Bad business or tough luck in the Open Source sector?

For the last week or two, Open Source religionists have been
working themselves into a lather about all the mean things Microsoft’s
Craig Mundie said about Linux and Open Source as a business model.
Frankly, I can’t figure out why Mundie’s masters even bothered to trot him out
to trash the competition. Over the last few days, Linux businesses made
most of Mundie’s points for him. If there was ever was a bottom to the Linux
market, we may have finally reached it. Is this just a temporary
setback or is there something fundamentally wrong with GPL and the Linux
development model?From a corporate standpoint, the past week has been as bad as any
week since the air first went out of the Internet and Linux stock bubble.

Rumors have been flying about layoffs at a prominent Linux
software house.

Hardware and server specialist (and our parent company) VA Linux,
reported a thundering net loss of $19 million, and reported a major
reduction in sales. Dot-com revenues shrunk by 40 percent. More worrisome,
corporate “enterprise” sales fell by over two-thirds. The company had been
warning for weeks about weak numbers and was expected to take big
write-downs. In fact, First Call listed VA Linux financials as a “positive surprise”
because market expectations had become so dismal, and the stock price
actually rose.

At MandrakeSoft, shaky numbers and an internal squabble over
strategic direction led to the exit of 25 of the company’s 150 employees, including
its CEO, Henri Poole, and its chief technology officer. The company’s financial viability and the fate of its MandrakeExpert and MandrakeCampus e-learning products has been thrown up in the air, just as the company has begun to prepare for a future IPO.

Over at Corel, no news continues to be bad news as negotiations over
the sale of its Linux business to Linux General Partners continues to
drag on and on. The value of the deal has been dropping by a half million
dollars a month.

Ugly, ugly ugly.

Unfortunately, the worst news was the collapse of the technically
promising if financially dubious enterprise called Eazel. Eazel had dared to
attempt what seemed like the hardest task in Open Source — to redefine the
office desktop. Eazel aspired to be the foundation of a fully realized
platform for integrating the Net and the desktop, supporting a product/service
mix that would attract independent software developers and revitalize
innovation on the corporate desktop. At least on paper, Eazel looked great. The
mockups and schematics were elegant. The integration of external and internal
resources looked superior to anything offered by Microsoft or Apple. The
core file system, Nautilus, actually worked.

Eazel had everything but money. Unfortunately the company proved
that building a corporate desktop was expensive. The company quickly burned
through $11 million of seed capital, and then starved to death. Despite
initial backing from Sun and Dell and a host of others, Eazel died
quietly this past week, because none of the company’s “supporters” were willing
to invest in its future.

The reason for this lack of support is simple. For all the scorching
rhetoric coming from the Open Source community defending the viability
of Open Source and GPL as a development platform, when it came to
spending money, investors and corporate CEOs agreed with Craig Mundie — even
CEOs “committed” to Linux.

In some ways, Eazel represented the toughest financial proposition
in Open Source. It was going to be very expensive to build and very
inexpensive to distribute over thousands of desktops. One reason that Linux has
done better in the server market than elsewhere is that server software is
relatively inexpensive to develop and server farms are somewhat difficult to set
up. This means that costs can be contained and there is a visible, if thin,
revenue stream to be had from systems integration and support services. By
contrast, Eazel looked like an expensive proposition to be supported by virtually
no visible revenue. It was a great idea and a bad business.

Despite that fact, industry leaders should have tried to figure out
a way to keep the Eazel project alive. They will regret its passing.

Microsoft correctly believes that the key to its success in the
future is retaining ownership of the desktop. It plans to offer a rich variety
of services for handheld devices that are effectively extensions of its
desktop offerings. Applications drive the marketplace. Without a
presence on the desktop, it will be hard for independent handheld device
hardware and software developers to resist working with Microsoft. Whatever
Eazel’s direct financial prospects, the company represented a potential
alternative for application software developers and handheld vendors. Without that
alternative, many of these commercial developers may have little choice
but to work with Microsoft.

Unfortunately, this was the one moment when the marketplace was
ready for a Linux desktop. Corporate IT managers have been deeply disturbed
by Microsoft’s transition from a sales to a rental model. Companies worry
that the .Net initiative means they will no longer be able
to control costs by postponing software upgrades during economic
downturns. Ceding control of the software means companies may be faced with
upgrading their hardware to meet Microsoft’s schedule, rather than their own.
Linux currently holds roughly 2% of the desktop market. If there was ever a
moment when Linux could make a run at the desktop market, this was it.

Too bad.

Tech writers such as LinuxPlanet columnist Kevin Reichard have suggested that we
write off Linux on the desktop, more or less permanently. Linux
cheerleaders, including me, have suggested that Open Source and Linux can survive
quite nicely in a market consisting of corporate servers and communications
devices.

There are two things wrong with that assessment. First, it trashes
Linux’s original champions. Linux began as a hobbyist tool and a desktop
alternative. Writing off the desktop means writing off the energy of all those
developers who may not be nearly as interested in writing cool Internet protocol applications.

In addition there’s disturbing evidence that the back office may not
be a safe harbor for Open Source development. While Linux has developed
into a robust platform that outperforms Microsoft server products, the
combined market for Linux and Unix servers actually shrunk 2% over the
last quarter. In addition, Dell, IBM and Sun Microsystems are taking an
increasingly large share of this shrinking market. The good
news is that many of these servers are running Linux. The bad news is
that, with these three companies owning a combined 86% of the Linux/Unix
server market, the Open Source community and its corporate representatives
have found themselves in a very poor position to dictate standards or to
create viable independent businesses.

Is all this just tough luck — or is it a bad business model?