February 2002
On my browser is CNET piece titled "Got Linux? Many companies say no", by Sergio G. Non. The subhead reads "News Analysis: Linux penguins are braying louder, but companies don't plan to adopt many of them in the near future". The opening text paints fact as speculation, then shoots it down: "...because the economy is still weak, many tech observers believe that Linux--and its price tag of "free"--will attract more businesses looking to cut costs. At least that's the theory. Practice indicates something else."
But the "practice" is anything but. It's a Goldman Sachs survey of 100 techonology executives who were asked to name their highest and lowest spending priorities. Tops were migrating to new versions of Windows, security software and UNIX servers. At the bottom were mainframes, supply-chain management software and Linux servers. Again, these were "spending priorities." No practice involved. Also no respect for Linux' virtues as an instrument of savings.
We finally run into practice about halfway into the piece, when the author reports that Amazon has just "saved millions by switching to Linux from UNIX in many areas of its business." Then we read quotes from an IT consultant who says "Many of our clients consider Linux to be a very real option for cost savings," and "There's a definite ROI." Later the same consultant adds, "I would agree that many comapnies are scaling back forward-looking projects... However, the feedback we get from our clients is that they are very interested in Linux."
In fact the weight of the evidence in the piece actually favors Linux, on the whole. But that's not the story this CNET writer wanted to tell. There's a backlash against Linux right now in the business press, and this guy was just doing his part. To be fair, there's a backlash against everything associated with the dot-com bubble, from venture capitalists to day traders to the countless wacky ideas that attracted hugely speculative investments.
But I think the backlash hits Linux worse for a number of reasons. One is that there aren't many pure "Linux companies" left. The Business Tux population has thinned to the point where it looks like an endangered species -- even while lots of big companies like IBM are crooning their love for Linux.
Last week I purged my collection of business cards collected over the last several years, and all but a few of them went in the trash. The ones I tossed were mostly from companies that are now dead, absorbed into other companies, or busy "repositioning" themselves as something else.
Andother problem was Linux' brief but spectacular moment in the spotlight on Wall Street. That spotlight fell on Tux right at the very peak of the Street's speculative fever. In late '99 no fleece was more golden than Linux, even though very few investors knew what Linux did, other than appearing to threaten Microsoft. Between August and December of 99, Red Hat, Cobalt, Andover and VA Linux all had huge IPO successes. VA's debut run-up was the biggest in Wall Street history.
Today the only one of those four companies that still proudly flies the Linux flag is Red Hat. And I say hats off to them. Cobalt was acquired by Sun. Andover was acquired by VA which recently filed to change its name to VA Software and even to rid its stock of the once-valuable LNUX name.
The larger problem was that too many dot com companies were not companies at all. They were projects. Ventures. They were like those businesses with false fronts that lined the streets of mining towns built overnight in the Old West and abandoned almost as fast when the lode ran out, if it was ever there in the first place. From the street they looked like real businesses in real towns, but in fact everything in sight was a gamble.
But Linux isn't a gamble. Not as a technology. But it also isn't a business. And that brings us to our third problem, which is trying to understand and explain it to real businesses. The Goldman Sachs study cited by that CNET article typifies the problem. Linux in its pure state isn't for sale. You can't look at its popularity in terms of sales.
Worse, prognostications have way of flat-out sucking anyway. Yesterday I went through boxes in the garage looking for a document I've been saving for a dozen years because it was so spectacularly wrong, right from the minute it was printed. I couldn't find it, even though it was thick as a phone book and chock full of evidence for the continued growth of OS/2, Novell and other brand-name LANs, the likely success of AT&T in the computer business, the continued growth and proliferation of a dozen different UNIX brands, especially AT&T's original UNIX, recently bought by Novell and rebranded as "UnixWare". Lots of stuff about Lotus Notes. The continuing share losses for Apple. There was nothing, of course, about the Internet. Nothing about Linux. No idea that a PC would ever be hooked up to anything more significant than a printer and a file server.
But I did find some wonderful other nuggets. Here's a Sports Illustrated from November 30, 1981 with a cover story on the preseason #1 North Carolina Tar Heels, with four of the team's stars -- James Worthy, Matt Doherty, Jimmy Black and Sam Perkins. Not shown is the player that would emerge not only as the team's biggest star, but the best player in the history of the sport: Michael Jordan. And here's a Popular Electronics from July 1976. "Eclusive! Now You Can Build a High-Quality Intelligent Terminal!" it says. I page through it. There are lots of ads for CB radios and stereo gear, electonics courses and antennas. But near the front is this two page spread. "Imagine a microcompuer" the headline reads. Below is a box with LEDs and switches on the outside and an inside that's all backplane and power supply. "Imagine a micocomputer" said the headline. The copy went on...
Imagine a microcomputer with al the design savvy, ruggedness and sophistiction of the best minicomputers.
Imagine a microcomputer supported by dozens of interface, memory and processor option boards. One that can be interfaced to an indefinite number of peripheral devices...
It was for the MITS Altair 8800-b. It didn't succeed, but the vision sounds mighty familiar, doesn't it?
Doc Searls is Senior Editor of Linux Journal. His monthly column is Linux For Suits. He is also a co-author of The Cluetrain Manifesto.