Letters: Linux, Love It Or Hate It
Frank Ohlhorst's recent column, "A Linux OS For All," asked readers what needed to be done for Linux to gain more traction. And, boy, did they have a lot to say about Linux: the key players, where the open-source platform is now and where it should be headed.
Here's a sample of the reader feedback (and be sure to read Ohlhorst's response at the end):
I agree with your point of view on Linux. I write about this stuff on my blog. The Linux guys need to really take their technology to the next level and take a user-centric approach rather then trying to be the jack-of-all trades to niche developers and hobbyists.
Alex Zaltsman
--------------------
I just read your column and have to say I agree. The potential of linux is really overlooked. You stated "How do you think Linux can gain more traction? Let me know." Here are my few thoughts on the subject. Being an avid Windows and Mac user, I have touched on Linux and tried it more than a few times but, like many, have sort of been intimidated.
1. It's great that you can do everything from a BASH Shell, but it needs the clear, simple ability to be able to do it from a GUI, too.
2. The "control panel"-type system on the Linux distros (and I've tried quite a few, always using the KDE look) is too confusing for the end user. Things aren't categorized into the areas the average end user would think to look. Easier navigation has to be considered if Linux is ever to become a real rival to Mac and Microsoft.
3. The last qualm I have, and the one I have always mentioned to friends: Linux is a geekish system. As a result, they're the ones compiling and designing the GUIs and everything else around it; Mac is a classic example of a "pretty" OS. The current look and feel (and, like you said, you can customize it, but the average end user is too scared) needs to be developed in such a way that its aesthetically pleasing. Windows XP's popularity was partially credited to its new GUI. And Vista is set to rival the Mac OS with its own alphablending and "aero" interface.
Linux utilizes the system resources fantastically. If these key issues are addressed, in my mind Linux stands a chance, and a true potential.
Hope this has helped, or given you some more angles to look from.
Kris
--------------------
Things Linux needs to be more successful include:
1. Better package installation (as you pointed out). Ability to install stand-alone packages rather than the current package repositories. Having packages in repositories is nice in theory but in practice no repository includes all available software. Only one package you need has to be missing for the whole setup to fail.
True, you often can download stand-alone RPMs and such but then you lose dependency resolution and the RPM hardly ever contains all required libraries, etc. What you need is a composite system whereby the application itself is in a stand-alone package that you can download from a Web site, and have the package contain the necessary metadata to do dependency resolution, then obtain automatically all the libraries and other dependencies from a repository.
2. One distro. Anything more than one distro is too many. There is only one Windows, one Mac OS-X. What we want at the end of the day is Windows, Mac OS-X and Linux, not Windows, Mac OS-X, Fedora, Ubuntu, Frugalware, Suse, Gentoo, Slackware, KateOS, DSL, Zenwalk, Debian, Knoppix, Gnoppix, Boppix, Sploppix, etc, etc, etc.
Why do we need 200 me-too Linux distros? Ultimately, they all have the same kernel, the same drivers, the same basic packages (Open Office, Evolution, Gnome, KDE, Gimp, etc). The only thing that differentiates them from one another is a different set of missing functionality, a different set of bugs and a different set of missing packages in the repository. Merge the best bits of Suse, Mandrive, Fedora, Slackware, Debian and Ubuntu and create a complete and bug-free Linux and get rid of all the rest. If people want to contribute to free software, let them all add to this one mega-distro. We don't need six or seven different package managers all of which do the exact same thing, but slightly differently.
NEXT: More Letters
--------------------
I'll probably have more insight for you next week--I am in the middle of installing Gentoo on my girlfriend's system right now. This will be the first time I have installed Linux for primary use by someone other than myself (other than as a Samba, Cups, Apache, etc. server).
I suppose I am a lot like you, I like to fiddle with the internals of my OS. I agree to some extent with your article "A Linux OS For All." Linux does need to be better with hardware support. I have six computers of various origin running Gentoo. All fully functional for their purpose: three server, three desktop of which two are laptops--quite the ratios! For me, it wasn't that my hardware wasn't supported, it was the hours of googling to figure out which kernel/other drivers were required. I agree improvement is necessary, but things are coming along nicely.
I don't think that the argument about the average user not wanting to deal with the internals is as valid as it seems on its face. As in many areas all across the country, I am the family computer guy. My mom/sister/nieces call me when questions arise. Not any of them are comfortable installing software in Windows by themselves, let alone fixing the install when Windows goes haywire. For example, I have a niece who loves her Instant Messenger. I reinstalled Windows (it was completely trashed) twice before I came up with the workable solution of one Windows OS partition, one partition for saved data and one small partition for Linux with PartImage installed.
The point: Make a b-zipped image of the Windows partition ready for restore whenever necessary. The advantage over traditional restore options: Restore the partition, update windows, re-backup the partition, then you don't have to get those stupid updates every time. Turns the reinstall from three hours, including updates, into 20 minutes.
The point of all of this is that I would have been doing the legwork anyhow. If the right software was there, I would have just installed Linux. Right means easily compatible with de facto standards, .doc, .xls and .qbw are the ones I run into in my daily life.
If something ever went wrong or she needed something installed, I could ssh in and do it! No more "Bring me you laptop for the night and I'll fix it.. Similar advantages are available for software support. First, I and most computer geeks can answer a typical user question about a piece of software that we have never seen. Second, I can ssh/vnc in and help.
I am, in my daily life a CPA. Believe it or not, that comes with a lot of tech support responsibilities. It gives me a lot of insight into the computer user's daily life. It is pretty simplistic. Internet, music, pictures, word processing. For small businesses (where Linux could really catch on) accounting software (Quickbooks).
That's it. Cover those bases, almost regardless of ease of administration, and you have a winner!
Aaron Hopp
--------------------
To give Linux more traction we should define what a small business needs to function and make these applications easy to install.
We could make a difference between different kinds of businesses like simple offices, graphical workers, etc. etc. Gnome and KDE should get closer to each other, they are on the right path,
The overall biggest problem is still the hardware vendors that do not deliver drivers. HP is on the right track, which is why I bought an HP printer and no Canon. Linux users should ask and ask and ask for drivers, not just consume what the community delivers. I give comment to all vendors that do not deliver me drivers and keep asking until they do.
Jan van Leeuwen
Dutchman in Austria
NEXT: More Letters
--------------------
While I agree to an extent with your "A Linux OS For All" column, my only problem is that driver support isn't in the Linux vendors' hands so much as it is in the hands of the hardware folks. Witness the problems with ATI, Wi-Fi in general and other bits and pieces.
If there were better third-party support from these closed source vendors, then yes, Linux would become as easy to use from the driver standpoint as the Microsoft products. We lag behind in general because a lot of these things have to be reverse-engineered, and while this works, it makes things much tougher.
All that said, you didn't mention Ubuntu, and for that ... shame on you!!
Casey J. Peter
P.S. The gnome interface is easy to understand. Easier than Apple's.
--------------------
I have a gripe with what you claim and suggest. Essentially you propose robbing users from choice and diversity. Monocolutre is what GNU/Linux is here to address/tackle. Isn't that what SLED is for? Corporate uniformity?
Why eliminate all others as contenders? And why spread FUD about compiling packages when there are such huge Ubuntu repositories?
Roy S. Schestowitz
I read your article and I couldn't agree more. I am a dedicated Linux user. Although I dual-boot XP on one machine, I rarely use it anymore.
I believe that Linux, in one flavor or another, is ideal for the total beginner, and for the total geek. Where it misses the boat is with the large segment of the market that resides in the middle.
I have set up several basic Ubuntu/Kubuntu boxes for friends and family who only use their computers to send e-mail, surf a bit, and share pictures. They love the fact that they don't have to worry about viruses and the like. My father-in-law never owned a PC until he retired. He uses Linux and loves it, probably because he had no preconceived ideas about operating systems.
I have also introduced Linux to college students (I teach Electronics Engineering Technology). They love it for the same reasons you and I do.
In fairness to the Linux distributions, they are at a disadvantage when it comes to the integration of hardware. Microsoft became so dominant so early in the development of the personal computer that the hardware manufacturers themselves did the bulk of the work (and bore most of the expense) to make the integration of their hardware seamless. It was either work with Windows or die. The nature of open-source software precludes that sort of market influence. I believe it is unfair to judge Linux OSes by the MS standard, since they never had (and never will have) control of the hardware manufacturers. Slow and steady wins this race, I think.
Th. E. Samuel
--------------------
I am a user of both Windows and Linux. The bottom line is that it simply depends on the job I am doing. For most server applications I stick with Linux. Why? It's more reliable than Windows, costs less (I don't care what any Gartner study says. If your administrators are literate in both then Linux is cheaper), and I can get to the source code if need be to make some "tweaks." For the desktop I have been mostly staying with Windows. The key reasons are: 1. My company runs on Windows so for company business, I use my company-provided Windows laptop.
2. Visio. Give me a viable Linux alternative and I will drop Microsoft Office.
3. Application Development. I like the Microsoft Visual Studio.Net 2005 development platform. Until I am able to take the time and learn C/C++ I'm sort of stuck. Plus, most of the customers out there have Windows desktops so my development doesn't suffer. I do see a point in the near future when I am going to have to port to Linux. I am tempted to migrate to Java (which I have some experience with).
4. Applications. Many are only available in Windows flavor. For Linux to take over the desktop, it requires end user training, training, training, training. ... Did I emphasize it enough? People hate change. I know because I implement telecommunication systems. No matter how much you train, if there is something different, the end user won't like it. The same goes for the desktop. Linux vendors need to put great emphasis on getting into elementary and middle schools. Raise someone on Linux and you'll have them for life. Support requires trained technical staffing and Linux still appears to be in limited supply. In the corporate environment, Linux is viable. Most deployments are controlled by the IT department. The end user doesn't need to know about drivers, etc. Unfortunately today's IT departments are either overstretched, or just lazy. Linux has been out there for several years along with various flavors of Unix yet most IT departments don't have the skill set needed to work with Linux. Many companies just won't spend the dollars on training.
In my office I have nine computers with the following operating systems loaded:
(2) Windows XP Home Edition (1) Windows XP Pro (1) Windows 2003 SBS (1) Solaris 8 (2) Linux (Red Hat) One of my Linux boxes is used as a server and the other as a desktop. As Windows moves forward, this entire train of thought may become completely irrelevant. Windows is becoming more Linux-like. It must, to reap the benefits from open-source development. I think Bill Gates learned his lesson with Kerberos. Interoperability is no longer optional but mandatory. Look at the uphill battle Microsoft has endured with the state of Massachusetts and the open document standard that Microsoft has refused to support.
James Middleton, MCSE/ACE
Xeta Technologies
NEXT: The Response
--------------------
Many of the readers bring up some good points. But, I think some clarification on my column may be in order. For the most part, my comments were aimed at the commercial distributions of Linux, those which the channel must pay for to resell. The vendors behind those distributions want to become larger players in the business desktop market and must act accordingly, if they mean to succeed.
As for the open-source distributions, those are wonderful experiments in user freedom and those users looking for flexibility and control should seek out one of those distributions to meet their needs. Here, Ubuntu could prove to the best starting point.
The corporate desktop is another story altogether. It is a place were uniformity and management are needed to bring costs down and provide IT staffers (and channel integrators) with the ability to deploy, manage and support end users.
The vendors behind the commercial releases of Linux need to remember that fact and act accordingly. Luckily, with encouragement from the channel, the purveyors of commercial Linux distributions will get it right, but time is running out and Microsoft's Vista is right around the corner.
As for the hardware driver issues, it will be up to the channel to put pressure on those vendors not supporting Linux. That may take some calls to your hardware vendor's channel reps and liaisons.
Frank J. Ohlhorst
director
CRN Labs