Microsoft is floating an "iPod killer " which will pack audio and video playback into a Windows CE-based device costing $400 to $700. Units will be available in mid-2004, but I'm going to go out on a limb and say this product (at least as it is being shown currently) will land with a dull thud. In fact, I'm going to go even farther, and suggest that this will become an excellent case study of how good theories don't translate into good products.

Theory: To beat iPod, you have to do it better. It is pretty hard to refine the digital music player much beyond the iPod, which fits the niche almost perfectly. An excellent combination of storage capacity, form factor and experience (aka user interface), the only thing it needs is a lower price, and that's supposed to be coming. So how do you do iPod one better?

Why, add video, of course!

Bzzzzzzt! Wrong answer. It is easy to see video as the logical progression beyond audio, but it isn't. Especially for a portable device. You can't watch a movie while jogging, bicycling, driving a car, etc. In fact, for video, you pretty much have to be sitting down in one place, since it commands all your attention. So, adding video doesn't make it a better iPod, it puts it in a whole new product category: portable video player.

Guess what? This category already exists, but perhaps that fact got overlooked in the zeal to beat iPod. Here's a whole page from Amazon . These units take regular DVDs (so you don't have to download the movie to your computer first), they have screens up to 5" or larger (vs 4" for the Microsoft version), they cost as little as $170 (vs. $400), and they don't run Windows. Which leads me to the next theory....

Theory: "Windows Everywhere" is Microsoft's strategy of success. Microsoft has been famous for its "Windows Everywhere" strategy, and justifiably feels that this strategy has been an overwhelming success. At least....if by "Windows Everywhere" you mean "Windows running on an Intel/AMD platform in a PC or laptop anywhere it makes sense."

Microsoft doesn't enjoy anywhere near the market share in the handheld market as it does in PCs, and that's arguably the only non-PC non-server market where Microsoft has achieved any success with Windows at all.

I am constantly amazed that Microsoft doesn't understand that people don't want Windows Everywhere. With all their usability engineers and human factors specialists and $50 billion in the bank, Microsoft hasn't figured out consumers want to push PLAY and have it PLAY. The last thing most people want to deal with is a Windows-style interface layered on top of a music player, video player, mobile phone, or refrigerator.

You will note, by the way, that the iPod does not run MacOS X, though it certainly interfaces well to MacOS-based computers. On the other hand, Apple does not seem to need to extend its MacOS franchise beyond actual Macintosh computers.

Theory: Microsoft makes software, other guys make hardware. The past few decades in the computer business have been all about moving from vertical integration to horizontal integration (or, in less jargony terms, going from buying everything from IBM to buying the computer from Dell, the processor from Intel, and the software from Microsoft). This is the strategy Microsoft is pursuing with the media player, but the limitations are apparent.

Why, for example, would anyone want to build a media player based on Windows in the first place? The short answer is because Microsoft is providing the software layer, and insists that the software be Windows. Unfortunately, the end result is a beast which nobody will want, largely because of the fact that Microsoft felt the need to extend its Windows franchise.

Now, if Microsoft could get beyond Windows and behave more like an embedded software company, you might have a good product. But that would require giving up the Windows stamp, and that's not consistent with Microsoft's strategic goals (after all, how many outside geekdom have ever heard of Wind River? How many of us own something which contains Wind River software?).

The lesson here is that horizontal integration works only when everyone's strategic and tactical goals are consistent with making a winning product. If not, better to have one company do everything and avoid the conflict between vendors.