The history of IT since the 1970s can be summed up as one giant quest to find a better place to put stuff -- hardware, software and data. First it was on mainframes, then on clients and servers, then in N-tier arrangements; then it was outsourced or offshored, then "virtualized," then put on external Web servers, then moved into the . There were even a few flashbacks: Mainframes were said to be back in style; thin clients were in, then out, then back again. Every year or so, it seems, someone comes up with a "better" idea of how to slice and dice computing's resources.

Here's how it works: An IT manager has a certain computing infrastructure in place for his company. It works OK, maybe better than the last thing he had, but of course it does have some problems. Onto the stage stride the vendors, the analyst blowhards, a few peers and maybe a few users, all saying they have a better idea. Indeed, the IT manager needs a better idea in order to deflect criticism for that downtime last week and to justify a budget increase. Plus, his ventral striatum says new things and risk-taking are good.

Our hapless IT manager finds that changing to the new idea is much more expensive and painful than anyone expected and that not all the promised benefits are realized. But at least he has a new "platform" on which to catch his breath until the next big thing comes along.

Am I suggesting that we IT people -- who live on the leading edge almost by definition -- should run away from the next big thing? Should we emulate our financial brethren, so recently drunk on risk and leverage, who are now afraid to get out of bed? No, I'm just suggesting that new computing paradigms always have a certain faddish quality, and they entail risks you don't always have to take. Caution is back in style.

So this would be a good year to make what you have work a little bit better -- a good year to invest in training, procedures, documentation and other boring things. Then, by 2010, you'll be ready for the next silver bullet.