Author: JT Smith
It’s tradition among journalists trying to avoid working over the holidays to knock out an article that sums up the big events of the past year or that makes bold predictions about
the future. I think that many writers miss the big events even as they write about them, and that most self-appointed pundits stay in business by hoping the public forgot all the stupid things they said in the past.
For the past several years, I’ve dedicated the first column of January to an examination of the stories I should have covered as well as the ones I wrote about and to confess about the stories where I was completely
wrong as well as the places where I got the story right, early.
This past year featured two predictions that were gloriously inaccurate and a few that, unfortunately, look to be more accurate than I would prefer. More troubling, there were a few stories that didn’t get covered as well as they should have.
To start with the worst, I got
pretty excited when ABN Amro, a hard -headed Dutch bank, made an investment
in the Conectiva, the Brazilian Linux shop. The investment was reportedly close to $100 million for 10 percent of the company. The fact that a bank would consider a company that had grossed less than $4 million to be worth a billion dollars was news to me. Perhaps the sky-high valuations placed on several Linux companies during the dot-com mania weren’t so crazy after all? Well, of course they were.
The valuation of Conectiva was crazy, too. Admittedly, the company
has created
some useful tools such as the APT addition to Red Hat Package
Manager. Conectiva has popularized Linux in local press and helped make
a pop star out of
Linux kernel maintainer Marcelo
Tosatti in Brazilian newspapers like O
Estado Sao Paulo.
As I mentioned last February, less developed countries are a hot market for Linux.
Unlike the States, Brazil is a place where you can find a PC with Linux
on a retail store shelf. Conectiva has made inroads into the Brazilian
state and local government markets and established a presence in many
parts of Latin America. However, the marketing effort has flagged recently.
Press releases and newsletters, which were coming out at the rate of two to three a
month earlier in the year have nearly stopped. The company got sidetracked by
a
lengthy political battle with the Brazilian government over whether
Linux rather than Windows should be installed on PCs purchased for
educational institutions. And while it has attracted talented hackers like Tosatti,
the teenage “successor” to Alan Cox on the 2.4 kernel project, it’s
still a smallish, struggling enterprise. What’s it worth? Who knows? As
I pointed out in August, the valuations of highly speculative
growth companies in the United States continue to be driven by self-interested sell-side
analysts, many of whom ought to be hauled before the SEC and beaten with a stick.
In an article that was nearly as bad, last spring I decided to be
more optimistic about Linux as a business proposition. I predicted that the
gains Linux had made in the Web server market meant that the platform
had a real shot to take over local networks in 2001. In a paragraph I
regretted the rest of the year, I
actually quoted Eric Raymond as some sort of authority about the
prospects for Linux in the enterprise server market. Whatever Eric’s attributes,
an acute business sense isn’t one of his strengths.
As it has turned out, the shock troops from Microsoft’s Consulting
Service worked with prospects to wall off outward facing architecture from
inward facing domains, particularly as companies moved away from Windows NT.
MCS also focused on business issues rather than technology. The result was
a growing domination of the internal enterprise architecture by Redmond’s
Windows 2000. That development didn’t get covered sufficiently, and it
should have. I wrote an article in October that touched on the methods MCS used to win
the market, but the bulk of the “Envisioneering” exercises Microsoft
used to win over business executives were run in late 2000 and the beginning
of 2001, and should have been covered then.
I spent a lot of the year covering a couple of trends I thought were
critical to the future of Linux. One was the stampede away from Open
Source by several Linux champions. I’ve covered what I call an evolving “Half Way Covenant” among many corporate developers. They pledge to
support Linux as an open road-bed on which to run their applications, but
most want to keep their own applications closed. I
reported that Caldera has gone down that road. In
an interview with Ly-Huong Pham, she said Turbolinux was, too. They have joined literally hundreds of firms who have ported Windows, and Unix applications to a Linux platform, but who have restricted access to their own code.
The other trend was the price Linux companies continued to pay for
amateurish marketing and less-than-experienced executive leadership. Whether it
was the vaporware that was the Yopy handheld, the fiasco
that was the Indrema game platform, the embarrassing noises made by SuSE as it engaged in year-long management struggle,
the willful refusal of Linux companies to
support a reseller channel, or to consider basic
enterprise marketing strategies (like solutions-based selling as
opposed to technology bake-offs) far too many of the problems faced by Linux
companies had nothing to do with Microsoft and everything to do with bad
management. It put at risk some
really good products.
Big opportunities were ignored. I spent a couple of columns
suggesting that the telecommunications network and mundane systems like network-connected cash registers were great opportunities for Linux-centered
developers. Since then, executives from Intel and several communications
semiconductor shops pleaded with the Linux community to take on these markets.
Instead Linux-centric hardware companies like Penguin Computing tried
challenging IBM and Compaq for the traditional server business, and got their
brains beat in. Because nobody else was interested, IBM and a couple of Japanese
firms began making serious money building Linux cash registers.
And, while
I reported that Red Hat was dabbling in the telecoms market, I’m
still waiting for a Linux-centered challenger to compete with the likes of
Cisco Systems.
The Linux start-ups weren’t the only ones with management problems.
The refusal of most hardware companies not named IBM to properly
support their Linux business was absolutely bewildering. The inability of Sony to prepare a coordinated response to Microsoft’s challenge
to their core business is even more puzzling given the unique
suitability of Linux to Sony’s biggest strategic problems. Maybe it will happen next
year, although the company’s Balkanized management structure and its
past fondness for proprietary solutions isn’t promising.
As for next year, I predict that Carly Fiorina won’t be running HP. I may
have been among the first
to question the logic of the HP/Compaq merger, but now
everyone is wondering, and if HP’s management team is still intact next year
I’ll be shocked. I only hope that Compaq or its successor will continue to
support Linux and give IBM a run for its money on the hardware side.
As for the Linux pioneers, I won’t be shocked if at least three of them are
absorbed or collapse entirely. One
obvious survivor is Red Hat. I’m just not sure who else will be
around.
Despite all that, I do think Linux has a bright future. I
agree with Robin Miller that 2003 is likely to be the make-or-break
year for the platform. The withdrawal of support for NT by Microsoft,
the forced subscription program envisioned by Redmond, the rising
possibility that the state
attorneys general or the
European Union might overturn
a rigged antitrust settlement, the rise of alternative technologies
such as XML all provide a window for alternative platform developers.
As long as Red Hat avoids the temptation to
fork the platform, and Alan Cox and Linus Torvalds can
manage to get along and not let egos get in the way of unified development,
Open Source and its champions can survive 2002 and make the following
year their own.
Happy New Year.
Category:
- Open Source
