Linux 2000

by Phil Hughes

I see Linux as a progressive movement as well as an operating system. And, with any movement you need to chart your direction. To help with that charting I decided that rather than write the April, 1994 editorial I would just go ahead and write the one for January, 2000. So, here it is...

In the past 7 years we have seen Linux go from an idea for a small Unix-like system into a movement to bring affordable, reliable multi-tasking software to anyone who could buy a rather minimal computer. In fact, we have seen that in some parts of the world people are more likely to have a Linux system than be connected to the electric power grid.

Now we see Linux and an Internet connection in over 100 million homes worldwide. How did this happen? Cost is the best answer. Some of you probably remember an old program loader called MS-DOS. Back in the 1980s it was being marketed as an operating system and it managed to establish a user base close to that of Linux today. But it had three fatal flaws:

  • it only ran on one type of computer system and could be expanded to support the full capabilities of new microprocessors

  • it cost money

  • it didn't support multi-tasking

We can excuse the first flaw as it was originally written for a project at a computer company and was never intended to be marketed to the general public. Although a more visionary company might have made a better decision we can only say that hindsight is 20-20.

The fact that people actually had to pay for a copy of MS-DOS (or, more properly stated, were supposed to pay for it) could also be considered as a very short-sighted decision on the part of Microsoft. As we all know now it is the added value-training, customization and, of course, the user-specific applications-that make the money. Giving away operating systems helps to sell these services along with hardware.

The final flaw, however, is what resulted in the demise of MS-DOS. I remember that back in 1986 I gave a talk at a personal computer user's group meeting in Seattle. I had brought along an IBM-AT (remember those-it had an Intel 80286 processor in it and people ran MS-DOS on them) and a couple of H19 terminals (Heath kit? Another flash from the past for some). I had a version of the Unix system running on this hardware.

I talked about Unix systems pointing out multi-tasking as being a primary benefit. I was amazed when these allegedly computer-literate people didn't understand why multi-tasking was absolutely necessary. In fact, one of the group members actually said “I don't like Unix because it accesses the disk when I'm not doing anything”. Today, 99% of the computer system users don't even know what a disk is, much less a disk access.

With the advent if ISDN in the early 1990s and personal satellite stations in the late 1990s, connectivity became the big issue. People quickly realized that they didn't want to know what their computer was doing, they just wanted to see the results. Could you, for example, imagine manually instructing your computer to call up another computer? Well, Unix systems pretty much pioneered the initial ideas behind these sorts of computer links with the advent of the uucp program suite back in the early 1970s.

When the average user, without using these words, asked for a multi-tasking computer system, Linux was there and waiting. We have to give credit to early Linux activists (and Linux Journal itself) for going out to companies that intended to market personal Internet stations and point out that Linux was a more capable and less expensive base to use for their products. The result, as you can see today, is that most personal Internet stations are based on the Linux operating system.

But there is more to the success of Linux.

People recognized they would rather pay for service than “things”. Linux, much like my first car, a `55 Chevy, offers a choice for the consumer. They can either fix it themselves or then can hire someone to fix it. That someone can be a representative of the manufacturer or the kid down the street. This was certainly not the case with proprietary operating systems or vehicles of the 1990s.

How is Linux Journal Doing?

We are pleased to announce that as of November, 1999 90% of our subscribers are now via the Internet rather than on paper. We do, however, see that other 10% as so important because that is where those who are new to computing (yes, there still are some) find out about Linux and how easy it is to get their Linux system on the Internet. Over the years, most of our subscribers have moved from paper copies of LJ to an Internet subscription once we got them up to speed.

To make this electronic version possible we had to upgrade our offices to a complete Linux network. Even though all of our editorial and advertising work was done on Linux systems from our humble beginnings in 1993, our production, subscription and accounting systems ran on other computers. Today our seamless ISDN connections (and the satellite link to my office outside of Yaak, Montana) make this startup seem like a nightmare rather than the reality of seven years ago.

What Else Make Linux a Success?

These two events were a significant help in getting Linux on its way.

  • formation of MoAmI Semiconductor from Motorola, AMD and Intel engineers in 1994. Linux was first operating system to run on their 32-bit and 64-bit chips in native mode in 1996.

  • Linux became the most popular operating system used in computer science classes in 1995. This meant that the pool of available talent in the Linux market was huge.

But we also need to consider the how events in the Unix community helped Linux. When Novell decided to buy USL in 1993 it was seen as a move to get their own product out there instead of following each Microsoft move. As we now know, this worked and NT (remember, it was going to take over the world) became the niche operating system. Likewise, the decision of many fence-sitting vendors to go with Linux gave it the needed push that caused it to become a mainstream system.

This could be considered a political decision. Unix, although open if you had an extra $100,000 for the purchase of source code meant that its openness was restricted to existing companies. Linux, with its $0 to get started characteristics made it possible for creative talent to get into the computer business-much like in the days of the Altair and the Apple II. The big difference is that the hardware was generic and inexpensive in 1994 so the creative work went into software.

This creativity made the larger vendors realize that they needed to stick with hardware and support as profit centers. Going to Linux as their operating system both reduced their software development costs and made it easier for them to find pre-trained systems programmers for the software work they still needed to perform.

Is there one specific application that I see as making everyone want a Linux-based computer system? Yes. I would say it is the availability of telephone directory information on-line. The fact that it is free and much easier to use than a traditional phone book has caused more people to elect for an Internet connection and Linux which offers a free reader for this information that a 5-year-old can easily use.

What happened to those old Linux activists and developers?

Strange as it may seem, most are still writing code or books. We don't see any who are CEOs or multi-billion dollar corporations. There are successes and many who were involved in the early days of Linux are making a very comfortable living but it seems that these people chose a career path based on their interests rather than attempting to become rich, famous or powerful.

Many of those who were minor players in Linux development or just got involved because Linux was so popular in colleges are now independent consultants. And they credit this happening with the Linux philosophy. The fact that they could get source code and learn about real software while in college gave them the necessary skills to more directly into the work of their choice.

What's Left to do?

We need to offer Internet connectivity to everyone. In the 1930s we started a rural electrification project that was to bring on-grid electricity to everyone in the United States. In the 1950s television broadcasting was seen as a way to get information to everyone. Both of these efforts had benefits but they also had an associated cost. They encouraged people to go to their own separate spaces, interact less with other people and passively consumei—consume both products and consume information.

In 1990 people were much less likely to know the names of their neighbors or world leaders than the names of fictitious characters on TV shows. Although Internet connectivity may not help people get to know their physical neighbors it does help them build a community of electronic neighbors. Using the Internet is active, not passive. Whether people elect to do research or electronically talk to another person they are now making real choices and possibly talking to real people.

Because Linux was so significant in getting tens of millions of people connected to the Internet in the past five years and Linux machines make up the majority of the machines connected to the Internet today, I see this as a project that the Linux community should take on. In 1993 and 1994 we were all out there telling people about linux. If today we all walked next door, introduced ourselves to those neighbors that have lived there since 1990 and then offered to help them get connected to the Internet we could claim another huge victory for the Linux progressive movement before year's end.

As always, send me e-mail at phil@ssc.com and tell me what you think. And mention LJ to your neighbor while your at it.

Load Disqus comments