Even so, I can understand the attraction of having a piece of the Hat. Our friends in Carolina were the first to successfully provide a front end to Linux that was comfortable to those who couldn't hack their own code. (And yes, I know about the early players such as Yggdrasil and others, but none of them did the job anywhere near as well).
But enough about Red Hat's stock going into orbit. What puzzles me far more is what appears to be a similar hubbub surrounding the coming IPO of VA Linux Systems.
VA trades in a very different commodity from Red Hat. Simply put, it sells computer hardware guaranteed to be supported by Linux. The company has done very well at this so far, and has already attracted investment from Intel and others. Even before the IPO, VA is doing well enough to be able to hire Linux notables such as Linux International Executive Director Jon 'maddog' Hall, as well as set up a number of offices outside the U.S.
But how long can it last?
The problem with VA's current approach to the market is that it can survive only as long as Linux remains a niche item. The more of a mainstream success Linux becomes, the worse VA's prospects become.
Why? Look at the competition, which now includes many of the heavyweights of the PC hardware biz. IBM, Compaq, Dell, Gateway and others have all announced Linux hardware support to varying degrees, and these are companies with far broader product and service offerings than VA. While VA specializes in Linux servers, IBM and Dell have Linux servers, Linux laptops, and maybe even a desktop or two. Which was the first hardware vendor to launch a Linux training program? Hint; it wasn't VA.
Now compare VA's prospects to those of Red Hat, which hasn't had to compete against such dominant industry leaders. Red Hat's main competitors -- such as Caldera and SuSE -- are also relative startups, just as small or even smaller than Red Hat. Until Corel comes onto the scene with its distribution this winter, none of the current Linux distributions have a reputation of doing anything non-Linux (with the possible exception of Caldera's DOS operation, and even that appears to be getting phased down in favor of embedded Linux).
The business plan
It used to be that hardware support for Linux was hit-or-miss, and you needed to trust a company like VA to provide hardware that you could count on to work with Linux. But now even commodity peripheral companies are getting in on the Linux act, and most distribution makers provide excellent hardware compatibility guides, such as this one from SuSE.
Moreover, in a world of commodity systems where flexibility is king, VA is a one-trick pony. The Linux support of IBM, Compaq, Dell and others just augments their existing lists of supported operating systems; most of these companies also support OS/2, NetWare, Solaris and UnixWare in addition to Linux and Windows.
VA may get some moral mileage out of saying that Linux is its only business, but that mileage runs out when VA tries to sell to customers who want a single supplier for all their hardware. Such clients won't like the idea of going to one source for Linux boxes and another source for everything else. The more popular Linux becomes, the more the big PC makers can exploit their multi-OS talents that VA lacks (and doesn't appear likely to gain anytime soon).
Consider this: exactly how many single-OS, Intel-based hardware manufacturers can you think of, let alone how many have succeeded? I can't recall anything since Pick-dedicated systems by Fujitsu and others went away many years ago.
Instead of demonstrating mainstream smarts, VA's latest publicity grabber was an announcement that it would be throwing money at the Debian folk and combining with SGI and O'Reilly to repackage Debian's distribution commercially. All well and good, but it does nothing to demonstrate that VA understands the needs of hardware buyers -- or even corporate Linux buyers -- one shred more than the day before the announcement.
I was hoping to talk to VA head Larry Augustin, but neither he nor anyone else would talk about VA's future. Normally difficult to contact at the best of times, Augustin is now muzzled by regulations since the announcement of the IPO.
The VA filing says that development of services such as integration, consulting and training will be critical to the company's success. But little of that infrastructure is in place now, and VA has but a fraction of the VAR networks so carefully cultivated by its competitors.
Don't get me wrong -- VA makes very high quality systems and is staffed by enthusiastic, skilled, hard-working people. But I just can't understand how anyone can see VA as a reasonable long-term investment, unless they're counting on it either buying, or getting bought by, a company that can take it beyond a model based on Linux-as-a-niche.
I wish VA well, wherever it ends up... so long as it isn't bought by NEC.