Of Skunkworks, Server Wars and the Webmaster Renaissance

by Mike Levin SEO & Datamaster, 05/21/2012

Maverick projects are fun and interesting. Skunkworks, rogue, call them what you will, but they are the little projects that make the big difference because ideas can take root that would normally be squelshed by the establishment. No matter how open-minded you think the establishment may be, any new idea or initiative is guaranteed to run afoul of some steak holder—usually the very ones you think should be in favor of it. The problem is the idea wasn’t theirs, and they feel threatened by it. And if, god-forbid, the idea works out, someone else might end up holding the keys to the kingdom.

Therefore, it is in your interest to be able to carry out maverick projects with your own tools and on your own servers—or at least your own cloud account—long enough so that the evidence of your brilliant work is undeniable and self-evident. It is with that timing that you can casually reveal your work and let the new reality to dawn on the establishment, essentially giving them the right of first refusal.

Cloud servers, micro servers, and the free and open source software movement make this all uniquely possible these days. In the past, you would have found yourself in a server war, because you would have needed a resource that was prohibitively expensive to provide on your own, so you would have to go with your hat in your hand to the draconian overlord fellow who administrates such equipment, at which time he would cut loose a hailstorm of FUD (fear, uncertainty & doubt) to your bosses, which would predictably put the kibosh on your plans. 

I should know. This has been a recurring theme for me, and I’m coming up on the ten-year anniversary of my first epic server war, which ousted a president, overhauled a company, and ultimately dragged it kicking and screaming to success. I needed real hardware in those days, and still bear the emotional scars of having to rip it out of the ghastly death-grip of sysadminzilla—nearly incurring the physical damage of the sledgehammer he still keeps in the corner of his office ostensibly to assemble bakers racks, but which everyone knows is for the intimidation effect.

Times have changed quite a bit since June 2002, the date of the watershed battle in my first epic server war, and in a delicious turn of fortune, June 2012 is when my $35 Raspberry Pi micro-server will be shipping—almost 10-years to the day after. This $35 device has nearly the same power as $5000 servers of a decade ago.

But it’s not only hardware that’s massively more accessible. Software has undergone equally dramatic evolution, with free and open source software (FOSS) having matured to the point where there almost no reason not to choose it over the costly proprietary alternatives. You used to have to be an uber-geek to get Linux running, and all the source code compiled and dependencies resolved to have a development platform. Today, you can download a canned virtual appliance and double-click it to run. 

All this weighs very favorably to the advantage of the maverick—the cowboy coder—who only a few years ago was considered a pesk, and not the company’s most exciting potential. I’m not even talking about the high flying stakes of startup business like YCombinator. I’m just talking about the interesting side-projects of today’s knowledge workers.

However, I’m not talking about sloppy code, either. Nor am I talking about that “just enough knowledge to be dangerous” phase that every career server-nazi thinks you’re at. It took me about 10-years to get to the point where I could deal with all the many issues that apps like the ones I write bring up—how not to leak memory, how not to bloat objects, how to minimize state, security, revision control, deployment, optimization, company politics, the merits of different Linux kernel patches, and the rest of the shitstorm that’s going to be blown up your ass in an attempt to knock you off your high-horse. 

It was hard-won knowledge during an era where the single-handed webmaster was on the way out. People think that the subdivision to specialties is the best way and a one-way street. It’s not. It’s the grounds for a new renaissance where visions are realized that couldn’t be in today’s guarded turf.