Dreaded question from friends and acquaintances #568: “What kind of computer should I buy?”

macallan wrote:

As an educational IT guy, I am frequently asked about when to buy a new machine. My advice is always: buy when you need a new machine, and get the “best” one for your needs at that time. There are lots of separate discussions about what the “best” machine for any given user will be, but I never advise people to wait more than a week, since there is always something better[1] coming soon[2]. Most people are constrained by the lead time on a new machine, or some other deadline, so I say just buy something already.

1] that’s what marketing calls every new thing
2] marketing will have you believe it’s going to be here tomorrow

I think that’s right; when I’m asked about “what computer should I buy?” I generally tell folks it’s best to buy based on:

  • When you need it, at the latest
  • What you need, at a minimum
  • What you can afford, at a maximum


Once you need it, buy whatever within your budget meets your needs. If you know you won’t really need it for awhile, I say wait—it’s certain that once you’re ready, the products will be cheaper, more full-featured, or (likely) both. (There are exceptions—I’ve been in the market for an HDTV panel for a few months, and am rather distressed at how prices have been increasing or at least holding steady of late. But it usually holds.)

The budget question is the only one that I think really takes some finesse depending on individuals’ situations. If the buyer is a non-techie and is buying a primary personal (or work) desktop computer, I say, no question, decide the model family you’re interested in, and then blow your budget on it—every additional dollar spent on RAM or CPU or storage or graphics or screen size is an investment that will delay the day you need to buy a replacement.

If on the other hand, the buyer is getting a special-purpose device (printer, scanner, music player, whatever), I say find the model that does what you need, then step up by one increment of speed or storage or whatever if you can afford it. But don’t bust the bank on an accessory.

Notebooks are a bit of a weird one. They have a much riskier lifespan—both in terms of the chance of destroying the device through mishandling and through the volatility of features from generation to generation—and so I tend to be a little bit more conservative on budgeting for them. They also tend to have fewer field-upgradeable parts, and something that doesn’t seem like a limitation today (for example, the lack of a builtin WiMax or 3G telephone card) could become a very big deal in a year or two. Personally, my last three laptops (all Macs) have been factory-refurbished models—an easy way for me to save some money on what I consider to be a riskier purchase. (I’ve never had any problems with refurbs, by the way—I don’t know if that’s just Apple, though.)

One other thing—I think techies have a much lower tolerance for the primary instrument of their craft being creaky or slow, and so where a non-techie might buy a desktop computer with a 4+ year lifespan expectation and a notebook or accessory with a 3-year lifespan expectation, I think techies should adjust those figures to 3 and 2 years, or even less depending on your desire for the latest and greatest. (Too bad the IRS won’t let us adjust our depreciation schedule to two years….)

This observation, I would posit, would recommend that unless you’re independently wealthy, you may want to confine your purchases to the midrange models. Today’s midrange is going to be equivalent to the high-end model of just a few months ago (a year at the outside), but will be significantly cheaper. So you’ll be able to buy more frequently at the midrange and keep the period when your computer seems “old, slow and crufty” as short as possible—maybe even to zero—without killing your wallet. It’ll also give you better control over forecasting your budget, since midrange prices are less volatile than the high-end.

Back to first-generation models: my last posting was a bit flippant, as I do think that pre-ordering a first-generation model before any have been shaken out in the real world is risky, and I also think that buying the week before an event like MacWorld is just asking for heartburn (or heartache, depending on your relationship with your technology).

If you recognize the risk, though, and buy with your eyes open, I wouldn’t get too dogmatic or absolutist about it. When I saw the very first iPod announced in October 2001, I put in a pre-order that day. I already had a little experience with MP3 players, having a flash-based Diamond Rio, and the parts in it—FireWire, a laptop hard disk, an RCA jack, a monochrome LCD screen—seemed pretty straightforward and well-tested parts. I got bitten when Apple’s promise to have iTunes for Windows “soon” turned into a two-year wait—I was running Linux and Windows exclusively at the time—so had to use some hacks to get my music on it for the first few months. But then I bought an iMac G4—again a preorder on announcement day—mostly because a) I was hearing good things about this new Unix-based Mac OS X and wanted to try it (and be able to sync my iPod without hacks), b) all my other computers were provided by work and I needed a truly “personal” computer, and c) the thing just looked so freaking cool, like no computer I’d ever seen.

I continued to use that mechanical-scroll wheel iPod until just last year, and the iMac G4 is *still* in use (though a bit hobbled since there appears to be no way to get a modern WiFi (a, g, or n) connection working with it). I knew I was taking a risk on those two purchases, but they worked out brilliantly. Again, you just have to do it with your eyes open.