Linux: What's the Big Deal?

If you haven't heard of Linux by now, well then, you just need to remove the propeller from your beanie. Start there. If you don't have a propeller, remove your beanie--you can have it back when you compile your first tarball successfully. If you don't have a beanie then we have too much to talk about to do it here. E-mail me (that's electronic mail ... it's on computers now).

"So what is Linux?" you ask.

It's an operating system.

"Well I know that, but what's the big deal? I mean, I heard it's free ... is that true?" you continue.

Yes. In many cases free, as in "free beer." In all cases free as in "free speech." They're different things, you know?

So what is the big deal? It's a revolution. That's not drama, and it's not sensationalism. It's the truth. It's revolutionary in so many ways it has become a revolution. It has frightened many, confused more, and impressed countless numbers. Linux. A free operating system. An operating system that many swear by on servers, a few swear by on desktops and laptops, and more swear by on anything from a gas pump to a Speak'N'Spell. Okay, maybe not the Speak'N'Spell, but hear this: just because I said that, someone is going to set out to port Linux to a Speak'N'Spell. You'll read about it on Slashdot in 6 weeks. I'll bet you.

Nearly 11 years ago a nerdy kid in Finland decided he was fed up with the anemic operating systems available for the x86 architecture. He used UNIX in his university classes, and decided he wanted to use UNIX at home. This was before Solaris, so he didn't have any means to run UNIX on his x86. Early that summer--the summer before he started university--he bought all the textbooks he would need for his fall classes. Here's where it gets really nerdy: he decided to read one of them to get a head start on the class. The book was Operating Systems, by Andrew S. Tanenbaum. In the book was the source code for an operating system that Tanenbaum had written himself as an object lesson for study. It was based on UNIX This kid, who would not leave his room for several years because of the project he was about to start, decided to write his own operating system. Linus Torvalds was his name. The basis of his work was a simplified, open version of UNIX--and the tools he would use would all be look-alikes from the UNIX set--so he called his operating system Linux.

"How do you pronounce that?" you ask.

I don't know.

I have read a few books on the subject of Linux and seen several competing theories on the correct pronunciation. It's a play on "Linus" and "UNIX." So ... "Lie-knicks?" No, that's the one they unanimously said was wrong. In Swedish (I know, I know, he's from Finland ... but a part of Finland that speaks Swedish) "Linus" is pronounced "Lee-noose." So "Lee-knicks?" Well, almost. For some reason the "nix" part is unanimously "nucks." So "Lee-nucks?" Yes. Also accepted in many places, including the U.S., is "Lynn-nucks."

So Linus wrote his operating system and in the process got some help from people via the seedling of the Internet: Usenet. He asked some questions, answered some questions, and even got into some throwdowns. Finally, after a lot of requests, he put his operating system up for download. A community was born.

To be fair, he had a lot of help. Linus happened upon Richard Stallman (you will see this name regularly in discussions of Linux, and usually referred to as RMS; he lives on a soapbox.), head of the Free Software Foundation. Stallman came just after the age of free computing. In the early days of computing everything was free ("free speech" here, not "free beer"). He and his hacker buddies at the MIT AI lab could ask to see the source to any of the software that came through the university, and the authors were more than happy to oblige. Then in the early '80s, companies began to realize that a great deal of money could be made copyrighting the concepts involved with software authoring. Stallman saw all of his peers hired into proprietary software companies, and MIT switched from the Open Source (before there was such a term) OS to a proprietary OS. Stallman decided that if the computer industry as he knew it would ever survive it would take a free OS, an open OS. Stallman set out to author a UNIX-like OS, because it would be easily portable and powerful. He was a skilled programmer, and came from a time when the source code had been available, so his software was very UNIX-like, but written from scratch.

Luckily for Linus, Stallman started at the other end of the puzzle, and they ended up meeting in the middle. Stallman started by writing core applications. He wrote the GCC--a compiler, the thing that actually "makes" a program--from scratch. Without GCC Linus would not have been able to build the kernel (until he built his own compiler, of course). Stallman wrote Emacs, widely regarded as "light years ahead" of its cumbersome UNIX counterparts, vi and ed. He also wrote BASH, a rewrite of the UNIX command-line. Linus supplied the kernel, and now they had an operating system.

Because of this shared work many people insist that the name of the operating system is GNU/Linux. GNU was the name of Stallman's suite of software, an acronym for "GNU's Not UNIX." Clever, eh? By the way, it's pronounced "Guh-New" no matter what you are told. It was from Stallman, also, that Linus got the licensing rhetoric that he would use for Linux:: the GPL. The GPL dictated that anyone could see the source code for Linux, and anyone could alter it, add to it, and do pretty much whatever he or she wanted with it, with one catch: whatever he or she did to it had to be made open to the public.

Meanwhile, a community grew online, centered around this OS. Linus worked hard at it all through university. He added networking to it when that technology became hot. He was now getting help on the coding from people online. Soon people wrote him to tell him they were running Linux full-time. Next he was told that companies were distributing Linux on floppies as a commercial venture. Under the GPL any new features or bug-fixes developed by this or any other distributor were freely available to the world. Everyone played along. Before Linus knew it there were thousands of people coding for Linux.

The next big step for Linux was the addition of X, which was ported from UNIX. X is a layer within an OS that acts as liaison between a window system and the operating system. In other words, Linux could now have a Graphical User Interface. Two major groups, among others, would go on to develop two successful GUIs for Linux: Gnome (pronounced "Guh-gnome" because of ties to GNU; it stands for "GNU Network Object Model Environment") and KDE (short for "K Desktop Environment"). Both GUIs are very Microsoft Windows-like, but don't tell a hard-core Linux-Geek that. Tell the geek they're "full-featured, robust GUIs."

"Okay," you say, "cute story so far, but what's the big deal?"

Well, what if I told you that this little nerd and his horde of online pals went from basement computer-desk trolls to millionaires in some cases, and that the little OS that could, HAS. What if I told you that in 10 years (Microsoft has been around for about 25 years) Linux has reached over 30% marketshare on servers, has made a dent in workstations, and has a fleet of handheld devices?

"Hmmph," you say, "not bad!"

Indeed.

Linux is based on UNIX, as mentioned, which was the foremost OS for networking. Servers are designed, in many cases, to run UNIX. Linux, however, is Open Source.

"So?" you ask.

I wish you wouldn't interrupt.

So ... that means people working on it a) were motivated by passion and fun, not deadlines, pressure, and panic, and b) had none of the limitations that a company has. For one thing, you were working with some of the most brilliant people in the world when you worked on Linux--from all over the world, not just the people that happen to work with you at your company. Here's an example of why that's important. Right now Microsoft is working on a project called Farsite that looks to rewrite the idea of the operating system from the ground up. The project members plan to write an operating system that, beyond the duties of the traditional operating system, monitors resources and reports to applications so that an entire network can look like one big shared device and so that applications can adapt to changing hardware on the fly.

Trouble (for Microsoft) is, Carnegie Mellon University and IBM are working on the same thing. The difference is, whatever findings come from Carnegie Mellon and IBM's work are shared resources. When the Carnegie boys hit a wall they can take a look freely at the IBM research. When IBM researchers start scratching their Big Blue heads they can offer the pickle to an online forum and get answers from Linus Torvalds, Richard Stallman, Alan Cox, and thousands and thousands of nerds who desire nothing more than to tell all of their friends about how they ingeniously solved the problem.

In the Linux movement, the "marketability," "risk of investment," and "projected duration of development" didn't matter. People used Linux in any way they saw fit. One guy showed Linus a project in which he used Linux to run a gas pump (Just for Fun, p.229)! If he had put that idea out for a focus group at Exxon they would have laughed him out of the building. He had the idea; he made it happen. This is largely how it happens, and the main reason why Linux will run on nearly anything with a circuit. Some guy wants to run Linux on his PowerMac ... he ports it. Now anyone with a PowerMac can run Linux.

Also, Linux is free, as in "free beer." People started realizing that if they could code worth a box of rocks then they could write their own apps on this platform. Further, if they used the resources that Linus was sharing, they could get thousands of people all over the world to help them! Have you ever heard of Apache? It's one of the best and one of the most commonly used Web server packages out there. Know where it got its name? It was written on the Internet, much like Linux, and each change that was made by one of the coders was called a "patch." People started calling it "A Patchy Web Server." The name, with modified spelling, stuck. With tools like Apache, The Gimp (an Open Source Photoshop-killer), and StarOffice (an Open Source MS Office-like suite), people were jumping on the Linux bandwagon. In Asia this year Linux is winning big on workstation installations as companies in China and Korea race to legitimize the software installations on their computers ... to escape the penalties from piracy. Free software--meant both ways.

Developers love Linux. For one thing, you can see the source. What better way to learn to program than to start in an operating system that is open? I don't program, but programmers have described the feeling to me of switching from a closed platform to an open platform. A common theme in these descriptions is "the light turns on." In a closed platform you are given some expensive tools to push things around inside a black box. In an Open Source platform you can see exactly what you are doing, not to mention that many of the tools are free.

On top of all of this are stories of people running insane server configurations on years-old hardware. Think about it: it's Open Source. You go in and remove anything in the kernel that you don't need and customize the OS to suit your needs. Is this a Web server? Fine. Remove everything except that which it needs to run to serve Web pages, and that can include the GUI. Suddenly, your PII 450MHz with decent RAM and hard drive support can server up a Web page pretty darn well. You can get a Linux OS so small that you can boot if from a floppy--I'm not talking about a startup disk, I'm talking about an OS. It's a crippled OS, but it's an OS.

The Open Source movement is so attractive to developers and corporations that want custom-made software that industry giants have been forced to consider Open Source. Apple made the first move, making Darwin (the UNIX core of OS X) Open Source. Basically that gives you access to the kernel of OS X, but the GUI is still completely protected. In a shocking move Microsoft has dipped its toe into the Open Source waters. Last May Microsoft announced the Shared Source Initiative (SSI), a program that opens Microsoft source code to select groups, including academia and key industry partners. The SSI was advanced last week to "System Integrators," select techies from major Microsoft sellers who will be allowed to buy access to the source code.

Servers are where Linux really began to grab some limelight. Huge, expensive server hardware is expensive for a small business--or a home. Slowly but steadily, people have realized that they can do it with a spare desktop, and small businesses have saved millions. There is a new paradigm in servers now: the single processor small business server. Where did that come from? Linux. Even Sun Microsystems has finally given in. The company will release soon a single processor desktop computer that comes with a Linux OS to act as a small business server. Sun Microsystems .. selling Linux. In case you hadn't heard, IBM has made a HUGE investment into Linux ... and has recouped the costs.

So that's the big deal. Do you see the picture? Open Source. Free Software. The chance to learn, the chance to teach, and the chance to do it all without paying the Microsoft, Sun, or Apple tax. The chance for need and interest and passion to dictate where the technology will go next, not men in ridiculously expensive suits sitting at big desks. The chance to work with brilliant minds from all over the world. Think about it: Linux has nearly caught up to Microsoft in 10 years, while Microsoft had a 15 year headstart. How long, do you think, before Linux surpasses Microsoft? Well, let's see: take a few hundred universities, times, oh, three professors each to be conservative, times about three solid geeky students, plus a few thousand hardcore geeks outside the school systems, plus professional programmers from companies like IBM and Red Hat. What do you get? Well, in my, once again very humble, estimation, a bigger machine than Microsoft. I give it three years. That's right, I just gave a prediction. It'll be three years before Linux is widely regarded to have surpassed Microsoft technologically speaking (many would say it already has).