They said it at LinuxWorld in Toronto a few months ago. They’ve buzzed it at analysts, and now the press is saying it to the public. Novell says this is the year of the Linux desktop, and I’m familiar with evidence showing gains in popularity for Linux. Yet, I disagree that this is the year. Nothing is happening this year to make it, specifically, the year of the Linux desktop and I’m going to hypothesize what could change that.
To me, there’s no contest, GNU/Linux systems have been offering more innovative, stable, easily productive, and pleasant desktop systems (KDE for example) for years. However, that’s not enough to move Linux to a place where it challenges the automatic momentum both Microsoft and Apple enjoy within the mindset of the general population (at least in North America–perhaps elsewhere this is different). The mindset of the user/customer environment is what is needed to turn it into the year of the Linux desktop–Novell isn’t making much of a dent in this regard.
Jem Matzen wrote why specialized systems as opposed to fancier eye candy would be a better answer to move in this direction (that’s my very over-simplified paraphrase). I appreciate that notion in part; I’d like to suggest something else though, something which I think would give GNU/Linux and FOSS applications a real poignant way to shift the public’s mindset toward their adoption. Even better, it’s a business model that could only, really work in its entirety within a Free and open source ecosystem. What I’m suggesting, is essentially like something James P. Womack and Daniel T. Jones recommend in their book, Lean Thinking, except applied within a FOSS ecosystem.
To catalyze the required mindset shift–and this may appear plain at first glance, let me flesh it out–if a customer could easily buy a computer system, stacked with the desired hardware, configured software, support expertise, update service, backup service, in addition to having automatic access to a range of web services (like music stores or VoIP services) optionally pre-setup, it would be a completely compelling solution. What’s so special? Don’t we see that from the likes of Apple or Dell? Not really. No company that I’m aware of actually does this to the degree I’m proposing but a GNU/Linux OS distribution is the one that would fit this model and allow it to work, now. I’ll continue by talking about what such a fictitious GNU/Linux solution provider would do and I’m going to refer to this fictitious company as Fictux.
A full computing solution should come from a company that pre-bundles everything its customers want, consistently supporting it, for the duration of ownership. It should not require anxious intervention from the owner when the owner desires a new component or new system, and the new system should have all data and applications from the old system installed, setup, and accessible upon delivery.
1) Getting the computer. It’s not impossible to find a company on-line that will sell a computer set up with Linux. There are some hardware vendors offering compelling Ubuntu and Linspire preinstalled systems. Every now and then you even hear about a big box store selling some Linux PCs. Some companies, like Dell, even let you pre-configure the hardware components to varying degrees. Fictux would make this selection easy, it would have pre-tested the hardware to be sure it all works together in combination with the applicable software. This is not a new idea but it must be combined perfectly with the rest of the service.
2) The right software, configured right. The system cannot simply be preloaded with a Linux distro! From the point-of-view of most average users, there probably isn’t a cognizance of getting anything extremely compelling from an OEM with Linux preinstalled, they might as well have Windows. Worse, getting a new system with the standard OS leaves too much effort to the user to seek and install all their desired applications (this is true of Windows, Macintosh, and Linux). Most standard Linux distributions get a running start (bundling thousands of apps) compared to Windows or Mac systems, but sometimes too many apps are a detriment. Worse is when the user gets apps targetting what s/he wants but they’re not necessarily the specific ones s/he wanted (say I want Kopete while my distro automatically gives me GAIM).
A long time ago, when I was a dedicated Mandrake (Mandriva) user, I remember suggesting (and I don’t recall if this was in a user forum, an e-mail, a comment form, or what) that they let users select every software package they want, in advance to downloading an installation ISO. Then the user could download a totally custom version of the distribution. That’s to say that Fictux would offer custom versions of its distribution, tailored to exactly what the user wants the instant the system is turned on. This must be done at the time of purchasing the hardware.
Could Microsoft or Apple get agreements, permanently ongoing agreements, from the thousands of potential proprietary software vendors a customer might want to have installed? Could Microsoft or Apple charge a humane price for such a system? It doesn’t seem plausible. However, a Linux-based manufacturer can do this because of its FOSS ecosystem.
If I was the customer, obviously over the computer’s lifetime I’d want to occasionally install something new, but currently when I, for example, install a Kubuntu system for the first time, I have to search through a package repository interface (though it’s an easily unified one) for whatever I want to install, then tell it to install–the consequence is that every time I set up a new computer with the operating system, I spend half a day just adding the applications I want and configuring them. Yet a Linux distribution is already a carefully selected collection of Free software applications, tied and tested together into a whole system. Why is practically every distribution offering its common system (sometimes there is a server or business version) and then asking the user to install all the options? Fictux would ask the options first and make the distribution, the user’s distribution. It could be an audio work-oriented distro, desktop publishing distro, file server distro, immediately upon powering on, and according to the user’s taste. Furthermore, and I’ll expand this when I get to backups, it should already be populated with the information about the user, his/her preferences, and files.
3) Provide the support expertise. Plenty of companies, especially in the open source world, have chosen a business model of providing support services. Why is this often an independent company from the hardware, software, or other services? Of course they’re not all independent companies, but Fictux, in providing each point I’m detailing here would also be the point of contact for any support-related issue. Software questions, hardware failures (even to the point of arranging pickup and delivery replacement service), possibly even in agreement with the ISP.
4) Manage the update service. If there is some sort of hardware recall, Fictux would be responsible. As new technology is available, Fictux stays on top of it and folds the new tech into its service. It’s got to preemptively know which hardware will best support new software and be able to let the user know, without requiring the user to research all kinds of options and configurations. I think the transparency of the many test releases in open source development might be especially helpful in this regard. As fixes for software bugs, security holes, and new versions become available, the company must manage these and make them simple for the user to be aware of and apply. This is essentially a no-brainer for Linux distributions, most of them already do this on the software side, it’s a matter of making this process as effortless on the hardware side. For example, current excitement is the Novell sponsored xgl/compiz combo. It requires certain graphics hardware. Fictux would offer this alongside its software update service so that the user immediately and easily understood what would be needed to get the latest fun features. Linux systems generally are able to support the hardware I throw at them (often more easily than Windows), though some exceptions stand out–as Linux systems gain in popularity, I expect this issue will continue to decrease.
5) Make the backup service easy and more useful than just a data backup. A number of different Internet-based backup services have been sprouting up, both for business and the regular home user, but these don’t interconnect as an integral part of the rest of the products and services I’ve mentioned for Fictux. Backing up data should come easily and automatically. It should be secure and accessible. But let it do more than just back-up data. It could be used for preconfiguring a system. Save all the configuration data throughout users’ computers’ lifetimes, even as new applications are installed. When it’s time to buy a new system, the customer won’t have to reselect all of his/her applications (like the first time) because it would already be known to Fictux. Even better, the computer system that the user receives would include all of his/her data, settings, bookmarks, etc. Many of these could even be imported from non-Linux systems at the first order. This would be like a dynamic “ghosting” system for companies that continually have to order new computers for employees. I’m sure there are vendors that already deliver similar services for large organizations but again, I’m not aware of a company that does it in conjunction with all of the rest of the items I’ve detailed and by scaling from one to hundreds or thousands of units.
6) Pre-setup web services. Deals used to come bundled by some manufacturers, months of AOL at a discount, just click the icon to activate it. Instead, allow the user to select the web services they use or would like to use (say VoIP services, on-line music stores, and even free services such as favourite Internet radio stations) in advance to receiving the computer, it would just be another configuration the company could easily arrange for its customers before the customers even start using their computers and more importantly it would allow Fictux to include the appropriate hardware to support these services (audio file player? headset?, etc.). It may be argued that these services are too vast to manage, but I think Fictux could find a way to bundle a service distribution in much the same manner it bundles the thousands of Free software applications in its repository.
Finally, as I said at the beginning, none of these ideas are necessarily new in-and-of themselves, they just haven’t all been offered together by one company. If each can be done by some company, why can’t they all be done by a single company? It should appeal from a business perspective because each provision of service or product helps the company further its sales effort within its own solution chain. The more important point, however is the customer/user. Each step of buying a computer, using it, managing to obtain and use software, hardware, and services, and finally, after a few years, buying a new one, is accompanied by anxiety, research efforts, and ultimately wasted time by the customer/user. A company should eliminate all of that extra effort. Most users only undertake these efforts because they have no choice (read, these steps themselves provide no value for the customer/user). As I mentioned in my second point, only a FOSS vendor can adequately offer such a solution. Furthermore I think a FOSS vendor would be especially suited to do the other steps well (such as the web services/hardware pre-configuration integration) because of its existing expertise in packaging complex and diverse software configurations.
A single vendor that can accomplish all of these steps would be offering something incredibly appealing for the masses (neophytes and computer experts alike) because it would be offering the only solution that is valuable from the start, with a minimum of wasted customer/user effort. I think this kind of solution would differentiate a company enough to challenge the automatic momentum Microsoft and Apple enjoy within the mindset of the general population. When it arrives, it might even shift the gradual gain in Linux adoption to a more pronounced, year of the Linux desktop.