There's more computing horsepower in a DVD player than an Apollo moon rocket. Governments have launched missiles with computers less powerful than a Nintendo game. The average new car has more technology than entire universities did in the 1980s. There's no question that computer hardware is getting faster, cheaper and more ubiquitous. But it's the software that really defines the consumer electronics experience.

Sometimes that software is hidden; sometimes it's the whole product. Clearly, the consumer entertainment electronics market offers a lucrative new frontier for smarter software. But software for consumer entertainment systems faces a number of challenges that PC software simply does not have to contend with.

The biggest challenge-but also the biggest opportunity-lies in the fact that consumer systems consume about as many chips as computers do, but, unlike computers, there is no standardized hardware or software "platform." No operating system or application dominates; there's no equivalent to Pentium or Windows. Every vendor has-and wants-unique features, functions and frills. Unlike PCs, Macs or workstations, consumer electronics is wide open for new development.

Yes, some consumer items do run recognizable operating systems, including Windows or Linux. TiVo, for instance (the device Federal Communications Commission Chairman Michael Powell enthusiastically described as "God's machine"), is a Linux box. TiVo's hardware is pretty unremarkable; it's the software that makes it such a success. Medical scanners may run Windows, too, and some in-dash navigation systems-and even radios-are based on Linux.

Yet, there's little need for a universal software standard across consumer items. That's partly because there's no such thing as a "standard" consumer device, so one operating system couldn't hope to satisfy the diverse needs of the medical, automotive and entertainment industries. Instead, the operating system market is fragmented among dozens of vendors. From Windows and Linux down to small real-time operating systems and kernels, there's enough variety and complexity to handle any imaginable consumer device. These operating systems are usually hidden from the user, so there's no need for a familiar interface. The software at this level is truly embedded.

Communications presents another dilemma that's mostly practical, not technical. "Easy to use is easy to say," goes the old adage. Plug-and-play stereo gear used to be pretty simple: You plugged in the cables and played your tunes. Now communication among consumer devices is getting far more complicated, with TV add-ons fighting over which box controls the incoming video, which one records it and which one ultimately formats it for the new plasma screen hanging on the wall. In-car networks (and, yes, they are networks) run throughout the vehicle and might easily include two dozen chips talking amongst themselves. If we're to impose any order on these warring factions, we'll have to improve the communications-and that's a software problem. Today's separate boxes and microcontrollers will need to merge into one, or some dominant company will need to step forward and impose order, or all the competing vendors have to agree on some guidelines. Neither of the first two alternatives seems likely, so we'll have to put our trust in the latter option.

Merging living room electronics into a single device makes compelling sense-to an engineer. The TV tuner, DVD player, satellite or cable receiver, game player and video recorder ought to be one box because they're essentially all the same hardware. But history suggests this "TV appliance" will never catch on. Instead, these boxes need to be detached yet communicate with one another, preferably without masses of cables and wires. That calls for some tricky software and a few new industry standards. Sharing remote controls among devices is difficult enough; squirting video streams from box to box would overwhelm current 802.11 Wi-Fi standards and might cause widespread interference with cordless phones-or the neighbors' video gear. There's also the odd side effect of broadcasting one's video playback to the neighborhood. In the future, everyone will be their own TV station.

Even simple consumer devices have software 'stacks,' with each interdependent layer of software developed independently.

Setting standards in the consumer electronics business is a tricky political game. Most manufacturers proudly differentiate themselves based on their software and try to avoid look-alike standards. So much of a consumer device's functionality is predetermined (the way a DVD is recorded, the way e-mail is stored and forwarded, the format of a TV broadcast, etc.) that manufacturers want to take advantage of any leeway they have. That generally boils down to the arrangement of the buttons on the front panel (the "user interface") and the color of the plastic.

User-interface guru Niall Murphy decried his new DVD player's user interface, which displayed the words "root menu" on its front panel. To an advanced computer user (and, specifically, to one familiar with the Unix operating system), this is a common expression. But to the other 99.99 percent of the world-including the product's target audience-this is a completely meaningless phrase that seems almost intentionally bewildering and intimidating. Who could have thought this was a good idea?

User interfaces shouldn't be standardized, at least not anytime soon. Experimentation and continual improvement are the order of the day, and some consumer electronics companies are doing a good job of "de-complicating" their gear. In the computer world, user interfaces are determined by fiat; Microsoft dominates the PC operating system market, so its GUI is the de facto standard. There's no particular need for stereo, navigation or pager interfaces to adhere to any standard. In fact, the variability among user interfaces is a major competitive differentiator.

The key to smarter software is not to make it too smart. "Creeping featurism" has doomed many an otherwise well-thought-out product. Manufacturers have more than enough processing power at their disposal to pack an entire computer, literally, into a dashboard navigation system or e-mail appliance. BMW has done just this and been raked over the coals for it. The key is to resist that temptation. We've all dealt with spam filters or e-mail rules that behaved unpredictably because the software was "too smart" for its own good. That's barely tolerable in the computer realm; it's instant market death with consumer electronics. The best features are the ones that are obvious, clear and unambiguous. This is no place for subtlety.

Networking among devices will be even harder. There's just no good way to present network topologies, domains and Internet Protocol masks to a consumer. It should all be invisible and work "thoughtlessly." Consumer items sometimes adopt PC-centric hardware, such as USB and Wi-Fi, that needs a central operating system and lots of tinkering, so making quick connections between simpler consumer items is very tough. Consumer electronics companies need to expand (or redefine) the existing standards, but that risks making current equipment incompatible or obsolete. Given enough software, all that complexity can be hidden, but it will take careful planning and more cooperation among vendors.

The details may be hazy, but the prevailing trend is clear: Consumer electronics is booming and there's no end in sight. In good economic times and in bad, we all like to have our toys. Today, the average middle-class household includes more than 40 computer-controlled systems.

High-end cars already have more than 100 microprocessors communicating over fiber-optic networks. Between the rapid advancements in hardware and the infinite flexibility of software, there's a vast blank slate for the imagination. And there's plenty of imagination to go around.