Is desktop Linux mission impossible?
December 5, 2007 
The other day I realized my kids don't know where the Mission Impossible theme came from. They know the tune – at least, my 14 year old does because she played it in orchestra. But neither she nor her younger sister has any idea there was ever a TV show by this name. The Tom Cruise movies are pretty much ancient history to them too – can you believe the first one of that trio came out in 1996?.
To measure how long ago that was, consider that desktop operating systems back then ran comfortably in 8 Mbytes of RAM. At least, OS 7 did on my Mac IIfx. I can't vouch for Windows 95. Those were the days! Nowadays I wouldn't recommend less than 2 gigs of RAM to anyone planning to run Vista or Mac OS X on a new machine – not even to my worst enemy (hmm, maybe I should tell him to go for 1.5 gigs – that will keep him busy for a while). Of course if you are using Linux you can get superlative performance with far less memory – but readers of this post already know that.
Anyway, my point is that every adult knows the story behind the Mission Impossible song: a team of elite but non-conformist superagents is sent out to perform some vital but hopeless task that no one else dares tackle. Everyone is sure they will fail. But the superteam somehow pulls all kinds of ingenious tricks from its bag of surprises, gets the job done in the nick of time and makes it back alive.
In my last post I argued that desktop Linux has slim prospects for ever becoming more than a niche alternative to Windows unless it can somehow come up with a viable business model. Free as in beer doesn't cut it. Judging from the reader responses to that post, it looks like some Linux fans are joining Tom Cruise and Peter Graves in taking on an impossible mission. They are betting that Linux can succeed simply by beating Microsoft in the marketplace of ideas, without getting real traction in the more humdrum marketplace of dollars and cents. But this is not going to happen. The software market regrettably does not belong to that fantasy world where Hollywood scriptwriters make the rules (when they're not on strike), where superspy heroes are rescued from impossible situations with a clickety-clack of the keyboard.
Many readers chided me for failing to mention what a technically superior desktop OS Linux makes compared to XP or Vista. It doesn't crash all the time. It boots faster. It uses less memory and less CPU. It doesn't require you to defrag your hard disk twice a week. Its file manager finds files in a snap, even in huge volumes. It emits less carbon. No animals were harmed in creating it (not counting the penguin). Oh, and did I mention that it doesn't crash all the time? Right. All this is well and good. I thank my readers for this useful information. Most of their points about the superiority of Linux are well taken, allowing for a little hyperbole here and there. (As for Reader X who complains that it takes him two minutes to attach a file to an e-mail in Windows, dude, you've got to stop e-mailing those hi res movies, try putting them on an FTP site. Or YouTube.)
But all this is beside the point. Defenders of the "Linux will win because it's a better OS" mantra are forgetting a fundamental fact of life as we know it: good enough is the enemy of best. It doesn't matter how superior Linux is, because Windows – XP or Vista, take your pick – is good enough for the vast majority of users.
I realize that for many Linux users this is a shocking thing to say. But it is the nub of the problem, and the key to understanding why the desktop Linux community is in some ways its own worst enemy. You see, most Linux users don’t want to think of themselves as part of the "vast majority" of PC users. On the contrary, they view themselves as members of a cutting edge elite that has seen the light and thrown off the chains of a stifling regime. Now they're running free and wild on the open range and rarely give a thought to the benighted peasants who still labor in the dark mills of the Windows desktop.
What they forget is that we loutish Windows-using serfs still need to get on with our daily tasks, and having the best desktop OS in the world won't help if the applications that run on it aren't up to snuff (for the server side of this issue, see It's the stack, stupid!). Quite a few commenters suggested that if I only tried Linux I would learn to love it, and so find the way to breaking my dependence on Microsoft Office. Well folks, I hate to tell you, but I have tried it, and the desktop Linux application stack still leaves me cold.
Right now I have sitting within reach on my desk a hip new Dell laptop with Ubuntu installed on it and a sleek and glamorous Mac Book Pro. But neither of these machines is currently booted up. I'm typing this post on my funky old ThinkPad running Windows XP. Why? Because every word I write is written using the outline function in Word. True, the outliner in Word for Mac seems pretty much equivalent to the one in WinWord (though neither hold a candle to Dave Winer's late great outliner More, which Apple wantonly abandoned when it dropped backward compatibility for legacy Motorola 68000 programs). But since I already owned a copy of Microsoft Office for XP when I got the Mac, I didn't see any reason to pay the Redmond freight twice.
As for OpenOffice on Ubuntu, the outlining functions in Writer are a long-standing insult to the needs and wants of anyone who writes structured texts for a living. This has been written about countless times and I won't rehash the subject here, but check out this writer's pro-outline rant which is typical of how a lot of us feel. You can also take a look at OpenOffice.org issue status page 3959 to see why adding decent outlining features to Writer is such a technical nightmare. (To be fair, some people who don't think outlining is so important find that Writer compares quite well to Word. But I suspect they are people who write mostly e-mail or code rather than longer texts.)
My point is simple. Like the vast majority of PC users, within certain broad limits, I don't care how good or bad the desktop OS is. All of them are good enough to do the basic tasks. None of them – and this includes Linux – comes close to the reliability or ease of use of my Honda Accord. Some are better than others, sure. But the differences are far less important than the quality and variety of applications that each OS supports. And here the perhaps unfortunate but indisputable truth is that the vast majority of PC users believe Microsoft has the best apps. OpenOffice is free as in beer and has been for years, but if tens of millions of people insist on paying serious bucks for Microsoft Office every year, that tells you something about how they rate the two suites. It's not as if people are too lazy to download and use free apps when they think they're better. Just look at Firefox. Unlike Writer, it really is better than what Microsoft offers, so millions of people use it. (I do myself - despite the fact that Firefox routinely loses its mind when I try to download too many PDF into tabs.)
At this point in the conversation diehard Linux fans will often argue that all those people using Windows and Office only do it because they're brainwashed, or because their employers force them to (maybe the employers have been brainwashed?), or because they're, well, stupid. There. I've said it. Too many people who use Linux believe that their Windows-using brethren are suffering from a deficit in the gray matter department. They just can't understand what other explanation there could be for the fact that so many people persist in making such a self-evidently wrong choice.
But in refusing to see that the Windows-using masses are neither stupid nor deluded nor coerced, the Linux elites commit a grave analytical error, one that I am sorry to say will very likely lead to the long-term marginalization of Linux as a client-side OS. Their error is to believe that Windows users have irrevocably ceded control of their PCs to Microsoft and couldn't take it back even if they wanted to. This is quite wrong. The users very well could take control back from Microsoft if they wanted to. So far, no one has offered them a better deal. That's the bottom line. The Linux aristocracy, peering down at the plodding Windows peasantry from the towering heights of its 1% market share, has forgotten that in the software game what counts the most at the end of the day is still product. Give Windows users better applications, and they will switch. OpenOffice isn't doing this.
Actually, there is another player in the software marketplace who understands perfectly well that it's still all about product and is pursuing bold new ideas quite effectively, albeit in a way that is both supremely profit-oriented and resolutely non-open. That would be Google, of course, and I predict they are going to give Microsoft a real run for their money.
Reader Comments (30)
You seem to be saying that Microsoft is "better" because it's what everybody uses and everybody would use Linux if, like Microsoft, everybody used it. And what Linux needs to do for everybody to use it is to get everybody to use it. And on, and on, and on. Now on to a more serious note. You made an error in your comparison of Open Office Writer with Firefox "Just look at Firefox. Unlike Writer, it really is better than what Microsoft offers, so millions of people use it." There are many other browsers too and people use them to with nary a problem. How could this be? Because the Web is based on standards, like HTML, but unless Open Document Format (ODF) becomes the accepted standard for office suites, Microsoft Office can enjoy total dominance in its self-made ecosystem. You can't compare the two, applications in a standards environment, applications in a vendor only environment.
Linux, or its descendant, will become the common operating, not because of its technical merit like you say, but because it's part of a much larger business method called Open Source. Ever hear of it? It's slowly but relentlessly poisoning the waters that proprietary vendors have had to themselves for far too long. And it's customer driven so don't expect to see any big ad campaigns to tip you off that it's coming.
In some form or another our words here will be around for future business history students to read. To them I just want to say, "Some of us could see it coming".
I agree completely with the notion that the OS really isn't all that important, it's the software.
My experience has been that resistance to trying to move people to Open Office has more to do with the fact that people simply don't like to learn new software, particularly somthing as mundane as Office applications. They like to do their job efficiently, and that means doing what you know.
I will say that our users are *very* receptive to Google Apps, even though it lacks functionality in spades. I also think that is good for Linux, because they only thing they rely on is the browser.
Lastly, almost all average Joe users I know are wary of Vista. So while I think that it's the applications that count, don't discount a shoddy OS. If you can't run your applications reliably, people( and businesses) will look for alternatives.
My concern is the hyperbole you complain about, you throw about. Constantly referring to Linux users as "nobility" and "elitists". This is just a different tack at trying to use emotional branding to scare users away or essentially, polarize even further the discussion groups.
Part of the issue for business desktop Linux is indeed lack of consistency in applications. Consistency in usability, in functionality and presentation in general.
Linux applications currently are at an "either/or" stage, they are either brilliant or lacking" in qualified details which are taken for granted in preexisting apps for the "mainstream".
The overall performance of Linux applications is much the same as any MS or other app. The devil is in the details. little things that people bump into semi frequently and just expect to see the tool that evolved to meet that need after years of development in MS apps. In Linux, the rush to meet or beat the "big" points in apps have overshadowed the little functions in those same apps and that is where the problems arise. The constant sniping about x small function or tool that is common in MS apps but haven't yet made it to Linux version.
Time will show for these Linux apps, as has already been seen being delivered on in the networking/server world, once the "big" stuff is done , the time can then be taken to go back and iron out the small stuff that people clamor about so often. Once that happens, as is exactly the case in your example of Firefox, once it had the details worked out, it became the more successful. the sought out one.
It's not that Linux users se themselves as "elitists" to use your phrase, but that they are patient enough to wait for things to be done right, instead of easily fooled into paying for incomplete work? ( Windows Millennium ring a bell?)
Big Bear
On the InterWebs, there are magical things called "hyperlinks". When you click them, they take you to new tubes showing you new information. If you're going to throw stones, such as the quoted passage above, you'll do your readers a great service by actually backing up your diatribe with examples. With them, you might be compelling. Without them, you give the impression that you're just a crank.
After all, if you can figure out how to research and link to an article on a television program, shouldn't you be capable of figuring out how to research and link to examples of the nefarious behavior you claim of others?
You are totally ignoring the whole licensing issue. As Linux becomes better & better (which it will), and MS continues with it's insane licensing schemes (which it will). Corporations and municipalities (and even users) will grow tired of paying for and tracking all their licenses, and worrying that they may be inviting a lawsuit just for using day to day software. Who needs that bother and expense?
Will Linux take over the world? I would like to see it, but at the very least it may force the big guys to modify their behavior to save their market share.
Assuming your a Linux user, your post just gave us a sterling example(the stupid part at least). Mission Acomplished!
Or was that not snark at the beginning of your post?
That's true.
A Linux elite^W user since 10 years.
Inertia kills initiative, and the inertia residing in the form of Windows in all its flavors is a huge impediment to ever getting people to switch, even if Linux applications were always superior to Windows apps, which they are decidedly not.
I really enjoy goofing around on my PCLinuxOS Dell 933r, which lost decent performance at WinME and can't run Vista with its 256mb RAM. However, I can't receive any business spreadsheets on it at home and work on them because they all contain critical and sometimes very sophisticated VBA macros. The OpenOffice spreadsheet product gags on them, and I have yet to find an opensource project out there to translate VBA macros in place, let me modify the sheets, save back in Excel format with the original macros restored, and get it back to the sender. Such a tool, IMHO, would be an Excel killer and the lever to get millions of businesses off the Windoze drug.
GoodEnough+HugeInertialMass is always > BestofBreed
Still rooting for the Penguinistas!
You and your stupid business market money throwing kind will all think that Micro$oft is better... Linux is built around freedom. It's not like Micro$oft, with tons of money for advertising (if Vista didn't have a million ads, it wouldn't sell as well) because Linux isn't about making money. It's about a free OS that anyone can simply download, install, and use, as opposed to pay an insanely high (200 bucks for JUST the OS!) price, install, worry about licenses, see that you have 2 computers and that the lame thing you bought only works on one of them, go buy ANOTHER copy (200 MORE bucks, for the Same OS!), install that, activate...then your motherboard dies, oh, what's that? You have to buy it AGAIN, since your new motherboard has a different serial...
Linux isn't tied up with that license user proprietary money taking nonsense. It shows that using a completely free OS still allows you to do everyday tasks quite well without requiring insanely priced software.
Oh, and it's updated almost every day...and it doesn't get viruses or spyware every 5 seconds you go to a suspicious website...and it's not a primary target for hackers (legitimate or not, like Sony's music CD rootkits) either...and you even consider Windows after that?
Sure, Office has some random stuff that works differently than other software, and it has some proprietary formats, but when OpenOffice has all the same stuff in different locations, isn't learning a new interface better than coming up with $400 for Microsoft's latest? And to that, most M$ office users I know that have Office 07 hate it because of it's new interface...so that's just doubly bad, $400 and also having to learn a new interface.
You sound just like those who believed that Linux would never make it in the server (or embedded) market 5 (or even) 10 years ago, and you're just as wrong as they were then.
- Brendan
Lots of thing you say are true, but you also make some mistakes.
I think that your analysis is far too based on a US point of view. In a lot of countries, european, south american, open source software are chosen because it became a strategical and political issue. All administrations, the military, parliaments, they all switch to opensource solutions (linux+firefox+openoffice) because they try to keep a strategical independance towards transnational firms and US intelligence offices (you can't blame a country for trying to keep a technological and informational independance, it's not necessarily the outcome of a paranoid delirium). You say people don't use OpenOffice, it might be true in the US but it's not in other countries. I happen to be a french student (nobody's perfect), and a lot of my classmates use open office, my university's library uses linux, firefox and open office. If i work in the french or european public sector, my office suite will be Openoffice, and it will be just fine ! OpenOffice is definitely on firefox's path. I have to add that I'm not some kind of geek studying computer science, I'm just the french average student studying geography at La Sorbonne University.
Your analysis on how people choose their software is absolutely true, but you use it to explain why OpenOffice will not replace the MS Office Suite for day to day use. But facts don't lie.
thanks for reading.
Desktop Linux in one fashion (computing) or the other (mobile market) will dominate over windows. XP and Vista are NOT 'good enough' for the common user, because more and more switch everyday, in IRC channels, people are amiss trying support MS users in Linux. The Windows Platform is polluted, and has been, for a long time before I stopped using it on purpose. Viri, Spam, Malware, and All the concoctions of proprietary software. Ask a vendor open source, or closed? ask google. ask yahoo. ask tivo. ask the nokia 810. ask the sscreen. Linux Will Dominate The Desktop 'Market'...
Linux desktop is much better, but it has still some way to go to be an effective alternative for windows. In my case I completely stopped using windows and only a linux user (i'm not a techy nor a newbie) and I have absolutely everything I need for my work but there are some issues in which only with some googling or forum search it can be corrected.
But linux is going in the right way. It's better to go foward with steady passes than to go foward to fast passed and in the end it's a crap.
I totally agree with you. This is highly realistic review - as linux as a kernel might be good, but the software is equally important and its the ease of use that matters. I am not saying that GNU/Linux is difficult but it still has to do a lot. I use ubuntu for my day to day work, but my wife can't stop hating the whole environment, although she used it for months. Now i am back to vista as os of choice at home (I do dual boot when required).
I have never contacted a virus (more than 15 years on MS OSes) and I have never been hacked (my previous company used Windows Server). If you are good enough, there is no problem with any OS its just freedom, ease of use and most importantly person loving the environment.
A note for sick minded linux advocates: Please accept the criticism be it +ve or -ve. It only helps. If you guys stop listening to criticism, its down hill.
Jeff-
This is quite weak, if not flame-bate. Though selectively on target, you imply that Linux desktop will never dominate.
It may not; but the kinds of work being done for example on SLED are filling in the missing pieces.
A different way of stating your position is that while Linux in general is a disruptive innovation, Desktop Linux is -- per se -- merely a sustaining innovation ... with respect to Windows.
I would suggest that as we move forward with major refreshes every 6 months, up against major Windows refreshes every 3 or 5 years, the quality lead will be unmistakeable and general folks will not be able to ignore the kinds of productivity gains created.
We are beginning to see this in laptops -- where already today (with hardware prices falling) Desktop Linux is more differentiated in cost/perfrmance and therefore more disruptive. The extreme case of this is OLPC where Windows is too big and lacking modularity to fit & run easily, but it is true for Dells & hps running Linux too.
Quality is defined by users through their requirements, but also by situations and by code.
The Desktop isn't computing's nor is it Linux's destiny -- it's just one place it can go, and it is going there effectively.
All you guys trashing the author are full of it. He's completely right--IT'S THE APPS, STUPID!!! I "hate" Microsoft's operating systems, but like it or not, the apps I need to use ONLY run on "Windows" (and no, the Mac will NOT cut it). I keep homing that Wine/Crossover will finally reach a degree of maturity that will let me run my Windows CAD software under Linux---but I ain't holding my breath.
jeff.......
what kind of old an stupid person did you become?
It seems to me you have two separate theses. One is that Linux on the desktop cannot flourish until someone figures out how to make money from it--it's not quite specified, but broadly indicated that this someone must be a Linux vendor. The other is that Linux on the desktop cannot flourish until the full constellation of applications are not only present, but at least equivalent to Windows equivalents feature for feature.
You back up the second thesis fairly well, although it's oversimplified. In fact you spend your entire post backing up that second thesis. The problem: You conclude that therefore your first thesis is correct. But you draw no actual connection between the two. You busily prove "B" and conclude that those who doubted you when you said "A" are fools.
To get specific--note that the application you point to as ready for liftoff and superior and therefore accumulating use share is Firefox--which was developed largely by typical if bright open source developers for free. Now consider that the application you point to as failing to reach potential because infected with the "elitist" and "superior" and out of touch attitude of open source ideologues--is OpenOffice, which is to this day mainly developed by in-house by Sun, using money, and as a plan for making money--near as I can see, Sun has concluded they aren't going to yank any desktop share from Microsoft (and by extension Dell etc. on the hardware side) unless they can dethrone MS Office, and OpenOffice is the tool they're using to try to do that. The point being that it's *not* developed by open source ideologue elitists, it's developed by precisely the kind of people you claim are needed for Linux to succeed. And yet it's your example of failure.
There is no connection between your "It's the apps, stupid" point, which has some force, and your "It's the money" point, which doesn't seem to. Linux in various areas has flourished largely from "pull" rather than "push"--Linux flourishes not by making money for providers, but by being a value proposition for customers. Providers come along and make moderate amounts of money once they realize the market is already forming on its own. I don't see any real reason why the desktop should be any different.
As to my point about oversimplification--Yes, applications are paramount. But so is segmentation. "The Desktop" is a complex market. Apple maintained market share through some lean years by having the best applications for certain important niches, such as graphics. Linux can make progress niche by niche through having the best applications for a given niche--although your basic good office suite is important for an awful lot of them. The "Gamer" niche, for instance, will probably be last, long after the "scientific user" niche, the "typical Joe user" niche, the "business desktop" niche, or even the "music professional" niche.
I don't understand. You say linux can make this place because of the MS office suite? (and other windows only apps).
You already can use MS Office suite on linux (and adobe photoshop for example) via Crossover/wine.
But i think you right on the fact that windows just fit the need of the average user. in fact the average user don't even know that something else exist (even Mac).
When the average user will have the choice to buy a pc with linux or a pc with windows, linux may be win some share on desktop.
For the record, linux don't want to win market share. Linux is an OS which fit for the need of some user. If linux destroy Microsoft, that will just be a completely unintentional side effect.
You have blown me clear off the map! Seems that for Linux to be THE Desktop according to you, it has to have MS Office because of its outlining feature. We are not professional writers and I can live without outlining for awhile. As with most open source projects the time will come when such a feature will be available.
In the mean time I will read email, surf the web, type letters and such with OpenOffice, play a game or two, update my system via my preferred package manager, wish that Linux will blue screen so it can remind me of my days gone by with the other OS, look for a virus have four or more terminals going so I can compile a new kernel, and yes see what the windows apps are doing in VMWare or Xen.
Yes, seems that I will have to call my Linux desktop by another name.
If Microsoft Software is so fucking excellent like you say, why won't they support open standards?
OOXML is NOT an open standard. No one else can work with it. [search the internet for proof]
ODF is an open standard yet they won't support it. Why? Because if they released the lock -in they have on document formats people would drop office like the bag of shit that it is!!!!
Linux will get desktop space. Problem is not exactly now. Linux Standard Base needs to complete more. Migration between Windows to Linux has to be made more a gradual process. Instead of the complete jump of the deep end. Wine has only minorly softened the thud of changing OS. http://openlina.org openlina allowing linux applications on Windows will soften the thud. KDE4 being cross platform also reduces the thud.
There is a risk with openlina since the applications run on All major platforms and most minor platforms it will force a problem. There is one other problem openlina applications can run faster on linux since they can just use all native parts.
Openlina project exists because we see our windows users missing out on open source programs some of them massively better than what any windows only user has ever seen. Same with KDE coming to windows. Next year will be interesting. Windows using companies will have the option of reducing there software bugget massively and in the process more linux friendly.
OpenOffice Writer does have an Outline feature. I'm not sure how it compares to MS Word but it's there and it works well. Go to menu item: Format...Bullets and Numbering...Outline Tab.
"good enough is the enemy of best." Agreed.
I can understand that as a professional writer you've become attached to the outline features of MS Word. But you seem to overlook the fact that the majority of computer users are not professional writers, and most of them don't need those outline features. I'd argue that OpenOffice Writer is a "good enough" wordprocessor for the majority of computer users, and you can download it for free.
And Linux has recently become a "good enough" operating system for the masses. Thanks to the desktop-oriented Linux distributions, Linux is no longer any more difficult to use than MS Windows. Also, the price of Linux is quite competitive because it's free.
So I don't think making Linux popular is a "mission impossible" at all. I predict that the popularity of Linux among the vast majority of computer users will rise considerably during the next few years. It's inevitable because Linux is "good enough" and it's free.
We can all argue day and night about which OS is better and which is worse, but ultimately it comes down to the end user. For example, in 2004 it was clear to everyone in the world that our current president was absolutely unqualified for the job, yet he received the necessary number of electoral votes to remain in power. A similar thing is going on in the use of operating systems and programs. I am a technically well versed Latin and French teacher in a private high school, and I have successfully installed Ubuntu on a student's machine who never complained about anything, but when something didn't work he brought the machine to me, just like all Windows users bring theirs to the tech department.
I am sad to say this, but not only our adults are hesitant to fix (or learn how to fix) their problems, the youth is also on the same path. Our youngsters don't care that Apple's music file format sucks or MP3s are patented, all they care is to play that song or watch the video or make that cell phone call. For those of us who like to study and know the machines we use Linux will always be the best, even though we agree that Macs make better movies and Windows is used by 95% of the world (although not for too long).