OK, it's time to check in the bottom of the birdcage and see how many of the last few year's predictions came true. Finished? Yes, they look like bird poop. Sort of like how they started …

For the first time in perhaps 4 or 5 years, I will not predict that Microsoft will be broken into pieces. It looks like it's finally "game over" for that one, thanks to a Justice Department whose slogan has become "Antitrust - we've heard of that". Only Massachusetts and a few other diehards, including Sun Microsystems and AOL, will carry on the fight to halt Bill Gate's coronation as Big Brother (new corporate logo: Innovation Till It Hurts). Ah well, it was fun while it lasted.

New attempts to find something in technology that captures the public's imagination will reach ludicrous heights, although topping the recent ads about taking pictures with your cellphone will be tough. Look for such technological innovations as cat/camera implants (send kitty in to see what your kids are really doing while you're not watching), instant messaging that reads brain waves to really save time (psychiatrists should love that one) or a microwave that dials your cell when it's finished heating dinner. Seem farfetched? Not more that having a wireless networking router that will actually talk to a laptop from the basement to the 3rd floor of a typical Columbia townhouse.

Moving back to the real world, expect a continued - but probably fruitless - push for grid computing. This is the idea of using multiple lower-power machines networked together to emulate a higher-power version. Recently, Gateway announced a plan to employ the unused power of all their demo machines sitting in showrooms as available for solving problems usually associated with supercomputers such as Cray. The theory is to break a computing problem down and parcel it out to units just sitting there scrolling logos or playing songs, and re-combine the results. Yes, it can be done - higher order servers often contain multiple processors, but usually in one machine. So why do I think this idea is another for the computer hall of great ideas that never made it? First, it requires a pretty uniform batch of computers to effectively divide the work - one reason why this idea will not be easily implemented in a typical office with the usual mixed bag of brands, ages, versions of Windows, etc. Secondly, the reliability factor: although grids are redundant and thus somewhat immune to failure of one (or several) computers, it still can be a problem. But primarily, the issue will be security: if you're developing a new drug and want to crunch a computer model that requires such vast power, are you going to want to give your data to Gateway and let it be spread throughout their network of stores? Really, really doubtful.

Also headed for the "faceplants in the sands of time" exhibit will be IBM's push for "computing on demand". Try reading their website to find out what this means, but only if you don't suffer from dizziness. It combines the idea of grid computing with remote access, probably to IBM suppliers who will "share resources" for a fee. Those of use with early memories of sitting at teletypes (yes, the ones with rolls of yellow paper and punchtape to serve as data input) doing time-sharing recognize this as updated version of an old idea. Again, it uses the idea of tapping into unused power of often idle machines to feed the grid. Maybe useful for very big companies who often use remote access to central server collections anyway, but strictly out of it for most small to mid-size companies. Most everyone would prefer to have their own computers under their own control, and rightfully so.

So, any ideas of what will work? Not today. Perhaps after several evenings before a cozy fire (a real one, not a video feed to a laptop) with a good bottle of wine … now there's a prediction I can really get behind.