Coffee: Systems have more to say to each other, but listening to users is still more important.
I see that the latest thing in laundry equipment is GEs Harmony washer and dryer set: a system in which the washer tells the dryer whats coming, and sets the dryer appropriately for the load thats just been washed. Im glad their engineers are keeping busy and out of trouble, but pardon me if I wait for something else before I fill out a Draper Prize nomination form.
When I saw GEs dubious innovation in the Sunday newspaper supplement from Sears, it reminded me of what someone at the old PC Week Labs once wisecracked about the hype over putting Ethernet in everything. I dont remember which of our founding analysts made the comment, but it went something like this: "Calling everything Ethernet-compatible is like calling all your appliances 117-volt AC compatible. Yes, they plug into the same connectors, but do they have anything to say to each other?"
Well, here we have a pair of appliances that arguably have enough common interests to get them through a dinner dates worth of conversation. But heres whats wrong with the idea, which is also the reason why Im taking your time to talk about laundry hardware: because this is a classic example of assuming that if a system can talk to itself, it should, instead of finding better ways to get input from its user.
In the specific context of laundry, I rarely do a simple wash/dry cycle: there are khakis that I want to hang from their cuffs to dry, there are dress shirts that get just a brief tumble before I put them on hangers, and so on. The Harmony premise is that users are predictable, when were not. And when I put my laundry into the dryer, Ill probably have to either bypass or disable its attempts to anticipate what I want.
I support the idea of systems that learn from users, but not at the price of making it harder for users to do the unexpected. I dont want to spend extra keystrokes--or even extra mouse clicks--telling dumb machines, "No, nice try, but thats not what I want." Im slightly interested in machines that try to figure out what I want by watching me, and that try to offer me what Ive wanted in the past--but not if the price of that cleverness is to make it more difficult for me to form efficient habits.
Heres what happens when someone has a superficial idea of how to build an adaptive system. I write a document; I save it to my hard disk; I save it again, without any changes, to a floppy disk so that I can keep a backup copy somewhere else. The next time I want to work on that document, I go to my "recent documents" menu and click on the document name--and I get an error message telling me that the floppy disk is not available. And I have to do something else, something out of the ordinary, because the system tried to learn by watching me--but didnt really understand what it was seeing.
As IT developers have more opportunities to feed information directly from one system to another, its going to be a challenge to think about when this is a cool but actually counterproductive thing to do--versus when its a genuinely useful improvement in making the user experience less complex. Stay focused, not on what users do today, but on what they would do more if the system made it easy. Thats how to develop technology that takes the prize.
What do you mutter to yourself about IT "innovation"?
Peter Coffee is Director of Platform Research at salesforce.com, where he serves as a liaison with the developer community to define the opportunity and clarify developers' technical requirements on the company's evolving Apex Platform. Peter previously spent 18 years with eWEEK (formerly PC Week), the national news magazine of enterprise technology practice, where he reviewed software development tools and methods and wrote regular columns on emerging technologies and professional community issues.Before he began writing full-time in 1989, Peter spent eleven years in technical and management positions at Exxon and The Aerospace Corporation, including management of the latter company's first desktop computing planning team and applied research in applications of artificial intelligence techniques. He holds an engineering degree from MIT and an MBA from Pepperdine University, he has held teaching appointments in computer science, business analytics and information systems management at Pepperdine, UCLA, and Chapman College.