Windows XP will, unless the privacy advocates have their way, be with us within a few days and there'll be a review on this site soon after it's launched. As with every other new piece of software, from Microsoft or anyone else, this new version is going to be better, faster, more intuitive, more feature-rich and user-friendly than its predecessor. Of course it is.

There's a problem, though, and it's not specific to Windows XP. How do you make something feature-rich and user-friendly? Somehow you have to hide those features that people aren't likely to use all the time, while simultaneously providing quick access for those users who do require them. 'Smart' menu bars in applications can help hide this problem, but not cure it (it's bad enough not being able to find something where you left it in real life, let alone in an 'Edit' menu).

But you can't blame the programmers. The whole point of new versions of software is to build in lots of fancy new things that the old software didn't have. Otherwise, why would anyone buy it? After all, there are plenty of minimalist word-processors and spreadsheet applications available, some of which we've reviewed on this site, but somehow they never have such mass appeal as the latest, most expensive and feature-laden tools from the market leaders. Partly this is due to the powers of marketing, partly it's due to user ignorance; people don't know exactly what they need so they play it safe by going with the recognised name.

The continuous upgrade cycle has its problems, though. Invariably, the first incarnation of the 'new' product will have bugs and it will take a few months to sort them out. Then a more fundamental problem may be discovered, but by then the next version will be on the way. Users have a choice; struggle on with the existing package or pay more for a new set of features - and a new set of bugs.

Somewhere in any product's life there's a version of the software that does just what the user will want and nothing more. Consider Microsoft Word. All that most people want from a word-processor is the ability to format and edit text, perhaps do mail merges and insert a few photographs. Having their spelling mistakes underlined is a nice touch too. So for 99 percent of the population, Word 97 would appear to be the ideal balance between features and usability (and if you don't want the spell-checker underlining, Word 6 would do). Anything more is likely to be a waste of processor power.

All of this is largely the reviewers' fault. Magazines have been so obsessed with comparative lists of features that they tend to ignore the important factors - is it fast, is it easy to use, does it produce good results for the majority of users? We forget sometimes that more is not necessarily better. It's the same with new processors. Looking at the computer magazines of five years ago, anything more than a 200MHz Pentium was seen as overkill; "only necessary for the CAD brigade". Now we have PCs with ten times that power being sold for use as glorified typewriters and Internet terminals.

But the 'early adopters' have to take their share of the blame. Early adopters just have to have the latest software and hardware in order to make themselves feel temporarily superior to their friends and colleagues.

"Only 1.4GHz? That's nothing. Mine's got 2GHz and 512MB of RAM", they'll say.

It's eerily reminiscent of the car park conversations of my youth.

"Only 140 horse power? That's nothing. Mine's got 200 BHP and a 5-speed gearbox".

Maybe version numbers and processor speeds are the new 0-60 times and horse power figures. If so, the recent downturn in PC sales indicates that perhaps people are realising that they don't need a Lamborghini Diablo at home; for most of them, a Nissan Micra will do.