Please feel free to forward this newsletter to as many friends as possible. I have a personal goal of reaching one million readers. I can achieve it with your help.

  the Software View: Linux, open sorcery. (Part III)

Welcome back, gentle readers. For those of you with Web access and a Netscape Navigator browser, please click here:
http://www.softwareview.com/
Scroll down the page and you will notice a link entitled, "Daily view weblog". The daily news page is also known as a "web log". It is en vogue and the fashion of these days to call it that. Click on the link, click "reload" on your browser or clear your browser cache to ensure that you always receive the freshest, hottest daily news concerning Java^TM, Linux®, XML, wireless, and the software industry! The link never changes, but I will be updating the HTML file page behind it every day. Please, do take a gander at it every day.

Also, gentle readers, the Software View is an Associate Internet World Wide Web site of Amazon.com. I'd like to extend my sincere, heartfelt gratitude and thanks for your patronage. I'm offering links to books, et cetera that you can purchase from my web site. I'd greatly appreciate it if you would purchase software industry books from my web site. Help support my newsletter and web site by purchasing items from Amazon.com from my web site. Here is the URL (Uniform Resource Locator):
Click here

Now, dear readers, on with this week's episode of the Software View!

ON THE BANDWAGON

Eugene Devereaux, an engineer with the Boeing Company, says those announcements marked a turning point for Linux supporters. "They gave the credibility that the technical people were looking for," he says. "Even if they didn't buy those applications, what they did was they put another reason on the list for technical people to argue Linux."

The new push for Linux applications was also taking place upon the desktop. Corel Corporation which already offered a Linux version of WordPerfect 7 shipped WordPerfect 8 Personal and Server editions in the summer of 1998 and developed a compatible suite of business applications. The Corel Computer subsidiary also sold development workstations based upon the StrongARM microprocessor that came with Linux pre-installed.

Linux is not just a fad. It's a reliable enterprise operating environment, that users can freely download, modify, and share, that meets the needs of the world's most demanding customers. There are thousands of Linux developers who are committed to making Linux a viable player in the enterprise. They're proud of their collaborative culture.

"When I started looking at the networking problems that I was having with Windows, I thought I couldn't make my development schedules," says Devereaux, who uses Linux to build onboard computer systems and testing tools for the Boeing Company. Devereaux says that by the time management even realizes Linux is running on company workstations, the project is often half over. "So long as it's appealing to the technical workforce," says Devereaux, "I think it's unstoppable."

THE WORLD IS NOT ENOUGH

Torvalds surfaces at various locations around the world to conduct "World Domination 101" seminars attended by the Linux faithful, who are determined to revolutionize the way software is built, licensed, and supported. He says, "We want to take over the world, but we don't have to do it by tomorrow - it's okay to do it by next week, or even next month."

Torvalds' favorite beer is Guinness. His daughters are named Patricia and Daniela, and his middle name is Benedict. Here is Torvalds in his own words ...

"I actually was introduced to programming by my grandfather. He bought a home computer for himself, essentially as a programmable calculator. I was ten or eleven years old at that point. He enlisted me as his helper, even though I wasn't much help. I eventually started to make my own changes.

I decided to write my own operating system really for my own use ... In the home personal computer world, you had the choice of cheap hardware - fairly powerful hardware with the PC - but then, when you wanted to do real work ... the operating system choice was very, very limited indeed. When you got a computer you got Microsoft DOS, when I started, and Windows was just starting to appear. I had already gotten used to the university machines, and they ran on Unix. In the end, I'd been programming for more than half my life. I was the traditional geekish person who easily sat in front of a computer for eight hours per day. I really didn't know what I was getting into. It was hard, but hard in the challenging sense.

I decided to seek feedback via the Internet in a fairly unplanned way. I'd been on the Internet about a year and a half before I made Linux available. It was just the way things were done. In a university, you really get recognition by publishing what you're doing.

Linux reached a point where enough people were using it that it made sense for certain companies to make releases for Linux. In 1998, partly thanks to Netscape's support, it broke through the next barrier to larger houses, like Oracle. It's a difficult time for any operating system - at the point where it's got enough people who are looking at it, but it's not popular enough for general purpose applications. What has been interesting in the last year and a half is that, with non-technical people starting to look into Linux, there has been more of a market for nontraditional Unix applications. Some of them have started to appear, but I think it's clear that there's a need for more Office personal productivity and small-office applications.

I think Linux has a different enough strategy that it doesn't compete with Microsoft on the same terms. With Microsoft having a majority market share, it's really hard to compete on Microsoft's terms. And that's why OS/2 and the Mac OS, I think, won't survive. Linux is competing in a different kind of marketplace. How does Microsoft compete against Linux? What can they do? They can try to spread fear, uncertainty, and doubt about Linux. But they can't undersell Linux.

I've never had any feelings that I wanted to do software applications. The operating system is kind of like the heart of the system. It's just a very fascinating part of the system. The operating system is like the laws and police force. In a good country, you're not supposed to care because you don't notice them. You're not supposed to see the OS, a and a lot of people aren't really aware of it. It's the underbelly of the civilized world.

Right now, Linux is an alternative, but it's mainly an alternative for people who have specific needs. I certainly hope Linux will be more of an alternative for normal people, too. It's not a technical issue. It's more an issue of perception. It just makes more sense for normal people to go with the flow. Hopefully, going with the flow will no longer mean Linux isn't an option."

EVERYBODY LOVES RAYMOND

Eric S. Raymond, an early proponent of open-source software development, has become an ambassador of sorts for the acceptance of Linux in corporate environments. When asked why he has taken on this mission, he dispenses with marketing niceties. "Because closed software sucks," says Raymond. "It's painful knowing that there are lots of people out there who are used to an astonishingly low level of quality and reliability of software and furthermore have been brainwashed into accepting this as normal. Open development means lower prices and better products."

Raymond, an editor of The New Hacker's Dictionary, authored the influential essay that prompted Netscape to release the source code to its Communicator browser earlier in 1998. Raymond's essay, "The Cathedral and the Bazaar", argues that most commercial software is built like cathedrals by small groups of artisans working in isolation. Open-source software, like Linux, is developed collectively over the Internet, which serves as an electronic bazaar for innovative ideas. "It's subversive," says Raymond of open-source software, "because it takes all of the thirty-year verities that we understand about software engineering and stands them upon their head.

The shot heard 'round the world in a quiet software revolution was the release of the Netscape Navigator Internet browser's source code at midnight on April Fools' Day of 1998. This brought to public attention a face-off between two dramatically different and fundamentally opposed styles of software development - a confrontation that had been in the making for thirty years but became inevitable after the advent of the World Wide Web and the explosion of the Internet's popularity in 1993 and 1994.

The first of the two programming styles is closed source - the traditional factory-production model of proprietary software, in which customers get a sealed block of computer binary bits that they cannot examine, modify, or evolve. Microsoft is the most famous practitioner of this approach. The other style is open-source, the Internet engineering tradition in which software source code is generally available for inspection, independent peer review, and rapid evolution. The standard-bearer of this approach is the Linux operating system environment.

The open-source model threatens to make closed-source software companies obsolete. To understand why, we need to step back from the particularities of Microsoft and Linux and consider that engineers and some high-technology executives are attracted to open-source development for three reasons: reliability, reduced total cost of ownership, and improved strategic business risk.

Historically, the way engineers and scientists have ensured high reliability of their products is through rigorous peer review. Physicists don't hide their experimental plans from each other; instead, they skeptically check each other's work. Engineers don't build dams or suspension bridges without having other engineers independent of the original design group vet the blueprints.

Software companies don't generally do peer review of their code. And in the software industry, reliability has always been terrible. Crashes, hangs, and lost data are still commonplace. These observations may seem unrelated, but the connection is clear when you look at the infrastructure of the Internet.

Most of the software that runs the Internet - like the Apache Web server, Perl programming language, and Sendmail e-mail system - was developed under the open-source model, and its reliability is extremely high. This level of reliability is even more significant given the fact that the Internet is multi-platform, heterogeneous, and international and has remained backward compatible through thirty years and several generations of technology.

A simple and compelling pattern is starting to emerge. Open-source software goes through rigorous peer review and has great reliability. Without peer review, software reliability suffers. This fact, in and of itself, will be sufficient to marginalize closed-source commercial software development.

Total cost of ownership is also drastically affected by open-source software. In a closed-source world, a producer charges for the bits and has an effective monopoly on service for its products. Accordingly, major closed-source packages cost thousands of dollars up front plus thousands of dollars a year in service and upgrade costs. In the open-source world, by contrast, the bits are free, the source code is available, and the provider doesn't have a lock on the service and upgrade business. Accordingly, both up-front and annual costs are low.

But the most important long-term effect of open-source software will be on strategic business risk. To see why, think again about the supplier's monopoly with closed-source software. The chief information officer of a Fortune 500 company typically spends millions of dollars on a strategic business system with software that no one in the company can modify. And that CIO must depend upon a single vendor to service the new system. Therefore, changes to those systems serve only the business plans of the software vendors rather than the companies who buy from them.

When companies use Windows operating systems, Microsoft is the only choice they have for service because only Microsoft has access to the Windows source code. If a company chooses Linux, Red Hat Software has to do its service job well or watch its business go to Caldera or the German software vendor, S.u.S.E. And if Red Hat, Caldera, and S.u.S.E. fail, tens of thousands of Internet developers will cheerfully launch new releases using the same free, common and exhaustively debugged software code base.

Open-source software is the best recipe for high reliability, dramatically lowers total cost of ownership, and effectively puts the software customer in the driver's seat. With these factors at work, it may not be long before buying closed-source software is considered fiduciary irresponsibility."

Raymond says that open-source software, created in what anthropologists call a "gift culture", is better at producing high quality software because status is gained by giving ideas away. Companies that value secrecy miss opportunities to get wealthier by sharing ideas and creating information pools. "That's a pragmatic statement," says Raymond. "Not an ideological one."

MAKING CENTS

Open-source software presents a host of new business opportunities, some more lucrative than others. Businesses will be motivated to embrace open-source development for a variety of reasons, ranging from the straightforward - to create new products and to continue supporting older releases - to the more indirect, like creating goodwill in the developer community that will make it easier for the company to recruit third-party developers and solutions integrators.

Raymond offers several business models for making money with open-source software like Linux. Revenues can be derived from open-source software. First, fees for the professional support of open-source software products. They include Linux distributors such as Red Hat Software and Cygnus Solutions, which essentially give away their software, but sell the distribution mechanism - Linux upon CD-ROM's - in addition to technical services and support. "Based upon the existing size of the Perl market, we'll have a $30 million annual run rate in two to three years," predicts Dick Hardt, Chief Executive Officer of Vancouver-based ActiveState, which is using the same model. "Our business model isn't that different from large software companies. The question is, 'Where is the expertise? In the code itself, or in the customer service and accompanying value you provide?'"

The second type of revenue stream is an indirect one, in which some companies will use open-source software as a loss leader and a market maker, giving away software in order to position themselves better for sales of their closed software. IBM, Netscape, and other software giants have gained substantial marketing benefits from this approach.

Then there's the "widget frosting" revenue strategy, where hardware companies go to open-source to acquire better drivers and free interface tools faster and for a smaller investment. The best example of this model is Corel Computer (a division of the software developer Corel), which in May of 1998 introduced its Netwinder line of Linux-based network computers. The company's executives say that licensing a competitive operating environment would have required $500,000 and six months to a year of development time.

Other companies experimenting with this strategy can make money selling complete systems with Linux pre-installed - a path pursued by younger companies such as VA Linux Research Incorporated in Mountain View, California, whose revenues tripled in 1998. This company designs hardware boxes specifically for Linux. Their VarServer line of servers and VarStation workstations come in a variety of configurations. But if none of those models suit you, the company offers custom systems with your choice of a central processing unit (CPU), cache, memory, and a Linux distribution.

The greatest opportunity for an open-source software business model may come from even less direct plays - companies that specialize in essential accessories like manuals stand to gain substantial profits from the demand for open-source software materials. O'Reilly & Associates, the leading publisher of technical resource books, has sold more than $94 million in open-source-related materials since it was founded, most of that since the year of 1997.

Raymond says that the open-source software business model also works well for software developers who can lower their overhead expenses and get access to more brains than most development shops can afford. He adds that there's an advantage to being the first open-source software developer in a given market niche because the pool of talent is finite. The first independent software vendors off the launch pad will attract the best co-developers and they can even use their customers' information technology staffs to help with software development. The pioneers of open-source business models are deriving some success from being early, including attracting the most talented developers. This development model seems to be a fruitful way for existing software companies to create products while continuing to derive revenues from more traditional sources. Open-source development will provide considerable benefits to those companies that tap into the valuable developer resources it provides.

At a recent Usenix Unix conference, an attendee asked Raymond for his criteria of success: "When everyone uses Linux or when we all drive Ferraris?" Raymond says he'll have won the battle when all software programs are open for inspection and modification and when no one is ever frustrated again because they can't fix broken software.

Raymond says official maintainer groups release numbered versions of Linux distributions. When a development team receives input, they freeze new features for that version. He says this type of controlled - yet constant - updating is superior to proprietary closed software vendors like Microsoft where the cost of the product doesn't include its ongoing evolution. "They don't have the money to support development after the sale and its reliability is crap," says Raymond. "It explains the current world." Here is Raymond in his own words ...

"The basic win for developers and programmers is the same: reliability, reliability, and reliability. The peer review aspect of the open-source method simply produces a superior product.

The additional potential for adding value through responsive customization to customers should be clear. It's also important that open-source code gives the Internet Service Provider (ISP) control over his own destiny. No longer are you at the mercy of a monopoly vendor's choices; you've got the source code, and no one can take that away from you. And whether or not you ever actually look at the source, this fact profoundly changes the marketing dynamic in your favor. It means lots of people will be competing to serve you.

I used to have beliefs about places where open-source wouldn't work. They've consistently turned out to be false, so I don't make predictions like that any more.

You'd be amazed at what people find interesting. If you get a large enough developer population, all the bases get covered; even stuff you or I would think is mind-numbingly boring."

Devereaux, the engineer at the Boeing Company, says he is more concerned about the endless Microsoft Windows update cycle and Word version 2.0 files that nobody can read. He adds that, in practical terms, he is able to archive the Linux toolsets and source code he uses to develop computer systems for specific airplanes. "I can never quit selling that airplane," says Devereaux. "But I can always resurrect that version and test it the same way, and I don't see a way to do that with other operating environments."

Randy Kessell, manager of technical analysis for a Southwestern Bell operation center, notes that because Linux allows his company to do more remote network administration and software loads than was possible with either Microsoft or NetWare products, it has driven down their network management costs. He adds that his company replaced their Windows network server with Linux. "Our tests show that the Linux solution was easily outrunning the Windows solution," says Kessell. "It's much faster."

BORN TO RUN

Open-source software offers flexibility. You're using a wide range of multi-vendor components to construct your systems. Linux overthrows Windows by selling its reliability. Ben Wittard, who is using Linux to run ninety print servers worldwide for Cisco Systems, says his managers are beginning to notice that his system has had no downtime for years. "Linux is an incredibly robust, very usable operating environment," says Wittard. "It was incredibly cheap and very effective for our needs."

It's not unheard of for a Linux server to go months or even years without requiring a reboot. This kind of reliability is the main reason the operating system has been widely used as an Internet server, whether for caching or Web services.

For cases in point, look to the relative stability of open-source workhorses like TCP/IP, DNS, and Perl, which keep the Web running more smoothly than any commercial equivalent. Eighty percent of all Internet e-mail is routed by an open-source software package called Sendmail. As more businesses start to depend upon the Web, access to software source code becomes more significant to programmers and Web site operators.

As the complexity of software increases, open-source software allows the number of people analyzing these programs to increase and scale proportionately. As proprietary software programs like Microsoft's get bigger, the relatively small number of in-house and beta testers means that more bugs will slip through. In the open-source software model, says Raymond, the large number of co-developers means there's a greater chance of catching problems, even in complex systems. With enough eyeballs, he says, all bugs are shallow. Boeing's Devereaux notes, for example, that if he has a problem with the way an optimization works under a compiler, sending the problem to Microsoft may or may not elicit a response. "But if I put it out as a problem report onto GNU or a newsgroup," he says, "I could have a fix to that within days."

Kessell of Southwestern Bell says reliability problems he had with Windows drove him to try Linux. His network monitoring group tried running PC's with Windows 95 and Windows 98, but the GUI for their monitoring tools kept locking up. "Because it was failing, our users never used the graphical interface which we had paid outrageous amounts of money for," says Kessell. "Linux not only runs the application well," he says, "but it's more secure." When Southwestern Bell's security team runs their intrusion tests on Linux, Kessell says he knows the versions they are testing are up-to-date. "Not just with one company, but the rest of the world."

Devereaux notes that the Boeing managers were impressed by how fast fixes are generated for Linux. "Now it's turning up in so many places, and it's being used in so many projects, they don't know how to get their hands around it. Who's really controlling this?" Linux has proven more secure because Boeing gets patches and fixes months before they get them from IBM or Hewlett-Packard. "And we can read though the source code," he says.

To be continued ...

Sincerely,
Mark Kuharich

Join my free e-mail newsletter called the Software View by clicking here or by sending an e-mail to thesoftwareview-owner@west-point.org

Linux is a registered trademark of Linus Torvalds