IEEE Internet Computing, March/April 2008:
Is anybody home?

A week or so ago, the wireless router in my home failed. This wasn't good news, but it didn't seem like a big deal. I'm reasonably knowledgeable about networking — the home variety, anyway — and I assumed that replacing it would only cost me a few bucks and an hour of my time.

Ha.

To keep a long and painful story short, I went through three different brands of routers before one worked. I'll leave out brand names to protect the guilty, but this was neither a pleasant experience nor a fun way to spend roughly eight hours of my life (not counting travel time to and from the store). So-called "easy install" programs didn't work, the enclosed documentation didn't document what I needed to do, and telephone support insisted on helping me solve a problem I didn't have. I was relieved when things starting working on my third try, but only in the sense of "I like hitting my head against the wall because it feels so good when I stop."

Shortly after getting my new router working, I read some press coverage of the recent consumer electronics show — CES — where the big story was about how all the devices we'll be buying in the very near future will be Internet-capable. I read this, thought about my router experience, and had a good laugh. Then I had a good cry.

Time has passed; I've calmed down a bit, and my new router is working fine. But I believe these experiences illustrate serious questions about how everyday devices will make their way into a connected world. I've touched on these issues before; my last column — "Stumbling Forward into a Connected Future" — was a somewhat pragmatic discussion about a particular set of devices and what I must know and do to have them work together. Now, given the comments coming out of CES, I think we should pull up a level and think about the broader notion of an ecology of connected devices: why this notion didn't work in the past, what it'll take to make it work this time, and what issues we, as designers of this ecology, should be thinking about as we move toward the future.

What happened the last time?

Some of you might recall that 1999 and 2000 were years filled with excitement about a new class of products called "Internet appliances." On the surface, the story was much like the one being told now: new classes of devices were emerging that could be connected to the Internet, expanding the role of content, information, and services, and offering greatly simplified use. This was a good idea with good intentions, but it failed. Why?

Too little value, too little loss of complexity. Many of these devices were simple PC replacements, offering Web browsing, email, and media and information management. Nevertheless, they had much of the complexity of a PC but a "real" PC was only a few hundred dollars more.

Too little networking. Back then, broadband was less common than it is today and in-home connectivity was often limited to wiring a couple of PCs into a router. Beyond that, few widely accepted standards existed for Web services and collaboration on which networked devices might be built, thus there was no guarantee that these devices would be able to talk to each other

Product-driven, not people-driven. Many of these products' advocates were large electronics companies who saw the growth of the home market as an opportunity to tie all of their products together. Unfortunately, these opportunities were often thinly veiled attempts to lock consumers into their products, and many of the product ideas were poorly conceived. At CES 2000, I saw at least three different demos showing how a company's offerings would let you control your microwave from your office computer. Personally, I think these demos came about less from consumer demand and more because the companies involved made both computers and microwaves and were looking for "synergy" between their product lines.

Nevertheless, the industry recognized the significance of the potential market, and a lively battle for control of the home began: PC companies, telephone and cable companies, content providers, and new companies with innovative product visions all had compelling arguments for why they should be the center of this new universe and why the rest of the industry should coalesce around them. Of course, instead of a unified market, we got confusion. Product visions centered not around true user needs, but rather one company's world view. The technical arena was flooded with competing technical standards that were designed to benefit the company proposing them, and the business arena was equally flooded with competing models for how, when, and to whom money should flow. The result of all this was a poorly defined and complicated world, one that was ultimately rejected by confused and even scared users who couldn't see the value in replacing all their appliances in order to gain remote control of their microwaves.

What needs to happen this time?

Today, we can hope that the conceptual, business, and technical progress we've made in the past few years has clarified some of these matters and given us a firmer base on which to create a new ecology of devices. It's now time for me to be constructive and use this soapbox to discuss some issues that, from a user-centered perspective, we should pay attention to in that creation process. Things have indeed changed since 2000, but we as an industry still have the opportunity to mess things up if we aren't careful.

Don't overlook the obvious

There are obvious things we can do, such as identify a real problem that users face and offer a complete solution to it. We can look at media management, and music in particular, as an area in which this has been done well. There's a clear problem to solve — how can you have access to all your music, anywhere — and the commercial success of Apple's combined offering of iPods, iTunes, and the iTunes Music Store shows the rewards of thinking carefully and offering a complete solution. This is, of course, a good thing for any project or product to do, but it will be especially important as consumer devices move onto the Internet because they're likely to build on a combination of personal devices, PCs or other kinds of PC-like devices, and services.

Equally clear is the importance of getting the physical user interface right. These new devices will vary greatly in their interface requirements and opportunities, and many of the interaction conventions we've become used to might not apply. As always, the best starting point is a careful task analysis of how your users will really use your device or service.

Minimize the amount and depth of knowledge neded.

Of course; you knew that already. But the particular challenge of working in this area is that some degree of technical complexity — subnet masks, gateway addresses, and encryption types — are unavoidable. Fortunately, there are ways to deal with them that don't require "ordinary" people becoming networking experts:

Fewer options are generally better. My wireless router offers me at least 10 different versions of encryption. They might all be necessary for legacy reasons, but they shouldn't all be offered as primary, equal alternatives. If you're a company that makes both computers and wireless equipment, you have an advantage; you can design both as part of a single solution that implements the best approach and hide the alternatives from users who don't need or care about them. If you're not such a company, I'd suggest that this is an area for collaboration and co-design between partners who are targeting different aspects of the same market.

Re-use what the user has already done. If, for instance, your phone's email setup should be the same as your PC's, it should be simple to push those settings down into the phone instead of making the user find them and enter them manually.

Let devices find and configure each other. There's been a lot of progress in this area in the past few years — consider the improvements that have come with Bluetooth, Zeroconf/Bonjour, and even dynamic host configuration protocol (DHCP) which let me start up and use a device without worrying about the configuration details. Moving forward, the challenge will be to repeat these successes at higher and more useful levels. For example, what would it take to build an email system that, given an email address and password, "just works," with the degree of functionality and security that people will and should expect? How would you retrofit it to all the mail servers and clients that currently exist? This is a broader notion of service discovery than that used in the devices just mentioned, but there are a lot of opportunities we can pursue here.

Think and act carefully about information security. The more that personal information finds its way into Internet-enabled devices, the more opportunities there are for problems. Some of this is pretty basic stuff — if my phone broadcasts my credit-card number to another device, that transmission should be secure. But, beyond that, huge questions arise about the kinds of information these devices collect, create, and distribute. What information gets collected? Who owns or controls it? What can they do with it? Who can get access to it? How do I control it? (Or, what part of it can I control?) There's a big trade-off here: the benefits coming from the great things that could be done with a complete aggregation of information versus the risks coming from the awful things that could be done. This is becoming increasingly evident, as mobile phones become ever more capable devices and Web-based services touch more of our online lives. There are no easy solutions; rather, a conversation has begun — a negotiation, really — about who owns this information, who owns the right to aggregate it, and who owns the inferences that can be drawn from those aggregations. If you're working in this area, you're a part of that negotiation whether you like it or not, and you should be thinking about how these issues affect your work.

Are you in the fashion business? Design and even fashion are important parts of personal devices. Many of the devices making their way to the Internet will be very personal, such that good physical design will be essential. Consider the recently announced Amazon Kindle e-book reader: how many of its early reviews commented — and not positively — about the device's physical design? Getting this right is essential to a product's success. Of course, attractive doesn't mean usable — remember my earlier story about my attractive but dysfunctional iPod case. Great physical design is an important part of the story, but it's not the only part.

Don't ignore legal matters. Personal Internet devices can have a broad reach — into media tied to contentious policies about access rights, and into personal information that raises legal and privacy issues, especially outside the US. In addition, challenging liability issues might arise when these devices tell other devices what to do. Recall the microwave remote control demos — I mentioned these demos once to a lawyer, who thought about it for a few seconds and began to laugh. "Wait — you want to make it possible to turn on a heat-producing device in a home when no one is there? Suppose something goes wrong — a software bug, a neighborhood kid breaking the device's security — and a fire starts? That's a lot of liability to look at." Even in these days of overeager lawyers, I had to admit that he had a point. There are difficult issues in this area and, although you might not be a lawyer, you can't afford to overlook them.

I wish I had more answers to the questions I've raised here. Huge opportunities are emerging that combine the Internet and home life, but we won't get there until we figure out how these new devices can be tamed to fit in effectively with a very broad user base that has a small fraction of the technical expertise that you and I do. The devices must do useful things, do them in ways that their users will understand, and still respect the parts of my life that they touch. We have no choice but to take these problems seriously — it's the right thing to do, and if we don't, this opportunity to expand the Internet into new markets and parts of our lives will fail as badly as the first one did. I hope that five years from now I'm not writing another column about the Great Home Internet Fallacy of 2008.