Does Desktop Linux need a Killer App?

31 Jul

OK, to start off, lets state something that everybody already knows: GNU/Linux on the Desktop hasn’t skyrocketed. "The Year of the Linux Desktop" has become a comical saying among the knowledgeable. On the other hand, in other areas such as the server market and the embedded market, it’s been one of the (if not the) most popular operating system, and more recently in the smart-phone market, it’s continued to grow in mind- and market-share despite several beatings with very big sticks from the competition. What’s the difference?

First: The Server

The first area in which Linux started to really become popular was in the Server market. In a really small nutshell, as the Internet started becoming more popular through the 80s and 90s, so did Linux adoption by ISPs and website managers. This is easily most attributable to the creation of the free-libre Apache web-server, which was one of the most capable servers of the time and ran best on top of GNU/Linux. It was essentially a no-brainer: you could buy custom server hardware, put it all together and link all your machines, then download/install a full GNU/Linux/Apache system and have a server running pretty painlessly, as well as at a considerably lower cost compared to buying or licensing a proprietary system with a proprietary server that was, by comparison with Apache, lacking in features. Apache was the application people wanted, and it was free; GNU and Linux simply served to sit under it and run it painlessly.

Then: The Embedded Markets

The embedded markets, such as routers with embedded firewalls, sat-navs and various other devices, is a trickier situation to analyse. Certainly, there doesn’t seem to be a, "killer app", rather than a, "killer feature": that you can put chunks into the system or, more appropriately, pull great big chunks out. GNU/Linux’s flexibility and overall leanness allowed it to technically fit into embedded systems much nicer than many proprietary systems that aren’t as modular. For example, with the router market, using a much more lightweight server than full-fat Apache (or a very refined Apache with very few loaded modules) for a browser-based configuration tool isn’t just technically possible, it’s bordering trivial. Similarly, getting firewall software in there was easy after you’ve removed, say, file management and filesystem management software as well as many other unused GNU userland utilities was hardly a squeeze. This allowed them to use cheaper hardware for the same task, sometimes even with more features. You might consider, in the case of the router market, that again, tiny server software and firewall software as the killer apps, but I don’t think the Linux kernel and GNU userland tools would have appealed as much had they not been as modular or as lean, factors that are very significant in the restrictions of embedded environments.

Later: The Phone Market

Google’s big push behind Android is the first thing that comes to mind when I bring up Linux on phones, and for good reason: it’s undeniably one of the most successful smart-phone operating systems today, especially considering Microsoft has been in this sector for a decade (counting Pocket PC 2000 as the, "first" version of Windows Mobile), and now we also have Nokia putting their bets on a full GNU/Linux system for their high-end smart-phone, too. Immediately, one can notice similarities between the smart-phone and the embedded markets: both need lean and streamlined software that can make the best use of the limited (though admittedly ever-growing) resources available to the smart-phone, for example, and you could also argue that input on smart-phone isn’t dissimilar to that of sat-navs. However, smart-phone have a huge differentiating factor: when people use a phone, they’re going to be consciously aware of the software they’re using, even if they don’t know what it’s called, how it’s made or who makes it: where sat-navs and routers are designed to be interacted with for a very limited time, then almost entirely ignored, smart-phones instead demand the full attention of the user for the duration of their use. While they can be left in the user’s pocket and ignored while not in use, users use these devices to browse the web, check their E-mail or otherwise interact with their, "social network" (in a more literal sense than dedicated websites, mind), and while they’re doing these activities the phone is the centre of their attention. These, then, require software that’s well-designed – on an operating-system level, and by that I mean the operating system has to get the user from a neutral/unused state to what the user wants to do with as little hassle and as few distractions as possible. In a sense, the software has to get out of the user’s way: if the user notices the rest of the operating system (e.g. it’s slow, something crashes, something is difficult to use or figure out, it’s somehow restricting them from getting to the tool they need to use etc.), it’s not very well designed. Here, then, the manufacturers are perhaps much more interested in the user-facing side of the phone than the underpinnings. Android, for example, does away with the GNU userland utilities and instead, has its own stack on top of the Linux kernel. This, to me, makes sense: while Unix was designed to be multi-user from the start, phones are very single-user, personal devices. You will rarely be creating new users for a phone. This means an awful lot of the tools of a typical Unix set-up may be redundant or unnecessary on a mobile phone: so now, we have Linux managing the hardware with a completely alien, specialised stack running on top.

This helps to demonstrate what the phone needed to be successful: not a single specialised application, but an entire experience. For this category of device, I think, there was no, "killer app": the experience of using a phone like this had to be good, or there was no sale. It needed to be comprehensive: it needed to be able to make phone calls reliably and in a way that is familiar or isn’t confusing; needed to be able to read, write, send and manage E-mails, SMS messages and multimedia messages, and; it needed to be able to browse the web, view and take pictures and play music and video. The more comprehensive the experience, the better: for example, Android has built-in Facebook and Twitter integration, along with picasa and flickr for images, as well as online publishing built-in thanks to Google’s technology (Picasa/Youtube). More importantly, all these functions had to be done in such a way that wasn’t overwhelming, confusing or strange to the user: it had to use familiar paradigms, from a QWERTY keyboard to multi-touch pinch-to-zoom; from the familiar back/pause/play/stop/next layout (or similar) in the music player from the single-tap-to-open interface in the menus. Android, then, doesn’t sell a single big-hit application: it sells an entire experience, and therein lies its success.

Now: The Desktop Market

So, what about the desktop? Does there need to be a killer app for the desktop, as there was the server market, or does it need to provide an experience, as the phone market does?

Personally, I believe the phone market is closest to what’s needed for desktop success: Linux on the desktop needs to provide a comprehensive experience. As it stands now, for example in Ubuntu or many of the other various distributions such as OpenSUSE or Fedora, on the software side they already do: they open most documents, music, video and web-pages with very little hassle, with exception to, perhaps, complex spreadsheets with various macros inside (which becomes more akin to running a program than opening a document, in my opinion, which is like complaining that it doesn’t run VB.Net applications). Many users will find most of their needs accounted for.

However, there are very few people actually providing the comprehensive experience: the comprehensive experience being using a computer. The missing component is the computer. In America, the excellently presented (if slightly overpriced) System76 store provides that experience. Remember, the typical computer user doesn’t consider hardware and software as two separate identities; they consider the whole thing as a computer. The reason people use Windows isn’t necessarily because it’s easier to use or it’s compatible with Microsoft Word, it’s because they bought a computer and Windows came with it: Windows was part of the experience they were paying for. For this reason, especially for netbooks and other low-power laptops, I would suggest people recommend stores like System76, or otherwise netbooks like the Dell Mini 10n with Ubuntu and other retailers selling Ubuntu on netbooks, to people when they buy a new computer: do so with more persistence than recommending Ubuntu itself. First, for the reason I stated above: because these stores sell an entire experience, and with that comes a few guarantees (like the software will work with the hardware with (near-)zero fuss) and thus, with the best first-hand experience. Second, because the more people that buy full computers from these stores, the more incentive the stores have to make Ubuntu part of their core range of products and, in the case of System76, the cheaper they can make that experience. Third, and really the most important part, to give users that have never heard of Ubuntu a very positive first experience of Ubutnu, a good understanding of the differences between Ubuntu and Windows and perhaps to get them buying Ubuntu computers in the future. In fact, you can make the experience even better for them and buy them an Ubuntu computer yourself, as a gift. For European (or, at least, United Kingdom) Ubuntu users, it’s a more difficult situation: there isn’t a single, comprehensive store (to my knowledge) with the same high-quality presentation as System76 over here, and the Ubuntu option seems to have disappeared from Dell or HP’s website, leaving us scouring Google Shop search for viable solutions (by the way, if any of this changes, please let me know!).

By all means, promote Ubuntu as a piece of software you can install on their computer for them (remember, installation and configuration is the hardest bit and making their lives easier is mostly the point of promoting Ubuntu in the first place), but don’t forget completely about the stores selling the whole experience: they are the most likely way Ubuntu and Linux on the Desktop are going to pick up traction. I mean, real traction. The, "Year of the Linux Desktop" type traction.

Powered by Blogilo


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: