Chasing the Dragon

It was just last week that my friend told me he was going to build his own computer. I asked him to understand in advance that in the weeks to come, after he put it together and it either didn’t work or suffered from a string of ongoing stability issues, I would mock him cruelly. But I would be doing so out of love.

Last night he IM’d me: “OK, you can start laughing now.”

I really do feel his pain, because the same thing happened to me. For whatever reason, a few years ago, I convinced myself that putting together a PC at home from assorted parts would be “fun.” And it was, up to a point. I researched carefully, and religiously followed the recommendations of the best of the hardware sites – you know the one, let’s call it “Tom’s AnandTechnica Extreme.” The parts came in, the pieces went together, the power went on, and I had a fully operational computer, all set to play the latest games of the day. And it was a little cheaper than a Dell!

Of course, it locked up about once an hour. So I spent several weeks de-tuning it until it was slow enough to stop locking up. I knew to do this because of the support forum, filled with thousands of other people who followed Tom’s AnandTechnica Extreme’s recommendation, who were suffering from the exact same problems, who eventually figured out through the process of elimination that the soundcard needed to be in slot 2, and you should only use these certain USB ports, and yadda yadda yadda.

This isn’t unusual. It’s par for the course. Just ask Tilt. Marvel at his plummet from the peak of wide-eyed optimism into the pit of despair.

The whole concept of “building a computer” is iffy. You’re not building anything. You’re assembling some parts where all of the interesting engineering has already been done for you. There is nothing technically challenging – or even interesting – about doing so. Anyone with a grammar school education, enough dexterity to handle a Phillips-head screwdriver, and a grounding strap can do it. “I put together my own computer” (if what we’re talking about is a mainstream Windows or Linux PC) is as impressive a technical achievement as “I installed some software from a CD-ROM.”

The reasons people (including my past self) give for building a computer fall into three categories:

The first argument is the one I have the most sympathy for. It’s the one that actually makes the most sense; if you’re going to spend a lot of money to create something that is inferior to what you can buy ready-made, it had better be fun to put together. But this argument never stands alone; it’s not like people want to build computers and then go on to the next project; they want to build them and then use them for stuff, and there’s the rub. If “fun putting something together” is one of your motivations, you would be better served by doing the following:

(1) Buy a Mac (or a Dell)

(2) Buy some Legos.

(3) When you want to put something together, build stuff with your Legos. When you want to use your computer, start up your Mac or Dell.

The “cheaper” argument only turns out to be true if you are willing to discount all risk of hardware failure any time within the warranty period that would cover a ready-made box. And if you’re willing to discount the shipping costs and hassle of dealing with separate vendors. And if you’re willing to assume that your time is worth nothing. Lots of people are, in fact, willing to do these things. There’s a kind of blame the victim mentality among the home-tuner crowd. The assumption is that if your homebuilt machine crashes, you must be doing something wrong. That’s sort of right: what you did was buy a bunch of parts that were never QAd by any serious vendor to work together for any substantial period of time, believing that it was a smart thing to do. Hardware is tricky. Just because some system survives long enough to run a graphics benchmark for Bladehunt: Deathspank 2: The Revenge does not mean it’s stable enough for use day in, day out, for years.

The performance argument is similarly an argument of false economy. Yes, you might indeed squeeze out a few extra frames per second than the roughly price- equivalent machine from Dell. Or, you could just wait about 30 days, and the next revision of the same Dell will outperform what you would have built yourself. The rate of change of performance in hardware you can obtain by just waiting a little while dwarfs the gains you’ll see by assuming greater risk in assembling your own. If you buy a preassembled PC, maybe Big Computer Company will sell you a lemon. If that happens, you send the whole thing back to them, at their expense, and wait for the working replacement. If your home-built hotrod has a problem, good luck sending individual components back (restocking fees, anyone?)

People like to analogize computer assembly hobbyists to guys who tinker on cars. In my experience, there’s a stark difference (in addition to the people who work on cars, generally, being more knowledgeable in their domain than people who assemble computers). Every guy I know who is serious about cars is completely clear-headed about whether or not a given car he is working on is actually drivable. You never see one of these guys saying “oh, yeah, the rear axle breaks if I drive it over 55, sometimes, but on the whole it’s reliable transportation!”

I don’t really expect to influence anyone with these observations. I still get the latent object sickness from time to time. Fortunately, my friends keep trying to assemble their own computers and encounter the obstacles I’ve described. This helps remind me that I shouldn’t do it again. So if you want to go ahead and roll your own, go ahead. When you are in the depths of your depression, trying to figure out why everything works fine until you try to play an MPEG movie, and then the machine reboots itself (Hi, ewm!) I will be by your side, mocking you.

Mocking you with love.