The Linux Desktop - The View from LinuxCon


I have just finished up three days at LinuxCon in Portland, put on by the Linux Foundation. As you might expect from such an event, there were discussions on a wide range of topics, some to get you thinking, some to excite you and some to challenge your notions. The levels of presentation varied by presenter and, overall, I would say it was a great success.

One topic that was discussed quite a bit was the Linux Desktop - the state of the desktop, the future of the desktop and the direction the desktop is taking. It was so popular a topic that it was mentioned in no less than three keynotes, and not always positively.

Bob Sutor from IBM took this tack about the desktop (and I am paraphrasing, along with some of my own thoughts). One option is that It goes away. We stop using desktops so who cares. This is an interesting tack to take, especially given the large advances of small, smart devices. More and more of us have smart phones that access data in ways that are hardly traditional desktop models. The UI is as alien to the model of the traditional desktop as windowing systems were to VT-220 terminal operators. As this trend continues, even the idea of what a desktop is along with what an application is will continue to evolve, and the idea of a Linux desktop as it is today may cease to be relevant. Personally, I do not see this happening any time soon, but my crystal ball runs Windows and I have not been able to upgrade it (anyone want a good used crystal ball cheap?).

Bob then hypothesized that the Linux desktop becomes a tactic instead of a strategy. So rather than watching it disappear, a single desktop distro might evolve to capture a dominant, percentage among those using Linux as their desktop, or hey, even crazier, a dominant share of all desktops. Given that Windows currently holds such a dominant place in the market, perhaps it is enough to reach parity and have an equal share between Windows, Mac, and Linux. The Linux desktop is good but as we discussed this week, it is not the great it needs to be to gain market share. In fact, current numbers indicate that Apple is kicking our butt, with an increase in market share of almost 10 percent while the Linux desktop has actually lost ground.

There are a number of reasons for this. One of the most interesting was put forth by Joe "Zonker" Brockmeier. Joe's words echo my own thinking on this subject, which is this: If the Linux desktop is going to get any traction in the end-user community, not just among us geeks, but the moms and pops, uncles, cousins and grandparents, the industry needs to stop hiding its availability on their sites like a trade secret. He further went on to suggest that the current model of downloading distributions as ISOs, burning them to disk or USB, installing them and configuring them is too hard and unrealistic for most end users, and I agree with him. Discard for the moment the feeling among some Linux people that Linux is for smart people, how many people do you do technical support for on Windows where they have tried to install the operating system from media? If people are not installing Windows from media, why should we be expecting them to install Linux from media, even if it is a relatively straightforward process?

He then went on to ask when was the last time you walked into a store and found a Linux-based system available for use. Most in the audience were left scratching their heads. It was pointed out that there are usually one or two netbooks in the big box stores configured with Ubuntu. This is a start, but when you consider that 2 of maybe 16 laptops and maybe one of a half-a-dozen desktop systems is available with Linux already installed, is becomes a real challenge to say the Linux desktop has arrived. Finally an insightful question was asked from the floor. How many of us know of a computer repair shop that sells preconfigured Linux systems? A valid question, but here is a follow up. How many people buy their machines from somewhere other than the usual suspects? This is not to discredit the question. It is valid, and certainly another way to get Linux into the mainstream, but is not the level of critical mass we need.

But things are not all bad. Ross Chevalier, CTO at Novell, brought some numbers to the table that indicate there are some bright spots ahead, especially as corporate revenues continue to decline and companies are looking at every avenue to slash costs. In a survey by IDG, 63% of the respondents indicated that they were looking at increasing their Linux desktop use in 2009. What was not discussed is whether this is as a replacement for existing Windows machines or in addition to new Windows implementation. I will be following up with Ross to see if he knows, but it indicates that the Linux desktop is no longer the pariah of the IT shop, and it is beginning to make in-roads as a realistic mainstream desktop system, and in the areas of manufacturing, retail and financial services because it leads to lower support costs. This is important - the Linux desktop is being adopted in industries outside of traditional IT.

Clearly IT, especially those in the desktop space, is looking at, if not an inflection point, then certainly the beginning of sea change. There is no question that the definition of a desktop is undergoing a change as more and more companies rethink the necessity to put a machine under each desk. Data breaches, maintenance costs, changing device preferences and increased connection speeds are making us consider the solutions we are asked to present; as maintenance and installation costs become the long pole in the budget tent, anything that we can do to shorten that pole is good. And a Linux desktop, whatever that means, is certainly going to play a big role providing those solutions.


David Lane, KG4GIY is a member of Linux Journal's Editorial Advisory Panel and the Control Op for Linux Journal's Virtual Ham Shack


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

install apps: how and where

Stefano Italy's picture

well, to me:
First: package managers must go in one direction and standardize themself. Let's select which one is good OR make ONE REALLY good and stop use all the others ( hey , suse , we already have yum and apt-get , did you really need zypper? f::k ! ).

Second idea: if i want to install a browser in /myapps/internet/whatever i want to do it. AND I do NOT want to get the source of the program and compile it for doing that ( and fighting with all the missing libraries if I do in this way..)


Anonymous's picture

I am in first year CompSci at UWO, and I can attest to the offerings from Microsoft. Anyone in my program can download and install (for free) XP, Vista, 7, 3 versions of Visual Studio, One Note, Office/Groove/Project/One Note and a few others. A lot of the people in my program think it's amazing that they can get all of this software for free and they eat it up like candy. That means that a large number of them will probably start developing for Windows. Unfortunately for Linux, there's nobody around to push the free software except other students.

Dump GNOME and PulseAudio

madman420's picture

The major distributions are all GNOME happy. This is not good. PulseAudio is a horrible, ugly, unnecessary beast. The only desktop environment that I have seen that is both stable and aesthetically pleasing is KDE 3.5.10. I think this environment would appeal to the people that hung on to Windows XP (the ones who avoided Vista). However, most distro's have switched to installing KDE 4.x as a second choice to the bloated GNOME.

I definitely agree with some standardization, like ALSA for sound (not many people truly need a "sound server" and it causes many problems). I have no idea why a "community" of GNU/Linux users & developers would put out some of these very poor excuses for a distro.

Another thing that would seem to help is "visualization". Since most of the applications for GNU/Linux/X are installed via package managers by "average" users, these package managers need to have complete and brief descriptions, along with *system requirements*, a screenshot or three, the purpose of the app (who would want to use it), etc. This is a new Linux user's window to the open-source world. Refining this area as well as installation and configuration / customization is crucial to getting the noobs' feet wet. And yes, that means a nice, clean, stable X Windowing version - not just a command-line! This is also the other side of the blade. We don't want to sacrifice any of the power user's features, NO MATTER WHAT.

Anyway, with Windows 7 just around the corner, the masses will probably look at GNU/Linux and wonder just where is the equivalent offering from this side of the fence? All it would take is *one* good distro to finally get it right(Debian, SuSe, ...?)

Linux Desktop evolution

SAMBO_Guy's picture

Well, I have been working in Linux since 1998, fought the cultural wars, tried all the latest and greatest the Redmond has to offer and still, I prefer my Ubuntu 9.10 desktop to anything coming out of Redmond, sure I also own a Mac Mini, a great machine, but my primary desktop is Gnome on Ubuntu. Sure my RHCE number is in the 300's, but time marches on. My server needs at home are quite modest at home and Ubuntu does a great job.

People say that Linux is an OS for geeks, truthfully I find it much simpler to create and maintain than an OS made to be convoluted. I remember the version of Windows Media Center I used in Vista Ultimate 64 Bit and under W7 RC2, both shared similar defects. Meanwhile on the Linux side I created a 5 line script to use vlc, and it is flawless.

My desktop machine is an AMD X2 and runs the "bloated" Gnome windows manager beautifully, no complaints, just stability. I guess the cultural wars will always exist, people will always be reluctant to change, but I have worked with Linux on both the server side and as a truly wonderful, stable desktop. Some folks just don't know what they are missing.

Ok, so if I had to choose a desktop, I would choose Ubuntu, since Hardy Heron, it has been a pleasure to use, before that, I was just another Fedora guy, happy and stable.

Future may be different??

Anand Vaidya's picture

I am more hopeful of a better future for Linux and one that is free of the suffocating stranglehold that MS Windows has on the desktop for the following reasons:

- nowadays, more people do things from mobile devices - atleast the trend seems to be iPhone and Linux based mobiles (Android Maemo, Moblin, WebOS etc)

- Virtualization is helping linux rather than windows (see TheRegister article on that)

- Linux desktop is getting exciting once again (KDE4, upcoming GNOME3) with the community unafraid of trying new paradigms

- More and more apps are available in hosted versions, accessible via a web browser, SaaS actually robs windows of lock-in advantages.

- Better (OSS) tools such as jQuery(+plugins), GWT, Pyjamas, OGL in a browser etc means it is possible to provide desktop like experience on the web...

- Countries (esp in EU) are worried about one American corp dominating the IT and that could mean serious problems should there be a conflict with USofA

Linux desktop

Anonymous's picture

Linux needs to standardize to survive.
Think about it, if you want to make a commercial program available for Linux which distro would you choose, most likely ubuntu since it has most of the market share.
Also most users only need a hand full of programs like Autocad, Garageband, Photoshop, Streets and trips ect. None of them work in Linux so your only solution is to duel boot into windows. Why not just stay there and save yourself some trouble.
Linux desktop has gone a long,long way, but needs to go further before you can turn windows off once and for all.

Linux needs more exposure

apexwm's picture

I think this article hits it right on the head, Linux definitely suffers from exposure. People aren't going to try out Linux unless they have a good reason to do so. A lot of people are going to go with what they are already familiar with, which is usually Windows. Others will hopefully realize there are better options to Windows, and will try Linux. I tried Linux and for the first time I was sold. When I ran Windows and Linux side by side, I found myself troubleshooting the Windows machines while the Linux machines ran untouched and everything just worked, all of the time. By using Linux, I soon found out how many advantages I had over Windows. But I'm in IT work, so I need something powerful and flexible and low maintenance. But, I do have to admit that within the past 4 years or so, Linux has made huge leaps forward in usability for the average user. It is actually simple now to install Linux with it's nice graphical installer. I use Fedora since it is widely supported, and I couldn't be happier. The nice thing is, you can use an older PC, install Fedora (or even use the Live CD and NOT install it but simply run it), and see how you like it, and the only thing you are missing is some of your time.


Zyyx's picture

Ive been using linux as my main desktop for the better part of 13 years and while its come a very long ways (ie: my mom uses it) it has also begun to adopt some of the stupid concepts from other desktops (windows/osx). This though is really not whats holding it back. Im a Unix Admin and there are problems using it as a primary desktop and they are all of the variety "incompatibility". Dell DRAC interfaces often have problems, clients using proprietary office formats, VMWare(and even Redhats KVM) tools windows only, other web interfaces requiring IE to everything from fabric switches to tape drives and so on. While this is not the fault of Linux IMHO nonetheless its a HUGE problem for IT and on a smaller scale end users. With regard to the average home user Linux is ready its just not "available" to them. For the gamer home user linux is all but not an option which is nothing more than poor political bs within the gaming industry.

Linux become end user desktop OS if strategy will be...

sackana's picture

Why today every where desktop with non-Linux OS?

Reason is that today's popular and top most (proprietary) OS have been targeted to a developers those who will be working for end user apps.

Generally end user will accept a system setup that developer will provide to them and now a day from small application providers to a big providers, provide only what they know or a setup in which they are hands-on.

And I have observed that most of the application providers are hands-on on a popular proprietary OS and that's the reason why all end user desktop having proprietary OS, not Linux.

So Linux need to target end user apps developers/providers and make sure that availability of fast development IDE (like NetBean, Qt Creator) and SDK that easy in development.

So Linux should be worked for a same strategy (to make easy development IDE & SDK ).

That will easy to develop end user apps and developers will make user friendly apps on Linux platform. And one day Linux become end user desktop OS.

Close but slightly off the target

Anonymous's picture

"And I have observed that most of the application providers are hands-on on a popular proprietary OS and that's the reason why all end user desktop having proprietary OS, not Linux."

Not precisely true. Many ISVs are excited that they do not have to remain stuck in backward compatibility - which is generally required if you want to develop for Windows. Linux is quite attractive to them.

Business side of things are of course different story. I yet to hear about free trainings and support provided by Linux companies.

"So Linux need to target end user apps developers/providers and make sure that availability of fast development IDE (like NetBean, Qt Creator) and SDK that easy in development."

SDKs will not solve all problems present today in Desktop Linux landscape.

I can only hope that Debian (or Ubuntu) would come out with something similar to e.g. FreeBSD style packaging. Biggest difference is that Linux flooded with technologies and APIs of all kind. Level of redundancy of APIs is astounding and often confusing. That's the freedom of choice's price. OSs like FreeBSD (or Mac OS X) provide developers with official blessed building blocks for their applications. And that what Linux IMO needs to make any dent in the desktop.

Desktop doesn't need the diversity of APIs. (After all end users do not care about that - as long as their applications do what they want them to do.) Desktop developers need to have a stable core GNU/Linux OS built of parts on which they can rely in development. Right now it's pretty much impossible to make a desktop application which would run smoothly on at least 95% of Linux desktops. Simply impossible.

Take a look at the chaos of sound and video acc. subsystems to understand the problem better. Some distros require app to use ALSA. Other target OSS (yes, there are such distros). Other require PulseAudio or Jack. Video acceleration in past three years was rewritten at least twice, meaning that multimedia applications have to support all of them to run efficiently.

Linux, if it ever to make to end user desktop, needs some sort of core, minimal desktop platform which would encompass essential to desktop components. Desktop centric distros would have to be built on top of the platform so that developers would be assured that their applications would work as intended.

Or direct response to your comment: SDK is meaningless if it is not widely supported. And you can't make a universal SDK given current state of diversity in desktop-centric distros. (Note that )

Desktop Linux should provide end-users with diversity of applications, while at the moment their provide developer with diversity of literally unportable libraries/APIs. As long as distro don't resolve latter, we wouldn't see the former.

Implementing the Linux desktop

Ram Sambamurthy's picture

As an ex-CIO of 1500-man IT company, I can tell you that unless CIOs and CTOs are involved in the decision to migrate to the Linux desktop, there won't be advances happening on this front.

Don't expect department managers or other levels to make this happen. It was easier moving servers to Linux but not desktops. Employees just want to get their jobs done and go back home. They don't care about viruses, maintenance, etc.; as far as they are concerned that's "IT's job".

Therefore, it is imperative that organizations like the Linux Foundation engage CIO/CTO/CEOs directly over this matter and help them pave the way for this to happen.

Otherwise, we can continue having this discussion over LinuxCons until the cows come home...

Quote: Therefore, it is

Anonymous's picture


Therefore, it is imperative that organizations like the Linux Foundation engage CIO/CTO/CEOs directly over this matter and help them pave the way for this to happen.

- - -

In the meantime, for us Linux users it is quite wonderful to see the Linux foundation engaging itself to various FUD campaigns coming directly from Mr. Zemlin himself.

I have little faith in such corporate playgrounds as the Linux foundation. If anything, these do more harm than good.

You are correct

David Lane's picture

Ram has hit the nail on the head. Of course, I discussed the issues of management involvement at the corporate level several weeks ago for those that missed it.

David Lane, KG4GIY is a member of Linux Journal's Editorial Advisory Panel and the Control Op for Linux Journal's Virtual Ham Shack

Geek Guide
The DevOps Toolbox

Tools and Technologies for Scale and Reliability
by Linux Journal Editor Bill Childers

Get your free copy today

Sponsored by IBM

Upcoming Webinar
8 Signs You're Beyond Cron

Scheduling Crontabs With an Enterprise Scheduler
11am CDT, April 29th
Moderated by Linux Journal Contributor Mike Diehl

Sign up now

Sponsored by Skybot