The Other Ninety Percent

About a year ago, a friend was revamping the back-end of his small computer services business, and debating which operating system would replace the long-obsolete version of Interactive Unix, which he ran. When he asked my opinion, he had narrowed his choices down to SCO's UnixWare, Sun's Solaris, and Microsoft's Windows NT 4.0. He did not like what he heard.


About a year ago, a friend was revamping the back-end of his small computer services business, and debating which operating system would replace the long-obsolete version of Interactive Unix, which he ran. When he asked my opinion, he had narrowed his choices down to SCO’s UnixWare, Sun’s Solaris, and Microsoft’s Windows NT 4.0. He did not like what he heard.

“Linux,” I said.”Linux, Linux,Linux.”

“Linux?” he asked.


“I don’t buy it.”

“Was that a pun, or are you being pig-headed?”

We’d talked about Linux before, and had never seen eye to eye. Where I saw a cheap, reliable, functional, expandable, and community-supported operating system, he saw a plaything — a science project gone wildly out of control. We’ve spent hours going back and forth, extolling virtues and damning limitations, and had never managed to reach a consensus.

“You’re a small businessman,” I said. “You should be overjoyed with Linux. It’s free! The return-on-investment is instant.”

“It’s not free,” he countered. “It will put me out of business the first time it goes down and there’s nobody to call. The risks are enormous. How can I trust something that I download off the Internet?”

Finally, we decided the problem was one of perspective. He spoke as a businessman, and I spoke as a geek. While I was satisfied with something that simply worked, his head was crowded with concerns that had nothing to do with the technical quality of the operating system — concerns that couldn’t be addressed with code. All of the benchmarks, testimonials, and soothing noises from distribution vendors could do nothing to ease the quivery feeling he got in his gut whenever the fate of his company was being considered.

What he wanted was the vague satisfaction that comes from knowing somebody else’s job is on the line if they screw up. He wanted the reassurance that comes from signing a contract and spending money. He wanted “official” support, a toll-free number, and nicely bound documentation. He wanted the warm fuzzies. Linux wasn’t giving it to him.

“What would make you use Linux?” I asked.

“A Micro–”

“Don’t say it.”

“–soft version.”

“I have to kill you now.”

As Linux runs out of technical limitations to overcome (it’s an awfully impressive science project that can offer SMP support and a journaling filesystem) the obvious next step is to start addressing psychological ones. Swaying the open-minded and technically adventurous, the OS has already won the easy battles. But, the rest of the population — literally millions of edgy businessmen, white-shirt-narrow-tie MIS professionals, and dismissive old-school administrators — still has to be won over for this dream of world domination to come true.

Much of what is needed to convince the straight-laced types of Linux’s place in their world is, unfortunately, antithetical to geekdom and the classic Linux way. Their warm fuzzies come from certifications and schedules, inefficiencies and bureaucracies — essentially, from frustrating stupidities. The benefits of providing them are potentially enormous, but to even begin the effort would require a fundamental shift in the way Linux is developed and released.

“So what do you want specifically?” I asked. “Give me something that could not also be used to describe a teddy bear.”

“I want development road maps,” he said. “I want an advertised schedule for kernel releases. I want to know that if this Linus guy decides that his mission in life is to build sandcastles, the whole process won’t fall apart. I want org-charts. I want a documented development process. I want ISO 9000 certification. I want…”

“But you don’t have ISO 9000 certification.”

“So? I’m the customer, remember? I’m always right.”

When someone wants something that Linux does not currently offer, the simple answer that’s usually given is, “Build it yourself.” While that’s a great method for motivating geeks, it’s a quick way to alienate nearly everybody else. The more complicated answer assumes that a host of companies will rise from the fertile ground to service customers with non-standard needs — the warm fuzzies, for instance. If you need to know that USB support will be out before the end of the year so that you can finish your budget, then you can go with Brand X Linux, which advertises its release schedule well ahead of time.

The trouble is, any warm fuzzies produced by Brand X’s schedule are instantly negated by the fact that they’re, well, from Brand X. By definition, a non-standard kernel produces non-standard warm fuzzies. Who knows if the Brand X USB support will eventually be rolled into the main development tree? Who knows if Brand X will even be around tomorrow? The Linux vendors may be solid companies run by smart people, but to the Luddites of the world, they still seem new and scary. If it’s psychic peace you seek, treading off the beaten path is not the way to get it.

“You’re being ridiculous,” I said.

“Boy, you’ve got that advocacy thing down, don’t you?”

“You don’t need all that stuff.”

“It’s not a question of need. I should be able to get what I want. If the Linux community isn’t interested in providing it, then I’ll go to someone who will. You may know what you think is best for me. You may even be right. But, it’s a decision I’m going to make for myself.”

In order for Linux to rule the world, the scratch-your-own-itch culture that has driven it to this point may need to evolve into a scratch-your-customer’s-itch culture. To be accepted into every niche of the industry mainstream (an industry filled with everything from jumpy business people to newbie home users), base-standard Linux needs to not only be better, stronger, and faster than the competition, it needs to actively address its potential customer’s concerns — or at least as many of them as is realistically possible.

Can the open source model really rule the market? Can it provide the features (or documentary non-features) that bore geeks to tears but customers demand? Can the Bazaar out-Cathedral the Cathedral? It’s easy to dismiss these concerns as old-school thinking, but these remain vital questions for the risk-averse people in charge of millions of potential installations. Now that Linux has provided the geeks with what they want, can it do the same for the suits? For example, is a kernel release schedule more than Linux is willing to provide? World domination may hang in the balance.

“And get your feet off my desk,” he said.

Greg Knauss is the Vice President of Research & Development for a small software company. He can be reached at greg@eod.com.

Comments are closed.