Five Common Misconceptions About Linux
Having used Linux solely for nearly four years now, I've gained a respect for what Linux can and can't do. By no means is it the perfect solution for every problem, but there are some misconceptions heard again and again that I plan to set straight.
1. Linux is Behind the Times
One comment heard often is “Linux was five years behind XP, and it's 10 years behind Vista!” Well, here are some facts:
- Windows began separating the basic user from the administrator account by default in Vista, over 15 years behind Linux.
- Windows added a firewall in 2001, over seven years behind Linux's 1994 addition of ipchains.
- Linux was the first operating system with x86_64 support, beating Windows XP Pro x64 by two years.
- Windows added an attractive 3D accelerated graphical interface in Vista, a full year behind Linux's XGL.
- Linux's package management system can install, uninstall, and update software from one interface. Everything installed from Apache to OpenOffice and Quake 4 may be updated with one press. Windows has nothing like this on the road map.
And Linux isn't slowing down. The Xen project has added an incredible level of virtualization to Linux, with more work going into the kernels development to add enterprise ready virtualization built-in [4]. Microsoft promised built-in Xen-like virtualization in Windows Server 2008 next year, but has announced that feature has been delayed and should be available sometime after launch [1], possibly in SP1, meaning Linux will lead with built-in virtualization by at least a couple of years before Windows catches up.
2. Linux is Hard to Use
Many have never realized they were using Linux, and haven't used it on a desktop. More troubling is the fact that lots of technically inclined persons tried Linux during the hype of the dot-com bubble, wrote it off and never revisited it. These along with other factors have left many thinking Linux is hard to use.
Well, enter the modern Linux distribution, such as Ubuntu. Ubuntu has an easy to use graphical interface that'll remind Macintosh fans of OS X. Optionally many other interfaces are available ranging from Windows XP duplicates to interfaces focused on certain areas, such as low system requirements or high-end graphical effects. Beyond this many common tasks and features, ranging from system updating to system wide indexing, are all handled automatically by default.
This all extends to every level of Linux use. Novell's Yast for example provides an easy to use GUI for everything from installing and updating software to managing DNS, email, and web servers, and basically anything else an administrator could think of. No command line or configuration files, unless desired.
To top it all off the installation is world class. The Ubuntu installation is done from within a fully functioning environment allowing web browsing, game playing, or or even the writing of a report all as the installation wizard ensures the install goes off not just without a hitch, but in a manor where the user doesn't need to know anything beyond how to click next, unless they want to.
3. Linux isn't Compatible with Anything
Everything from Maya and Oracle [7] to Firefox run on Linux natively. Games ranging from the Doom, Quake, and Unreal Tournament series to smaller gems like Darwinia all run native on Linux as well [8].
Beyond native applications free (non)emulation software called WINE, as well as commercially supported options like CrossOver and VMWare, allow users to run everything from iTunes to MS Office and Photoshop, and the $5 a month Cedega lets gamers play hundreds of Windows only titles, from Battlefield 2142 to World of Warcraft.
Finally alternatives to Windows only software can replace current systems with little to no extra work. Apache can run ASP code [13], OpenOffice can read and save Microsoft formats, and every major distribution can join a domain, or just browse Windows file and printer shares, with ease.
Hardware support is equally incredible, in fact Linux supports more hardware than any other operating system. From hand-helds to mainframes and everything in between, including equipment considered legacy and no longer supported by Windows, the chances are if connected to a Linux box it'll just work. Despite popular belief this does include a vast majority of consumer equipment as well, from digital cameras to iPods and 3D accelerators to wireless cards.
4. Linux isn't Enterprise Ready / No One Uses Linux
Amazon and Google [15] would disagree as they've built their technology off Linux. PSA Peugeot Citroen, the second largest car manufacturer in Europe, also announced earlier this year they'll be moving not only their 2,500 servers over to Linux, but also their 20,000 desktops [16]. Other companies like IBM and Novell have reinvented themselves using Linux as the base, and government deployments from Brazil [19] and India [20] to China and others promise to add tens of millions of new users to the Linux community.
This isn't even including the countless smaller government deployments like the city of Munich [22], the Indiana school system [23], or the U.S. Army's Land Warrior program. Paired with millions of users via the One Laptop per Child initiative and massive academic deployments, this means that outside of the United States the world is positioning Linux to be the foundation of computing for their children.
Of course Linux works fine outside of the enterprise. Whether it's browsing a website, chatting on a cellphone, checking email, getting cash from an ATM, or even just kicking some anti-lock brakes into action, there's a fair chance Linux is in control.
Then again Linux also accounts for an estimated 70% of the super computing market [25]. That means Linux has huge footholds in the embedded, server, and high-end market, leaving the desktop arena clearly in its sights.
5. Linux isn't Professionally Developed or Supported
It's true Linux started at the hands of a single college student, but that's not true today. Linux is now a multi-billion dollar global technology. The vast majority of code is now contributed by professional programmers [26]. Over the last year major code modifications have been submitted by IBM, Intel, Novell, VMware, and countless other big tech players. Beyond actively developing code others, such as Dell, have begun pushing vendors to develop higher quality Linux software [27]. And this isn't even going into the academic or government development, such as the security patch set developed and deployed by the U.S. National Security Agency for internal use, but available to anyone who wishes to use it, SELinux.
Support has taken on the same level of professionalism. Countless forums, IRC channels, and wikis are of course still available. But beyond that help can be sought from one of many books, certifications, or by contacting any one of the major players from IBM and Oracle [29] to Novell and Red Hat. If 24/7 global support in a dozen different languages is needed, it's just as available as free community support.
That about covers it. There are many other areas of interest, but those listed above are certainly some of the biggest misconceptions heard about Linux on a regular basis. Overall it just comes down to ignorance, be it having never used Linux, or having not used it in the past few years of heavy evolution. Of course Linux isn't without its faults, just like all software, but that's for another article. What it really boils down to is a responsible administrator has to do what's best for the company cutting the checks, and that includes keeping an open mind and evaluating all options, even open source ones.
------
01. http://www.microsoft-watch.com/content/server/longhorn_server_comes_when...
04. http://www.linux.com/articles/59150
07. http://www.oracle.com/technologies/linux/index.html
08. http://doc.gwos.org/index.php/Native_Games
13. http://www.apache-asp.org/
15. http://www.wired.com/science/discoveries/news/2002/08/54504?currentPage=1
16. http://www.tectonic.co.za/view.php?id=1347
19. http://www.wired.com/techbiz/it/news/2003/11/61257
20. http://www.digitaltippingpoint.com/?q=node/106
22. http://news.zdnet.com/2100-3513_22-6119153.html
23. http://news.com.com/Indiana+schools+enroll+Linux/2100-7344_3-5820237.html
25. http://www.top500.org/stats/list/29/os
26. http://www.computerworlduk.com/technology/hardware/processors/news-analy...
27. http://www.phoronix.com/?page=news_item&px=NTkxOA
29. http://www.oracle.com/technologies/linux/index.html
Printer-friendly version
Comments
Features of Windows Server 2008
Not pro or anti-microsoft, but here's some info on the bleeding edge of windows server 2008
Betanews - Top 10 features in Windows Server 2008
Where's the Innovation?
First off in all fairness to Microsoft – I like where they're going. They're finally doing some things I think they should have done awhile ago (like standardizing errors), but that brings the question: why are they playing catch-up with Linux instead of innovating?
To address a few of the items from that article. Why is Microsoft still using NTFS? Linux solved the problem of file fragmentation years ago but Microsoft has canned WinFS so they just keep hacking away at something that's almost 15 years.
I'm also turned off by the mention of the kernel transaction manager as its a fix for registry corruption. Why not focus on the core problem: why does Windows still use a binary registry? You can corrupt Linux configuration files all day and you won't lose any more data than the part of the given file you corrupted. They can accomplish the same job and configuration files also have th added bonus of being much more user friendly to power users as they allow commenting. There simply is no advantage to the Windows registry.
Additionally to build upon my main article Linux has had ASLR for seven years now via the PaX kernel patch – meaning Linux will have beat Windows there by over eight years by the time Server 2008 is out the door (though to be fair Vista has this, meaning Linux was only seven years ahead on the client-side). Then they're also playing catch-up with PowerShell and the Server Core feature – both features Linux has had since inception. These are simply ideas I can't imagine have taken so long to get into Windows.
http://www.mi80.com/
Nice write-up. :) Just a
Nice write-up. :)
Just a couple additional points:
RE #1: Not to mention, Linuz was born around 1992 and still had true pre-emptive multitasking before Windows did in '95.
RE #2: I've gotten to where, depending on the definition, I find Windows much HARDER and I'm sure many others have too. The more I'm exposed to Linux, sure the solutions can be more difficult (for the things I'm trying to do, that is, not normal computer junk) but once they work, they generally just work. Forever. Windows is much "harder" in the sense that there is not single place where error messages are logged, no single place where configurations are stored and can be copied, etc.
Lately, I've come to two important realizations. One, I'm never going to catch up to the "Linux gurus" no matter how hard I try. Two, as long as I can read documentation, search knowledge bases, and ask decent questions, I'll never have to be a Linux guru.
RE #3: Most Linux distros are compatible with just about anything you throw at them right out of the box. Whether a ten year old token ring card or a month old video card, it's rarely stumped on the hardware side and on the software side look at all the crap to chase down after a Windows install: support for various compression formats, PDF readers, DOC and XLS editors, NFS support, etc.
The only places where I've seen Linux struggle are (1) full compatibility with Exchange in any mail reader, (2) support for chipsets where manufacturers don't open the specs, and (3) full interoperability with obscure network "protocols" like MS-DFS and NTFRS. Not that any of these are the fault of Linux developers, but it's one hell of a track record considering the lack of help and outright bomb-throwing that so many vendors have done.
RE #4: People have no clue how common Linux is in the enterprise class. Although most of my clients don't get this Linux stuff at all, they certainly like how I almost never ask them for money to cover software at the server level.
In fact, there are only a few reasons why I have yet to advocate wholesale movement to Linux servers for EVERYTHING:
1. Exchange/Outlook (and that stupid calendar everyone seems to love) is so damned pervasive and there just isn't a clean transition
2. There are a few oddball things like DFS (Samba supports most of it but not replication... only took me a year of tinkering to realize that) that I can't migrate over to Linux without making drastic changes to entire infrastructures overnight
3. Active Directory makes 90% of the domain administrating tasks so freaking easy that it's hard to beat as a "LDAP" (note the quotation marks) server. Too bad the other 10% is virtually impossible or requires some sort of hack.
RE #5: It's hard to set foot on the Web without tripping on Linux support. ;) From what I've seen, the main problem is people are pretty clueless on where to/not to go for help.
None of these are 100% true in all cases, but for example:
1. Don't go to a Debian forum for help when you are installed Samba from source and you have a configuration problem.
2. Don't go to a Gentoo mailing list and expect help with Yum or APT.
3. Don't go to a BSD chatroom and ask whether (linux based) OpenFiler or (BSD based) FreeNAS is a better choice.
4. Don't go to a Mac user (most of whom can't seem to go 5 minutes without making some comment about OSX being based on Unix or some such thing) and expect them to know anything about Linux. They'll probably just convince you that BSD would be better but they won't know much about that either.
5. Don't go anywhere to complain about a driver not being available other than the bloody manufacturer.
But again, nice article. :) It's good to see some fellow level-headed Linux advocates out there.
Great Points!
Very good addition on point number one. I didn't even touch on multitasking and scheduling. I think currently Linux has a bit of an edge but an edge none the less – but then lately there have been some big strides in the Linux scheduling scene with the CFS and such which I think will help Linux pull ahead even more.
Also I agree with your second comment, I find Windows harder as well. It's actually simple: in Linux I have the ability to fix things, and never does a problem occur which I can't track down. In Windows something like a BSOD will happen and I simply have no way to fix it. I don't have the source code, and in Vista some processes are locked away so they can't even be debugged. I just have to update things and hope someone fixes the problem. I don't like feeling powerless and that's how Windows leaves me.
As far as compatibility I agree it's traditionally been a huge problem. But with recent additions like full NTFS read/write support, mainstream Broadcom WiFi support, and vendors opening wide to get Linux support (like the recent AMD/ATI announcement that they're going to slowly open source their drivers) this is quickly changing. Also for exchange support check out Evolution – it's by Novell and has an excellent Exchange plugin. It's not perfect but has worked well for me for a year or two now. Also I'd recommend checking out SLES10 and ZENworks. They're not perfect but they'll do most of what you mentioned regarding Active Directory.
True on the support area though. I'm always careful on that topic because there are in fact some bad places to go, but in my experience I've found many, many more positive experiences so I try to focus on that.
Any ways – great comments, I love getting some discussion going!
http://www.mi80.com/
The last couple days has
The last couple days has been a perfect example of how frustrating Windows can be. I have a client set up with multiple DFS servers in different sites. About every two months, *something* screws up and I end up having to recreate a DFS root or do a "non-authoratative repair" or whatever (all ends up the same in the end).
Currently, DFS is the best solution that I have without overhauling everything. I'd like to move to some sort of iSCSI setup but I have to work out the details, build the servers, and do everything at once. I'm stuck in Windows with this because Samba still doesn't do DFS replication. I may eventually do something with rsync but I can't find anyone who's done it yet and don't want to just throw it out there on a client so that's where I am.
Anyway, something happened. :) It looks like the Jet DB that handles the DFS on one server got corrupted (by and antivirus I suspect... that's a whole other issue) and so it wouldn't replicate. As it happens, there is no way to recover. The M$ workaround is to do a D2 repair... which is essentially starting from scratch. You can't pre-sync or anything like that and there really isn't a nice way to monitor any of the progress. So I'm stuck watching all day trying to catch people logging in as their profiles sync and possibly corrupting ntuser.dat or getting a folder collision or whatever.
I could go on all day but suffice to say it's a nonstop freaking nightmare. People keep telling me how glad they are I understand this stuff, but the truth is I shouldn't *have* to understand that crap because it's totally ludicrous; this crap shouldn't happen at all. I end up having to check the whole DFS mesh every day to make sure it's replicating and that it didn't just randomly delete a bunch of folders. It's typical Windows crap, but I've NEVER lost sleep or gotten up a 4am to check on a Linux server, that's even the ones on crappy hardware.
That reminds me of another point. :)
If I have a Linux server with some sort of RAID setup, I'm pretty comfortable because as long as part of the RAID works, absolute worst case I can slap the drives in a new machine, make a few adjustments, and I'm back in business. However, with most of the Windows servers I'm exposed to (and their OEM licenses), I can do no such thing without (a) using a pirated key, (b) hacking product activation, or (c) waiting for a replacement part or parts. In real life, it's usually a combination that occurs to buy time while waiting for parts. None of those are particularly appealing, but even more there is no reason for it to be that complicated. WPA and licensing won't stop any criminals, but it sure makes Windows hard to provide support for.
US Navy
You can also add in there that the US Navy uses Linux to power the fire control and sonar systems on board the latest Nuclear Submarines. These are critical and heavily relied upon systems for the safety and security of the entire ship.
Linux/Windows
I have three PC's (one is a laptop for my wife ) I have XPPro on one and Ubuntu Feisty Fawn on the other ,
I need the XP machine for ONE program (a genealogy program that is Windows dependent)
At this stage I get Everything I want from Ubuntu and it is getting better with every release .
I personally don't care what I am running , the fact that I can do Office , Browsing , I Podding , Graphics ,
Mail , and a host of other Apps for the financial investment of nil or a few dollars donation obviously makes the argument a non contest . I have Java , Realworks , Media players with all the codecs , all apps install or remove with ease these days , current Firefox and Thunderbird and just about any thing you can think of.
The Day that Gates and Ballmer offer me that would be the day I might go back .
And before anyone screams lack of technical support ...in 10 years of using /Unix/Linux , I have not needed any and I am no Guru I can assure you of that .
As for driver support , I bought a $40.00 Hp Bubblejet last week , pugged in the USB and away she went .
It's Ubuntu hands down !!
Almost in agreement,
I would say that most of your conclusions are pretty obvious once you start looking at the history of Linux and Windows, but I would say that your harder point is frankly invalid, and I will tell you why.
Most people doing non-technical work who are under 30, grew up using Windows. That means you have this huge inertia of Windows users who know how to do what they want (Internet, email, music now and then, and maybe a DVD) and they are very stubborn at changing because they will have to memorize a new system. Notice I said memorize and not learn, and this is the crux of it, most people who use computers for entertainment or to achieve very obvious things like using spreadsheets, typing up documents, manipulating images, etc. And so do not actually understand the platform they are working on.
It is like expecting a person to understand a playstation when all they want to do is play games. I don't think it will happen very soon, I think most of the people are just going to upgrade to vista unless some distro is pre-installed on their box by the OEM.
The Larger Scale
I think the key to note there is that that's inside the United States. Traction of Linux, and open source in general, isn't nearly at the percentage levels in the U.S. that it's approaching throughout Africa, Asia, South American, and even Europe. I don't think the U.S. is going to lead the open source world simply because why would we? One of the biggest motivators of Linux is cost, and we're the richest nation on Earth.
But if you look at the global scale where the other five billion or so of people who don't own computers that's where I'm talking about. Outside of Beijing the average Chinese person makes US$2-3 a day. Even at a 90% price reduction that means Vista Ultimate would still take a fair portion of their monthly income assuming well, they stopped eating and such. Additionally again on the global scale hardware isn't up to par with ours. If you tour Brazil you won't find the average computer to be a multi-core monster. And that's assuming Windows IS an option there. There are over two hundreds languages currently in use around the world and I assure you Windows doesn't come in all of them. As new markets emerge Windows simply isn't an option for some. These aren't problems we Americans have, so it's why we probably won't see Windows being replaced heavily anytime soon.
I would say you were right
I would say you were right until just recently. Windows WAS easier for the average user exactly because the average user was used to Windows. However, all that changes with Vista.
The first time I used Vista (and Office 2007) I found it to be one of the most difficult (computer environment) adaptations I've had to make in recent history. I'm a Mac OS X user at home, Win XP user at work, and I've tried out several flavors of Linux. Each time I've tried a new environment, there were things to learn and get used to, but they were easy to pick up. The last time I had a shock anything like Vista-shock was actually the transition between the old Mac OS and Mac OS X. Two things helped with that: 1) I tend to believe Apple knows what they're doing concerning UIs, and was willing to give them the benefit of the doubt, and 2) Apple really does know what they're doing with UIs. I think the Vista shock is actually even greater, and so far seems pretty senseless from a UI perspective. Perhaps my opinion on that will change once I've used it more.
There are a couple of interesting factors that come into play with this. Microsoft is trying something that Apple tried successfully several years ago. However, there are several differences between the two situations:
1) Microsoft's user base is not as loyal as Apple's. Rational or irrational as it may be, this is pretty clearly true, and will almost ALWAYS be the case with the market leader. This could hurt MS. Additionally, Microsoft's damaged reputation for security and quality (deserved or not) has alienated some of its customers, who may now be looking for a reason to leave.
2) Microsoft's user base is, on average, not as computer literate or "computer intelligent" as Apple's. This is not to say that everyone who chooses Windows is a moron. Rather, it is once again a function of being the market leader. It is because the average user will typically just buy whatever his friend or family member has, or whatever is most readily available at the store. Until recently, very few computer stores sold anything but Windows PCs. Choosing something else often indicates that you've thought about the choice, although choosing a Windows system does not imply that you have NOT thought about the choice. This could hurt MS, since low computer literacy makes a person more resistant to changes in their computing environment, and could keep them from upgrading until absolutely necessary.
3) Microsoft is the "safe" choice. That is, the brand is known. People feel safe choosing it because it's big. They've already invested in it (though they're losing the investment of learning it). This will help them, especially in the long run when people are forced to choose between upgrading and switching to something else. This, more than anything else, will help Microsoft, and might be the most powerful factor. I actually believe that this factor alone will drastically minimize everything else.
4) It's not at all clear to me that Microsoft has improved their UI. The change represents an immediate loss for Microsoft, due to the unfamiliarity of the interface, and the barrier that erects to upgrading, in exchange for no clear long-term gain. Again, I could be wrong about this, but when it comes to UI decisions, I don't see the sense in what they've done, and I don't trust them as much as I trust Apple in this area.