The urge, so widespread during this technological era, to improve things continually has resulted in a weird sort of entropy called "creeping featurism." The term was coined by writer and software designer Jim Edlin in the early 1980s, when the phenomenon first appeared.

The idea is that continually improving things just adds more and more complexity. Eventually there comes a point where a product is unusable except to those people who have kept up with each new change. Anyone trying to start with the program from scratch late in its evolution will find it impossible to understand.

GAMES THAT ARE TOO HARD TO PLAY. This problem is exacerbated with games. Anyone who loses a generation of upgrades can find it hard to play the latest iteration with any skill. This is one of the reasons that Nintendo is beginning to rise again from its third-place position in the gaming world. It has decided to concentrate on developing new game models that are both challenging to experts and approachable by beginners. The other platforms are suffering from creeping featurism, making games so difficult that only old pros can play. Newbies are discouraged because of the evolution of these established games.

When the concept of creeping featurism gets debated, I always like to ask what the Mona Lisa would look like if it were never finished but were continually "improved" with new features. Can you imagine? In some instances, an entire company moves away from its core competencies in a giant macro version of creeping featurism. Take Microsoft (please!). Here is a company that began as an alternative to corporate software. Early TV ads for Excel showed employees skulking around the office using Excel on the side. It was subversive. Now it's just the opposite. It has become what it sought to avoid. That all culminated in the .NET strategies, which have since been subverted by outsiders and LAMP (Linux, Apache, MySQL, and PHP). And it appears, to me at least, that PHP too may be beginning a long journey down the creeping-featurism tunnel of doom. We'll see.

The problem is that nobody will freeze code and call it a day. This is partly because a company has to respond when someone comes along with a competitive product that has a unique feature that people desire. But those features inevitably pile up, making the product difficult to use. Every once in a while, I actually want to use some fancy feature in Microsoft Word and cannot figure out how to do it. The item is in the wrong menu or buried beneath a number of drill-downs.

The drill-down menu itself is a creeping feature. Anyone who visits a lot of Web sites will quickly revert to the site map (if there is one), since it offers all the clickables of various drill-down menus in one well-organized location. I'm surprised the idea of the site map hasn't found its way back into the world of regular software. A site map of Word features would be easier to use than all the drill-down menus and buttons. But I digress.

HOORAY FOR THE PLUG-IN. Something does need to be done about the creeping featurism problem, and the only thing practical (albeit nontrivial) is to add features only as plug-ins. Microsoft is always looking for ways to make money, so consider this idea: Sell only bare-bones software, and all new features can only be sold as plug-ins.

Plug-in-driven software has always been an interesting phenomenon. I credit Adobe Photoshop for popularizing the concept. Adobe's use of the idea, though, is to let third-party folks be able to add specialized functionality to the product. Instead of freezing Photoshop code and adding plug-ins to upgrade the product, Adobe has continually upgraded the base code, making the base product harder to use. It kind of missed the point. If instead there were a skeletal product that evolved only through plug-ins, then users would never be left struggling with ever harder-to-use feature-rich versions.

The drawback to this, of course, is the clone software with the fancy features built in. In other words, since the Adobe skeletal software has fixed features, it could be cloned. But I'm not so sure that would be a disaster. If the core of the skeletal product were continually improved, not with creeping featurism but with performance enhancements, the competition would still have difficulty. People still want the best-of-breed product.