By Joe Barr
Originally published in 1997

linux 1. the most important and influential operating system of the 1990's 2. a hacker's heaven 3. the internet's revenge on Bill Gates

Let me say this right up front: Linux is no threat to replace Windows as the default OS of the masses. I will say, however, that it is a superior OS for many people and for many purposes. By superior I mean in relation to either Win9x at the low end or Windows NT at the high end. And beyond its limited threat to Windows for additional seats, its power, scalability, and cost are constant reminders of the sad reality of an over-hyped, overpriced, over-bloated morass of Microsoft mediocrity.

Who would have dreamed six years ago that the greatest threat to Windows in 1998 would not be coming from Steve Jobs, or Apple, or IBM. But instead from a primitive Unix clone being created by a computer science student in Helsinki, Finland, named Linus Torvalds. But that's exactly what has happened.

Linux began its life in 1990 as a one-man project. Linus Torvalds was learning about operating system design by adding parts, functions and features to Minix, a minimalistic OS designed by Andrew Tanenbaum as a teaching aid. Released in 1987, Minix tapped into the same internet pulsebeat then that Linux does today. But Tanenbaum had no desire to make it a fullfledged anything. That would be left to others.

Within two months of the release of Minix, there were 40,000 people participating its newsgroup. Minix was, in fact, the inspiration for Linus to get his first PC. Until that time he had resisted. There was nothing in the PC world to entice him. It would simply bog him down in the land of MS-DOS: no source code, no multitasking, nothing to inform.

So with his first PC, a four meg 386, Torvalds began hacking and extending Minix. And it wasn't long after Linus mentioned the existence of his creation on the Minix newsgroup that it ended its larval stage as a one-man project. When the source code was made available on an FTP server at the University of Helsinki, Linux was born. Linus remained the father, the creator, but there were many hands touching the code. Soon there were a hundred users, most contributing fixes, comments, and suggestions. That was early in 1992.

Nobody knows for sure how many Linux users there are today, but it is probably somewhere between five and ten million. It's available for free from public download sites around the world. This makes it wildly popular with dweebs and students, especially in developing countries. It's also available for purchase through a growing number of commercial distributions: Slackware, RedHat, Debian, Caldera, and SuSE to name a few. This makes it much more convenient to find and install.

No matter how you get it, the source code and the copyleft notice of the Free Software Foundation should be included. Copyleft is a licensing scheme which protects the free part of free software. Under the terms of the license, you are able to use and modify the source code, but if you distribute the result you must make your modifications available under the same terms. This prevents a common pitfall of code placed in the public domain. Such code can be incorporated into a proprietary product, but the resulting product isn't required to be placed back into the public domain.

Here in Austin, with its high dweeb-density ratio, there are two Linux users groups. The one I belong to meets weekly, drawing between 15 and 30 regulars and a couple of newbies to each meeting. There are similar groups all over the world. I call it my weekly dweeb meeting. It is all that and more.

Actually the group, and this part of the Linux phenomena, is a throwback to an earlier stage of the computer age. In the early days, most users were hobbyists or professionals who really loved computers and computing. Software was written on purpose, not as a means to make money. Quite often the software was shared, freely and openly exchanged. This group reminds very much of that era.

It is the sort of group where, if I walked in one night and announced I had modified and recompiled the kernel, it wouldn't cause much of a stir. And if anyone did bother to ask me why, and I replied that I did it simply because I could, there would be an instant smile of understanding.

The group has some of the most knowledgeable Unix/Linux/operating system gurus in Central Texas attending. Digital professionals, retired IBM'ers, high school students, housewives, security consultants, lawyers, programmers, tech support specialists, all sorts of folk attend.

The main focus is always in assisting newbies through the install process. But more than anything, it is a social event. These are people sharing a communal notion of computing. If it takes a village to raise a child, then it often takes a community to install Linux. These folks, and others like them in cyberspace and around the world, are the village elders. Attend a meeting or two and you will understand completely why InfoWorld gave the prize for 1997 Best Tech Support Award to the Linux user community.

Don't dare look down your nose at the housewives, either. They know their stuff. I'll never forget how my slack my jaw got one evening as I listened to one let go a rant against Netscape. "The damn thing is so un-Unix like," she fumed. "It goes against every important Unix philosophy: it tries to do more than one thing, and its footprint is huge." In the Linux community, and hopefully Netscape and others will learn this, it is going to take more than just not being Microsoft to make a good impression.

As I've done the research for this column, I've been amazed at the rate of change occurring in the Linux world. When I built my web server on a K-5 AMD chip and spare parts a couple of years ago, I started with a Red Hat distribution. Caldera came on the scene with its Network Desktop and I went with that for a year or so. Now the server is back on Red Hat.

I have had no failures of the website due to problems with either Linux or Apache. None. Yes, there was a hard drive crash. ISDN has been problematic at times. But the software platform has been rock solid. This over two years of constant use.

I haven't even installed an X-Window manager on my server. When I first put it together there were too few video cards supported and too many stories of crashes. A GUI would add nothing to my web server anyway, so why bother? Today, I bother.

On my desktop machine I run Windows 95 and the SuSE 5.1 distribution of Linux, using LILO for boot management. With the Applixware office suite from Red Hat, I suddenly have a full-featured word processor and spreadsheet as alternatives to Windows 95 applications. With xIrc, I have a GUI for chatting on IRC. Mail clients and dozens of other applications are springing up like wildflowers. The latest Linux distributions make the once daunting task of X windows configuration much easier. The SuSE distribution adds support for video cards that have never been tamed for Linux before. Linux and gorgeous desktop layouts are a happening kind of thing right now.

What's the biggest barrier to faster expansion of the Linux user base? Installation of the operating system has to rank number one. It's a two- or threefold problem. First of all, if you can't buy machines with the operating system preloaded, you are never going to reach the great mass of users. Microsoft knows this lesson well. It helped killed hopes for viable competition from OS/2 by keeping it from getting preloaded. In fact that action may be the only thing that preserved Microsoft's death-grip monopoly on the desktop.

Bryan Sparks, CEO of Caldera, Inc., charged recently that Microsoft has threatened a major PC maker with the Redmond Death Penalty (loss of its license to preload Windows) if they went ahead with plans to offer Linux preloaded.

According to Sparks, it was one of the top five PC makers in the world. Strangely enough, Microsoft did not flatly deny the charge. Instead they made off the mark replies about a CEO responding to hearsay. One can only assume that Microsoft, as always, will do anything and everything in its power, legal or not, to preserve its deadly embrace of the desktop.

The first addressable problem, then, is the installation itself. Many times it goes as smoothly as silk. Other times, even for experienced Linux users, there are long, often frustrating trial and error sessions before getting a successful install.

A case in point: my desktop installation of SuSE 5.1. I could not for the life of me get Linux to recognize my plain-Jane NE2000 compatible PCI ethernet card. It would see the card during the installation process, but not after booting my configured system.

The problem turned out to be that the tunable parameter I needed to tweak for proper installation was not where I expected it to be. I was going through every possible option under PCI cards that I could find. The argument I needed to change was listed under the ISA ethernet card section. Intuitive it is not.

The books are getting new covers with each release, but as one sharp-eyed dweeb at the meeting Thursday night pointed out, what's between the covers stays pretty much the same, including the mis spellings. What all this means, of course, is that Linux is still dweeb country. That's why it will never replace Windows.

But Linux can take a sizable slice of the pie, and it is. That is what is sending chills down the backs of the gangster geeks in Redmond. As personal computing itself matures, and more and more users move from newbie status to experienced, the baby-boomers of the computer era will find Linux more and more to their liking.

Why? It's fast, small, powerful. It supports SMP (multiple CPU's in the same box) right out of the box. This in spite of what you may have read recently in a Ziff-Davis review. It scales better from one processor to two than Microsoft's premium product, Windows NT. It is, as my experience has shown, incredibly solid. And as the InfoWorld award points out, it has the best technical support in the world.

Further, Microsoft seems to have played out its string in terms of operating system "advances." Windows 9x will die with the release of Win98, whenever that Trojan horse hits the market. The plan then seems to be to move everyone up to Windows NT. That's fine, I suppose, but it's fat and old hat.

So where will Microsoft turn for inspiration? My guess is that the next great "innovation" at Microsoft will be remarkably similar to what Linux is today.