Comdex report: The time is now for Linux

By: J.D. Hildebrand

Abstract: Three days of conversation and study have convinced your humble editor that Linux is here to stay. By J.D. Hildebrand

The time is now for Linux

Three days of conversation and study have convinced your humble editor that Linux is here to stay.

I've spent the past three days exploring the products, companies, and technologies on display at Linux Business Expo, which was held concurrently with Comdex Fall '99 in Las Vegas. Many of the companies at the expo made product announcements; I've tried to detail them in previous dispatches. Tonight, before I pack up to return to Developer News world headquarters, I'd like to share some impressions -- details that you could pick up only by being here.

Sea change

In every single interview I conducted, vendors volunteered that they had noticed a recent shift in customers' attitudes about Linux.

In the early days, virtually all of the interest in the OS came from programmers, hackers who were interested in exploring the operating system for technology's sake. Open source zealots and random gearheads loved Linux.

About a year ago, the tech guys started sneaking Linux into the corporations where they worked. They used 'em as firewalls or servers in limited-scale pilot programs. They compared the operating system's stability and operating costs to Windows NT's. They wrote simple apps and compared distributions.

Then suddenly -- vendors agree that it happened all at once, at August's Linux World conference -- Linux was no longer an experimental technology. It was proven, and corporations were ready to get serious about deploying it.

It's difficult to say anything substantive about the corporate world's moods or attitudes...but every Linux vendor I spoke to noticed the same shift. It wasn't one or two guys pitching me an angle, but a real phenomenon.

If Linux World established Linux as a viable technology for deployment in corporate IT environments, this week's show saw something just as remarkable happen. If there's one impression everyone took away from the show floor, it was this: Linux is a business. Big companies spending big money to promote their products. Big money on development. Big money on exhibit space and glossy brochures. Big IPOs, big competition, big opportunities. Linux may have once been part of a cash-free open source utopia, but no longer. Now it's responsible for spawning a market.

We've been through this before, perhaps you recall. If you didn't start studying Windows programming in 1991 or 1992, you had a lot of catching up to do when Windows became a dominant platform. It's clear that now is the time to start studying Linux. It's for real, it's important, and it's not going away.

That is not to say that Linux will replace Windows, though some vendors were candid in claiming Linux's superiority over Windows NT Server. With the Department of Justice gearing up for a knockout blow against Microsoft, anything could happen to Windows -- it could come back stronger than ever as the sole product of MS Operating Systems Inc. Ignore the press reports; this isn't a war and it isn't a football game. It's business. And the plain fact is that Windows does not have to lose for Linux to win.

Complications

With so much money at stake, the publishers of Linux distributions are in an awkward position. They can't tell if they're members of a collaborative open source family in which all developers share their code for the good of the community as a whole, or they are competitors fighting over who will get which piece of the financial pie. Each vendor has established a different strategy for combining open and proprietary technologies in future releases.

Compatibility, portability, and interoperability are emerging as large concerns. The explosion of Linux distributions leaves software publishers in a quandary because applications that run under Red Hat, for instance, may not run under Caldera. Publishers have to create and maintain distinct binaries for each operating system or just distribute the source and let customers tweak and compile the code for themselves.

The best hope for a solution to this problem comes from the Linux Standards Base initiative. The LSB is a cooperative effort among Linux vendors to define a set of APIs and services that represent a baseline standard for the operating system. If the group is successful, developers will be able to write to the LSB spec, confident that their applications will run on every compliant distribution. And vendors will be free to differentiate their products by adding value above the LSB foundation.

Whether LSB is intended to prevent forking, promote safe forking, or enable renewed enthusiastic forking is a debate best left to open source philosophers. The main thing this important initiative will do is prevent Linux from replacing Unix as the Yugoslavia of operating systems, full of incompatible warring republics that don't even speak the same language.

Most vendors agreed -- with varied mixes of regret and happy anticipation -- that the influx of money, competition, and corporate attention will tend to drive Linux's evolution more toward the closed Windows model and away from its open source roots. Virtually all publishers of Linux distributions also offer proprietary extensions, frameworks, and services that make one question whether any commercially usable Linux distribution will be offered free of charge.

Keep an eye on Shelf

RAD environments like Delphi and Visual Studio are not readily available for Linux. Today's leading development tools are published by Cygnus (now Red Hat), Metrowerks, and Gnu. Applix offers a nice little scripting language -- Shelf -- that works with its Applixware applications, but Shelf is not currently positioned as a general-purpose language. Shelf could be a sleeper, however -- functionally, it's somewhere between Visual Basic 1.0 and 2.0. Applix is working on extensions that could make Shelf a simple development tool that will let anyone be a Linux programmer, just as Visual Basic lowered the bar in the early days of Windows programming. So it may be worth taking a look...at least until the Inprise Kylix team gets the Linux versions of Delphi and C++Builder out the door.

Vendor news

The biggest recent news in the Linux community is undoubtedly the merger of Red Hat with Cygnus. Close behind is the release of Corel's widely anticipated Linux distribution. Reviews have already begun to appear on the net -- mostly favorable, with some minor criticism of installation defaults and configuration details.

I had a cup of coffee with Pick Systems marketing product manager Bud Coyne. Pick is a 30-year-old supplier of database technology and operating systems to value-added resellers, with an active user base of 850 to 900 VARs. Pick offers VARs two routes to Linux.

The first is a native implementation of the D3 database for Linux. This database is widely used as a foundation for vertical-market applications. VARs who would like to deploy applications on Linux are the target market for D3.

The second is D3 Pro Plus, which was originally intended to be a transition product for Pick's existing customer base. Pro Plus is an implementation of the Pick operating environment that happens to use Linux as a host. VARs who use Pro Plus need never interact with Linux; they can write all of their code as if they were running under the Pick operating system they have used for decades.

Once D3 Pro Plus got into the marketplace, Coyne told me, Pick discovered that it was generally useful as an application server appliance. Without the benefit of an IT department, VARs and small businesses could use Pro Plus as a host for custom applications. Pick's FlashConnect add-on transforms D3 applications into Web-based programs.

The benefits of Pick's Linux products, Coyne said, are three. First, they give VARs and developers access to the Pick database as the core of their applications. Second, Pick customers can developer under D3 and deploy in the simpler D3 Pro Plus environment, which does not require that users ever touch the base operating system. Finally, Pick customers have access to a library of more than 4,000 complete vertical-market applications written by other Pick users. With these prepackaged solutions, Coyne said, VARs can expand their reach into new vertical markets without writing a single line of code.

Catching up with TurboLinux

I spent a pleasant hour talking tech and open source philosophy with TurboLinux development VP John Terpstra.

TurboLinux currently offers three Linux core products:

  • TurboLinux Workstation 4.0 is configured as a developers' workstation with a suite of development tools: egcs, GNU C and C++, Objective C, Fortran, Lisp, gdb, ddd, xxgdb, as, id, Strip, Python, Perl, Tcl/Tk, Java, and CVS.
  • TurboLinux Server 4.0 offers a comprehensive set of predefined configuration options. During installation, users can configure the system as a mail, DNS, Web, news, caching, file, print, or ftp server -- in any combination.
  • TurboLinux TurboCluster Server 4.0 adds clustering support to TurboLinux Server, with dynamic load balancing, automatic IP failover, management tools, and support for Linux, NT, and Solaris cluster nodes.

An outspoken Australian, Terpstra told me that the real value of the open source movement is that it spreads the cost of development and maintenance across a large user base while maximizing the opportunity for innovation. Just the opposite happens with proprietary systems like Windows.

Terpstra also offered his view of Linux's chances in the market, along with an assessment of the key challenges it would face. "For the Linux community to survive for long," he said, "the operating system's core source code must be completely open. The platform must be dependably ubiquitous, meaning that applications written for one distribution can reasonably be expected to run on other distributions. And above all, the technology must effect a quantum shift in the way information is used."

That final point is the key, Terpstra said. The PC, the Internet, and the telephone were all the right idea at the right time. None has turned out to be an end in itself -- they're just tools that support a quantum shift in the way people live and work. Linux, Terpstra believes, is on the verge of supporting a similar change.

This is pretty heavy stuff. But with the big booths, the air of enthusiastic optimism, and the high interest from corporate decision-makers, Linux vendors can be excused for thinking big. The Linux boom has officially begun.


Server Response from: ETNASC03