDE haven't done well in on the desktop.

Plasma Active feels much closer to a desktop OS than any of the others we've looked at here. Perhaps it could be considered a closer competitor to Microsoft's Surface Pro than Android and iOS powered tablets? It's the one that we feel has the best chance of bridging the gap between content consumption devices and content creation devices. Being able to switch activities depending on what you are currently doing would be a real boon in this area. For example, having one activity setup for checking email and surfing the web on the go, and another for desktop-like activities with a keyboard plugged in.

Perhaps, because of the more feature-filled system, Plasma Active was the only system we tested that made the Nexus 7 feel too small. The extra screen space on a nine or ten inch tablet would really help make the most of the environment, as would a good case and keyboard.

Sailfish

Rising out of the ashes of MeeGo.

Jolla, a Finnish company founded by former Nokia employees, is building a new mobile phone operating system using Qt technologies on top of the Mer platform. However, that is where the similarities with the Plasma Active system end. Sailfish (Jolla's system) is definitely a mobile phone OS with none of the more complex features of Plasma Active.

In a similar way to Ubuntu Touch, the main selling point is the user interface, and the designers and engineers at Jolla have been rethinking how we interact with our devices - like Canonical. Jolla has come up with a far richer set of controls than both Android and iOS use.

Perhaps, though, the two biggest areas of interest are the screen that shows the currently running apps, and the ability to build themes based on pictures. The first roughly takes the place of the Home screen in most other mobile OSes, but acts more like a window switcher with a little more power. The second allows you to change the interface based on the colours from the background picture. This theme permeates the apps running as well as the Home screen and app pickers.

Jolla are betting on the fact that people want to customise the colour schemes of their phones, and the first devices will come with interchangeable cases to let users pick and choose the colour. We have been unable to get our hands on one of the phones yet, but the videos online and the Sailfish VM that's part of the SDK show a slick Qt interface that we're sure will be a delight to use. Devices are expected to start shipping around the end of 2013.

Tizen

From the deep pockets of Samsung et al.

Most of the systems here are created by either small companies (such as Jolla's Sailfish), or organisations that are new to the mobile space (such as Mozilla's Firefox OS or Canonical's Ubuntu Touch). Tizen stands out as a glaring anomaly in that it's being built by such behemoths as Samsung and Intel who have come together to form the Tizen Association that's overseen by the Linux Foundation. If money and market clout are all it takes to break into the mobile OS market, then Tizen is sure of success.

However it's not, and for proof of this, you need look no further than the difficulties the combined powers of Microsoft and Nokia are having pushing Windows Phone on an uninterested market. The Tizen interface is clean and easy to use, but frankly, it's hard to see what it has to recommend it over existing products, unless perhaps if it performs well at the low-price end of the market.

Our cynical side is tempted to conjecture that the main purpose of Tizen is to help Samsung gain a better negotiating position over Google for Android. Our less cynical side is excited by the possibilities for Tizen as a powerful embedded solution.

The platform overview states "Tizen aims to introduce a robust unified experience across multiple devices - smartphone, tablet, Smart TV, PC and in-vehicle infotainment - with a focus on mobile platforms." We can't help but suspect that the success of the platform, if not the focus, will lie in things like TVs and in-vehicle entertainment systems. This is an area that, unlike mobile phones, is under served and an open source platform could really benefit both manufacturers and consumers.

Desktop Linux

A handy alternative.

So far we've been looking at OSes built specifically for mobile devices, but actually there's no reason you can't use a regular version of Linux on a tablet (a mobile phone screen is a bit too small for a normal desktop). Enlightenment is one popular choice that we've been using. Aside from a few niggles, everything works well.

All the familiar applications work, and things go in familiar places. However, it quickly becomes apparent that there's a reason all the major players have developed systems for touchscreens. Hierarchical menus, which are great for mouse-driven user interfaces, feel unnatural and fiddly when used with a touch mechanic. There's the problem that you can't see through your finger to what the pointer is currently on.

Things that do work well on touch, but don't work with a mouse (such as pinch to zoom) aren't generally in these interfaces. Scroll bars are another widget that doesn't translate between touch and mice. Using a regular desktop environment on a tablet is a nice idea that works in practice, just not as well as the alternatives.

While Ubuntu has shown that similar user interfaces can work in various form factors, we remain unconvinced that a single environment can work well on both desktop and portable form factors. Today's hi-def tablet screens, though, do hold enough detail for desktops to work at this size. Though we can't see ourselves switching over to a desktop for our main tablet interface, it could be a useful second interface to use when an external keyboard and mouse are available. Although like Plasma Active, this is probably best left to larger tablets.

The challenge

How mobile Linux OSes can win.

In researching this article we installed four of the six main mobile Linux operating system on a 7-inch tablet. A further one (Firefox OS) is already available on phones commercially. The only one we haven't been able to try on a real device is Sailfish, and plans are well advanced for this to be released soon (possibly even by the time you read this). Together, they represent the work of at least half a dozen companies and plenty of dedicated volunteers.

We were most impressed by the diversity on show. With the exception of CyanogenMod, these aren't simple clones of existing solutions, but real innovations. They offer users a real choice about how they want their mobile devices to work, and allow you to optimise the user experience for your particular uses.

Of course, over a decade of slow growth in desktop Linux usage has shown us that this alone isn't enough to ensure success. The current major players in the market - Google and Apple - are two of the biggest companies on the planet, and neither of them will welcome a new participant, especially if the new participant starts to do well.

In some ways, the mobile marketplace now is a lot like the desktop market place in the early 2000s, at least from an open source operating system perspective. For mobile Linux OSes to succeed, they need to learn from the mistakes of desktop Linux. They need to be available in shops, they need to be backed up by advertising budgets, they need to be seen as cool by the average man in the street, not just by geeks. In short, they don't just need to be better than the existing players, they need to be seen to be better by a large proportion of the market.

It's a tough challenge, but there are some promising signs. Industry rumours are that hardware manufacturers are keen to support nascent OSes to break the duopoly of Apple and Android. Carriers are keen to offer new choice to make them stand out. Couple these with the fact that the average smartphone owner upgrades their device every two years and you have a fast-paced industry that's ripe for change.

As Blackberry can attest, it's a marketplace that can change fast and change hard, but as Microsoft and Nokia have learned, it's also a tough market to break into.

• Now why not read Can Firefox OS be the new Android?