|[Home] [Credit Search] [Category Browser] [Staff Roll Call]||The LINUX.COM Article Archive|
|Originally Published: Tuesday, 18 May 1999||Author: Jeff Alami|
|Published to: featured_articles/Featured Articles||Page: 1/1 - [Std View]|
In the commercial world, where Windows reigns supreme on corporate desktops, many FUD (Fear, Uncertainty, and Doubt) mongers like to discuss the weaknesses of the Linux operating system. This article isn't about FUD. All Linux users know that Linux has its weaknesses, and the community's goal is to reduce those weaknesses. It's not enough to compare Linux to Windows or other OS's. Linux must be good on its own, and we must compare it to itself....
Many of the actual weaknesses of Linux are exaggerated in FUD articles, but they still exist to some extent. The lack of end-user applications, user friendly interfaces, and incompatibilities between distributions are among those weaknesses. These weaknesses aren't severe enough to make Linux users look elsewhere, but they are nonetheless important issues to address. After witnessing many of the Linux community's accomplishments, I'm convinced that there is nothing that the community cannot achieve.
"I'll use Linux when I can run Microsoft Office on it"
I believe that Linux's biggest weaknesses is its lack of end-user applications. This weakness is most pronounced in the realm of financial software, games, and engineering tools. Fortunately for Linux users, this problem has the easiest solution. As more people use Linux, more Linux developers will create the necessary applications. Companies that currently provide products solely for Windows will definitely port to Linux when they see a large enough demand. After all, it's not normal for a for-profit corporation to overlook potential markets, especially if they're large enough to be worthwhile.
One problem with applications for Linux is that Windows users are so ingrained in Windows productivity applications such as Microsoft Office or Quicken that they will not give other applications a try. Sure, KOffice and GnuCash don't match the features of their Windows counterparts, but they will generally provide the necessary functionality for average users. And it's only a matter of time before the Linux applications (especially the Free Software offerings) match and even surpass the functionality of the popular Windows applications.
It's comforting to know that Linux's biggest problem has an evident solution. Of course, there will always be the need for emulators. Applications such as Bochs, VMware, and WINE offer the ability to run Windows applications on the Linux platform. By the time these emulators will be fully complete, the lack of Linux applications will be a moot point.
Hard to Install and Use
"Of course Linux is hard to use. After all, it isn't Windows."
I've heard this complaint many times. Any operating system is easy to use if you know how to use it. I admit that the learning curve for Linux is more steep than that for Windows, but learning Linux is more of an exercise in open-mindedness that anything else.
First of all, Linux is hard to use because all operating systems are hard to use. Many Windows users had their operating systems pre-installed on their systems. If they had to install the operating systems themselves, they probably wouldn't complain so much about installing Linux. However, we must not stop at that. If we want Linux to be used by more people, we must make it even easier to install than Windows would be.
A prime example of these solutions is Caldera OpenLinux 2.2. With a Windows-based installation starting point, Caldera's distribution recognised that many Linux installations would be carried out on Windows systems. Not only that, it takes into account the fact that most Linux systems are dual-boot. Other enhancements such as hardware autodetection and a game of Tetris during the extraction of packages make the install process as painless as possible.
Desktop environments such as GNOME and KDE are paving the way toward an easy-to-use Linux. The future challenge for these desktop environments is to become as user friendly as possible without becoming a burden to the productivity of power users. A user interface that's easy to learn yet hard to use has failed in its goals. However, the advantage of Linux is that having a particular desktop environment is not a requirement.
Incompatibilities Between Distributions
"Linux will become splintered just like UNIX did."
Because Linux isn't under the stranglehold of a single company, a wide diversity between Linux offerings is available in the various distributions. Unfortunately, this diversity can pose a problem of incompatibility between Linux distributions, with respect to interfaces and system libraries. As Linux becomes more widespread, more and different distributions are released, making the potential issue more pressing.
The solution is a set of standards. However, the bad thing about standards is that there are many from which to choose. Additionally, any standard must be careful to avoid stifling the innovation of the community. Projects such as the Linux Standard Base are working toward some form of standard between Linux distributions, especially in the system libraries and functions.
UNIX was developed under standards, and why are various UNIX flavours so incompatible? The problem is twofold. First, various UNIX systems use different processor architectures: for example, IRIX uses MIPS, HP-UX uses PA-RISC, Solaris uses SPARC, Tru64Unix uses Alpha, SCO uses Intel x86, and AIX uses PowerPC. While Linux supports all these architectures in some form (PA-RISC in the form of MkLinux/PA-RISC), Linux is most popular on the Intel x86. Second, extensions provided by commercial UNIX vendors are not Free Software, so it takes much longer for other UNIX flavours to develop those extensions, if at all. The fact that Linux is Free Software makes incompatibility a much smaller problem across distributions as well as processors.
"Linux has no support. Who do I sue if it goes wrong?"
Many Linux beginners have had the fortune to discover the incredible community-based support for Linux. Usenet newsgroups, Web sites, and even IRC channels are dedicated to helping with Linux issues. So why is it that many FUD articles discuss Linux's apparent lack of support?
The truth is, from corporate America's point of view, Linux has no support. Corporations are looking for 24x7 telephone and on-site support from the same company that owns, develops, and sells the product. You're not going to find that with Linux. No one source owns the operating system, so you're not going to find a single source of support. In a way, this difference is good; if you're not satisfied with one support provider, you can use another.
"Linux doesn't have good SMP or SCSI support."
With kernel 2.2, Linux moved further into the enterprise market, with enhanced symmetric multi-processing (SMP) and enterprise optimisations. For some demanding systems, however, the features in 2.2 aren't enough. What Linux needs is the high-end features only found in commercial UNIX systems, such as extremely scalable SMP (up to 16 or even 64 CPU's) and journaling filesystems.
Many of the enterprise tools are already available for Linux, from commercial vendors such as Oracle and SAP. Now, Linux needs to become a powerful enough server to provide mission-critical capability for large companies.
Despite its weaknesses, Linux is an excellent operating system with many more strengths. As a developer's desktop or Web and file server, Linux shines. It is unparalleled in stability, performance, low cost, and user support. But best of all, the Free Software nature of Linux means that these weaknesses' days are numbered.