Slashdot Mirror
Linux will have 20% desktop market share by 2008?
unmadindu writes "Siemens Business Systems, after conducting an extensive survey on non technical workers ("secretaries and managers, not IT people") is predicting that the Linux desktop will capture 20% of the market for desktop computers in large enterprises within the next 5 years. Senior program manager Duncan McNutt, who has overseen Siemens's testing of Linux desktops with users and administrators in enterprise settings, believes that the Ximian desktop and application suite, running on either SuSE or Red Hat, requires two days of training, which is the same as what most enterprises budget for a Windows/MS Office version upgrade. Interestingly, they used Ximian Desktop, instead of KDE, because Gnome, particularly Ximian's version, was "different enough" to set user expectations that the experience would be less like Windows. "
What if the kernel used year 2008 is the Hurd? Is it still "linux". We should really speak about free unix like operating systems.
"Siemens found KDE to be more "Windows-like" than Gnome, but that lead to problems when non-technical users expected a more Windows-like experience. Gnome, particularly Ximian's version, was "different enough" to set user expectations that the experience would be less like Windows, which led to fewer adoption problems."
Need more reasons to have at least two different desktops?
Interestingly, they used Ximian Desktop, instead of KDE, because Gnome, particularly Ximian's version, was "different enough" to set user expectations that the experience would be less like Windows.
Mark me down as flamebait, but perhaps this is truly important. Perhaps we as a community should stop trying to mimic existing applications and begin innovating instead. Certainly, a good user interface is necessary, but is Windows truly the best user experience? OF course, it's ridiculously hard to come up with a new user interface that is logical and easy to use. After all, a button is a button. It can't really get much better than that, but perhaps there is room for improvement.
I still remember the first time my girlfriend saw me running Linux and said that that looked exactly like Windows and then asked why would I bother going through the hassle of installing Linux when I could just use Windows, which was preinstalled and already worked. Keep in mind that she saw me using KDE and Gnome. (I do realize there are other window managers in this world.)
She had a good point. Windows 2000 and XP have been much less crash-prone, and I find myself increasingly using Windows XP and Mac OS X instead of *nix as my desktop OS of choice. Instead, only servers that I must work on use Linux, and I simply SSH into them, skipping all of the GUI nonsense. For me, the best user interface in Linux is the command-line - not the GUI that looks like Windows anyway.
Nobody can really predict the direction the computer industry is moving in the next 5 years. The technology is still very young and futher has a very high innovation speed. Prediction over such a long time range are rubbish.
Just remember the classical examples of such predictions getting fucked: AI, "processors beyond 300 MHz are physically impossible", "640 kB is enough for everyone", "OS/2 is the system of the future" etc.
And for Linux: there is hot stuff like Grid computing, immersive VR, Quantum computing etc. on the way and I don't see even the smallest efford to integrate this into Linux.
The only thing we can predict for the next 5 years is crackpot MBA doing academic, oops non-academic of course (we can't insult academics), circle-jerks and spewing out rubbish predictions.
Ha, outsource everyone to India.
Owner of a Mensa membership card.
To me, the really puzzling thing is why people who claim to not even bother to read the predictions bother to write about them.
...but this is extremely unlikely. In the event that they pull MS's entrenched ass out of the corporate world, maybe. People would be a lot more willing to run it at home if they ran it at work. Furthermore if Linux holds 20% you're going to have compatibility problems up the wazoo(sp?) The reason everyone uses Microsoft products is because it works[sic] so well together.
--fetch daddy's blue fright wig, i must be handsome when i release my rage
The two days of training may not seem significant. But you have to realize that most people already know how to use MS office/outlook by using it at home or school, so by the time they get in the workforce they don't need training. For Linux, they will need a day or two of training.
BTW, anyone have a link to Ximian desktop? I use Ximian evolution for email, and think it's a nice program. I tried Thunderbird, but it's still not quite there yet (I know it's only 0.1 or something, but I want something that works, not something I have to bug-test)
You've hit the mark. The only way people are going to dump Windows for Linux or any other OS is if that OS has a compelling difference that makes it worth the change. While one can argue that Linux or *BSD is more stable that is hard to demostrate without prolonged use and if the system is too close to Windows yet not quite there as is the case with KDE then users will be frustrated and leave it before the realize that it is more stable and more secure. The very things that make Linux the better OS are the hardest for end users to see.
Slashdot, home of supporters of free software, free music, and free speech.Except for Moderators that disagree with you.
How quickly we forget. Just a couple days ago a gent wrote to slashdot stating his company would pay, what was it, $350K to RedHat for their latest pricing scheme. That's free for very high prices of free. Oddly, it seems that the higher price tag adds some credibility. While Debian was a very popular choice for the replies, Debian lacks official support and the software vendors stamp of approval, which many (most) companies consider essential, hence was a moot point made over and over and over again.
Yesterday I put in some overtime by working in our manufacturing plant. Wore my nifty CopyLeft baseball cap with the backwards C. Several of the college kids asked me what it was. When I told them, not a one had a clue as to what I was talking about. They never heard of GNU or linux or opensource/free software in general. That didn't leave me with a warm fuzzy feeling for the future.
In our IT department, while I'm not what one would call an advocate, whenever I mention linux or anything in the free software genre, I quickly get an "oh shut up, it's junk" reply.
20% by 2008 is just a silly pipedream. The major problem seems to be mind share. Folks who aren't fanatical about Windows are seen as the bad guys who infect computers with viri and engage in other immoral activities that cause trouble. I picture a witch hunt type senario against linux et al rather than a major acceptance.
-----
The mob moves like demons possessed
Quiet in conscience, calm in their right
Confident their ways are best
The righteous rise
With burning eyes
Of hatred and ill-will
Madmen fed on fear and lies
-----
We've seen it over and over again, the better choice doesn't always win. The only way I see linux succeeding is by first dominating non-US markets. e.g. Company X sets up a shop in Country Y where linux is number one, they use linux there, it proves to be a good value, then is implemented in other locations. But given the Windows network design, I think that is even pretty far stretched, as incorporating non-Winodws network segments in their forest/tree thing can be a real pain. And, let us not forget, the US is basically the only place on earth that continues to reject the metric system.
Well, that's enough Sunday morning babbling :P
Here's a hint. The large businesses don't care what you think about "the right way to gain marketshare". If a business' IT department decides that it wants to cut costs and switch to Linux, you're going to have to accept it. This article isn't talking about home desktop computers. It's about business computers. Why is it unlikely that companies won't switch to free software that is perfectly capable of being used for email, writing, R&D, etc.?
It doesn't matter how easy it is for YOU to install, beacuse that's the IT department's job in a business environment. They do the maintenance. They set up the machines. They put apps on your desktop so that you can brainlessly point and click your way to "productivity."
Finally... Why must all versions of MS Office work seamlessly when they DON'T EVEN WORK THAT WAY TO BEGIN WITH?
Guess what. Your company is going to switch to Linux in a few years, and you'll be eating your fucking hat.
"Perhaps we as a community should stop trying to mimic existing applications and begin innovating instead."
Like Apple?
that Linux shouldn't necessarily be trying to emulate MS Windows' dekstop so much as making one that's better even though different.
Generally too many choices for the end user (read jane secretary, or joe PHB) are BAD because it confuses them and creates IT maintenence nightmares.
It is true that if you try to create a gui interface that is just like MS windows, except you differ in some crucial areas, the user will be put off by the "well windows doesn't do/have that" comparison. However, if the user expects it to be different (because it looks that way or obviously acts that way) then their expectation base is "Hm... this is neat, I will have to learn it" which creates a whole different set of expectations that are not driven by comparisons to MS Windows.
Is the juice worth the sqeeze?
A point is made here. GNU/Linux distributions should look and feel differently than Windows or other proprietary OSes, not radically but still enough to avoid the kind frustration that I've seen when people expect graphically equivalent desktops like GNOME or KDE to behave like Windows. They will never be perfect replacements for it, nor should they be. Interestingly, I've seen much more success when demonstrating window managers like Fluxbox. People immediately fall for their simplicity. They just love to have one simple, well organized desktop menu and no annoying icons or toolbars to push around. No nonsense user interface is what they like. They even tell me afterwards than Windows desktop is a mess compared to them! I hope that these lightweight WMs will gain grounds in the future, because frankly even if their libraries have become essential parts of many applications, GNOME and KDE do not look or feel like the right desktops for GNU/Linux (IMHO). Something about the UN!X philosophy of having one small and efficient tool for the job makes me (and a surprising amount of novices) more comfortable without them.
./configure --enable-shared --disable-static && make world clean
....[McNutt] believes that the Ximian desktop and application suite, running on either SuSE or Red Hat, requires two days of training, which is the same as what most enterprises budget for a Windows/MS Office version upgrade.
Now I know why people call in to tech support with such rediculous problems. Perhaps M$ apps could be made more useful if the people that relied on them were better trained in the techniques of using a windows system.
So what will happen if businesses were to migrate to a linux platform that's completely different from windows? Would the average desktop user really be able to pick up the necessary skills to use linux effectively in a matter of a couple of days?
I have a feeling that insufficient knowledge and interest in learning a new system is the reason that linux hasn't already claimed a larger share of the desktop market.
Before you can have a smarter desktop, you need to make smarter users.
~Mike
Mike Rizzo
Although I may be being shortsighted, I don't think it can get much better than a wheel. However, there are other things that could use improving that people work on all the time - the placement of dials and gauges, signal controls, headlight controls, position of the shifter, environmental controls, HUDs...
So, like Windows and other desktops, for a GUI, there are certain givens (until other hardware becomes commonplace, like maybe gloves), there is a pointer, there is a graphical background, you click on things to make something happen. In that way, yes, they are all the same. But the positions of items, the labeling, the color schemes, gestures, voice interfaces, these are all things that someone could (and should) run with to make the "standard" in the future.
I guess what we are (and should be) driving towards, no pun intended, is customizablilty. That's one of the strong suits. There is something to be said for consistency across applications (like menu hot-keys), but there can be a whole world of things besides that. What we should be driving more towards (and again, you see this more with Linux and Open Source applications) is being able to assign the hot-keys yourself.
Perhaps a system wide defination to make all applications behave the same way, but make that definition customizable. I don't mean "skins", although that's part of it, but skins don't necessarily change programs functionally.
In any event, it IS worth trying. With all that said, GUI is NOT a better interface than the command line all the time. Certainly for a lot of things, and if I had to choose I'd just stick with the GUI, but I tend to have at least one shell open (in Windows or Linux). The best thing on Windows is having CYGWIN and a "Bash Here" (or "Bash This") context option when you right click on a folder.
Stupid sexy Flanders.
> If the development trend goes that way, then we'll start hearing "For Linux to be accepted in the home and enterprise it must be much more like Windows"
You mean like we have been hearing for the last, what? 5 years.
1st rule of Linux club: You must agree on a single distribution
2nd rule of Linux club: You must agree on a single packaging format
3rd rule of Linux club: You must agree on a single desktop environment
4th rule of Linux club: You musy agree on a single web browser
5th rule of Linux club: You must develop a groupware suite
6th rule of Linux club: You must NOT mimic rival OS's
7th rule of Linux club: You must lay down your holier than thou ego's
8th and final rule: if this do this, we will all bask in the glory of Linux on the desktop.
</QUOTE>
I might add the following:
9th rule of Linux club: You must not reinvent the wheel
10th rule of Linux club: You must get rid of YetAnother***
hell-spawn program derivatives
Just my 0.02 euro
More importantly, by that time why would a business user have anything more then a dumb terminal. X server already has proven itsself to work with this model, why would anyone have to open applications at all locally?
Largo Florida already has done this, saving millions of dollars and is the easiest system to administer. Its users just use it, they dont care it its windows or linux (its KDE).
All of these discussions about desktop Linux seem to take place in a vacuum, or at least an ideal situation where office workers need no more than a suite package and a web browser and the training to go along with them. Sure, it's trivially true that if We Had To It Over Again, we'd come up with something better than Windows (even MS probably agrees on that point).
However, most organizations are tied down to the massive installed base of Windows applications. In larger corporations, the numbers of internal/vertical applications can even rival the number of employees. The OS is nearly meaningless compared to the importance of these apps to a business. The cost of porting to Unix/Java/Web would be absolutely gigantic -- if the tools were even there to do it (and it's not clear they are).
So, while incrementally lower administration costs are nice, it's simply impossible for COA considerations to overwhelm the absolutely massive transition cost of rewriting all of those apps.
[...] most organizations are tied down to the massive installed base of Windows applications. In larger corporations, the numbers of internal/vertical applications can even rival the number of employees. The OS is nearly meaningless compared to the importance of these apps to a business.
This is absolutely true. Once the cost of the migration exceeds the recurring yearly gain over a five year time horizon or so (doesn't matter why; could be porting apps, hardware transition, licensing) then you've got a no-go situation. Wine might be a partial solution to that specific problem, but each app would have to be rigorously tested in house and certified with management as to functionality. That testing would certainly cost and should be factored into the migration costs.
I'm not sure a large organization with many prior internally written and platform specific apps would save significant money from a transition to Linux. It's catch as catch can. Per desktop, Linux is certainly cheaper to manage in large deployments than Windows. It gets significantly cheaper the larger one scales up. But if your apps don't run and the workforce can't work, you've lost the whole point behind a deployment. So, I would certainly agree and wouldn't recommend Linux in that situation; especially if the organization in question wasn't interested in a potential wine solution to running their apps.
Now, back to the yardwork,
Cheers,
--Maynard
You get the same thing on windows. I do support for windows and frequently tell people to RTFM, not in those words, but something along the lines of I have no idea what application foo does, why don't you try googling for it etc.
Basically my point is that windows has it's arcane voodoo just as much as linux. People in large corporations get trained on their internal systems such as billing systems, databases etc. etc. Most of the problems non-technical people seem to have with a new user interface is a fear of looking around, they're told to click on file and open in their training, so they expect it to be there.
What all desktop environment projects should focus more on is to develop applications, that are/might be used by school kids and students. Educating the youngsters with the beauty of Linux is the most essential thing. Kids at my age (I'm 22) are using mostly Windows, because it was *the* OS for playing games and doing homework back then when I was at high school.
Changes do not happen overnight. It took many many years for Microsoft to get to the point where they now are (desktop dominance) and Linux may have to face similar, unfortunately even longer growth phase and time.
To answer a question that will probably pop up in a reply to my post, yes, I did read the article and actually printed it out. It was greater than any work of Shakespeare! :)
Interestingly, they used Ximian Desktop, instead of KDE, because Gnome, particularly Ximian's version, was "different enough" to set user expectations that the experience would be less like Windows.
You see, all you people who think the Linux desktop needs to be "more like Windows?" If you go the path of "like Windows" then you have to make Linux exactly the same as Windows or ex-Windows users (99% of the population) get confused.
On the other hand, as this story says, if the desktop is different enough from Windows, people automatically (because of psychological reasons) know it is not Windows so they expect things to be different, and are more open to the change.
Incidentally, they mention that training lusers on Linux takes 2 days, the same as a Windows upgrade, but I don't remember if they mentioned this: Upgrades to the Linux system (other than automatically administered bug patches for security reasons) won't need to take place as often as for Windows systems.
- Linux makes better use of the hardware.
- Open standards and open source on Linux means that nobody is forcing you to upgrade.
- Unlike in the Windows world, where you must upgrade because the rest of the world is doing it, there is no such requirement on Linux, except for security related patches which can be remotely administered by the IT department without the user even knowing it.
This means that companies will have to spend many less two days to get users acquainted with changes to their computer systems.Even if more horsepower is required for some reason (which would, in the Windows world, require all 50,000,000,000 computers in a company to be replaced with faster models and new software), the company can install one or more big huge servers running Linux or any other UNIX and use the resources on that machine, leaving all or most of the users' machines alone. Again, the users wouldn't even know anything was changed... and that means savings in cost. (If you have 45,000 employees on computers and you have to train them for two days, that's likely to cost twice as much as buying six million dollars in servers. (Figure 45,000 people making $18 an hour, 8 hours per day, for 2 days... Add to that all the taxes, insurances and benefits you have to pay and you've got two really expensive days!)
Furthermore, the free software community reduces costs for companies, not only because of licensing fees but because bugs and security problems get found and fixed quickly, and new features are added when someone needs them... I imagine that as more "enterprises" make the switch, they'll hire some folks into their IT department to do nothing but develop Linux to meet their special needs, and that means that with thousands of companies worldwide doing this, in addition to tech companies like IBM and HP, and in addition to the already existing (and growing) developer community... Linux is going to continue picking up speed and inertia, and Microsoft, with their "little team" of 30,000 programmers, soon won't be able to keep up.
It is for all the above reasons that I firmly believe that companies that don't invest in Linux now will scamper to invest in it later... or be left in the dust.
Linux will go the same path.
The reason that training is more successful on a very different looking distro is mode change. Apple was going after this with OS X. All the whiners that complain it looks different and acts different from OS 9 for no reason don't understand that they are trying to get people out of their old habits and fully appreciate the power of the new experience. The same is true if your Linux distro looks decidedly unlike Windows. People won't have as many "Windows" expectations.
Big companies are already going through the portings, IBM rewriting almost everything in Java or Web services for example, SAP is being rewritten as web services, etc. All these can then be used either through a browser or other thin client, making the OS almost irrelevant.
Once that happens, companies can look at the cost of buying the next 3 year cycle of Windows+application licenses, and compare it to the cost of porting over or replacing the remaining windows only applications, and do the maths. For some it will be easy, dump windows, get in Sun or IBM to do a Linux thin-client or workstation implementation with servers doing the real work.
Ewan
No offense, but I think the fact that MS said "We aren't going to make any more OS's like '9x" anymore forced game developers to start making games that ran under the NT/2000/XP OS's.
"Look! There! Evil, pure and simple from the Eighth Dimension!" --Buckaroo Banzai
[The same Anonymous Coward]
Because Notes and Groupwise are already "embedded" in many corporations. Sure they may not be bringing home many new customers but once you've got everyone's email, calendar + X many Notes apps built, it's a big hurdle to switch to any competing Groupware product (Exchange).
Arguably, for companies, it's Groupware which locks you to vendor more strongly than your office suite (for which there are now very viable alternatives).
With Notes and Groupwise clients well supported on Linux, the barrier to entry is much lower, specially for those still on NT 4 (which often goes hand in hand with Notes)
I have been trolled. Have a nice day.
Briefly, point one is true, mostly because windows now crashes roughly as linux always has, i.e. not very much. Point two is meaningless; how do you have a "meg" of security problems? MS have a reputation in the security field and they have it for a reason. Point three is hyperbole, with a kernel of truth. But so what? I'm sure you occasionally find two windows programs that do the same thing. At least, until MS decide to own that particular market. Point four is relevant to use of linux in large enterprises exactly how? Point five - well, you might do that. Large enterprises tend to restrict their purchasing to supported hardware.
Nothing to fix and no-one wants to talk to you anyway. What a shame.