The Next Generation of Geeks – Passing It On

The most interesting, and possibly the most disturbing, thing that I’ve ever done at a Tech Ed was to help judge the Imagine Cup programming contest in Barcelona in the summer of 2003. Microsoft sponsors this contest for teams of computer science students. It’s entirely up to the students to choose a problem, and then design and implement a software solution to it using the latest Microsoft technology. Besides having fun and hopefully learning something, the sixteen finalists get a free trip to Tech Ed and the top three teams share $40,000 of prize money. All teams retain ownership of their inventions.

Eight of us judges, three from the US and the rest from universities around the world had the job of ranking the contestants. Sixteen teams had advanced to this round; two from the US, the balance from Europe, Asia, and South America. All displayed good programming techniques and had surmounted their technical hurdles with ingenuity and imagination. But to this day I still cringe at their lack of problem definition skills, the way they chose which problems they’d surmount. Almost everyone worshiped technology for its own sake, my bugaboo as you’ve probably learned by now, and yours too.  I think a faculty adviser who allows that is letting his students down severely. I shudder to think of these guys getting unleashed on the world in their current state.

For example, the German team (dressed in blazers and ties, the only team so attired) had tried to solve the problem of unchecked password proliferation – requiring different user IDs and passwords for all your many web accounts, as I wrote about in Chapter 4. Fine, that’s a real and nasty problem, but their attempt at solving it demonstrated that they didn’t understand why it’s so nasty. Instead of typing in a password to gain access to a merchant’s web site, you’d call the team’s authentication server from your cell phone, and the caller ID system would authenticate your identity to the merchant web site. It’s the single sign-on idea that I described in Chapter 4, using a piece of hardware that they hoped the customer already has in his pocket.

They didn’t think from the standpoint of their customers, neither the users nor the merchant web sites. To buy a book from amazon.com, which would you rather do? Type in a password as you do today. Or  remember where you last left the &%^$ phone, go get it, charge the battery (finding the charger if necessary), turn it on, wait for it to boot up, dial a ten-digit number, then enter another twelve-digit order number. When I asked them why a user would buy from a site that made them go to all that trouble, they said, “Because it’s safer.” They don’t realize that ease of use is the main problem in any security system. If users find it too difficult to unlock a secure door, they’ll prop it open and disable the alarm sensor, or they’ll find a different pathway that isn’t locked, or they won’t go through it at all. The team also never thought about the chicken-and-egg problem inherent in their design – which sites would pay for their authentication service until a critical mass of customers register for it, but which customers will register for it until some useful sites accept it? This was a classic example of inventing something without stopping to ask why it would be a useful thing to have. It’s about what I’d expect from a country whose language gave us the crossover word schadenfreude.[1]

The team from Taiwan had developed a system that identified musical tunes, or tried to. Did you ever wake up in the morning with a tune running through your head that you just can’t identify? Me neither. But this team wrote a program that allowed you to dial a phone number and hum the tune into the computer on the other end.  The program would figure out which song it was, play it for you, and offer to download it as a ring tone for your cell phone if you punched in your credit card number. It worked OK for the pretty girl team member with the nice voice, but somehow not for me. Maybe my jet-lagged croaking scrambled its circuits, or maybe its database didn’t include “Three Day Drunk Looking For A Place To Happen.” (Fish Head Music, 1999, www.jim-morris.com. Hey, obscure tunes are, by definition, the ones you’re least likely to be able to remember. I’ll bet they don't have his 2005 smash hit, “Booze Is The Duct Tape Of Life”, either.) Again, I found the technology somewhat interesting for its own sake, especially the signal processing mechanisms (OK, I’m a geek). But I’d be lying if I said that fear of waking up with an unidentifiable tune in my head keeps me awake on many nights.

Worst of all, to my mind, was the team from Singapore. They said that waiting in a checkout line at a Singapore supermarket can take 20 minutes or more. How that can be true in hyper-efficient Singapore I don’t know, but I freely concede they ought to know that sort of thing much better than I do, and I agree that eliminating a 20-minute grocery checkout line would be a great boon to humanity. They’d worked incredibly hard, probably the hardest of any of teams. They’d written some excellent technical software, and even partnered with the electronic engineering department to develop some new hardware they needed to run it. And they’d come up with the sort of horrible monster that desperately needs drowning at birth.

They’d taken the idea of self-checkout, which I detest and refuse to use when I encounter it at Home Depot or my local supermarket, and put it on wheels, where I detest it even more. They mounted a portable laser scanner onto a shopping cart. When you pick a product off the supermarket shelf, you’d scan it before tossing it into your cart. To keep you honest, the cart would have a weight sensor under the basket, and the weight would have to tally with the items you had scanned.  To check out, you’d slide your credit card through a reader on the cart. You’d redeem coupons by placing your infrared-equipped palmtop or cell phone against a smart kiosk display, then holding it under the scanner.

May the gods protect you and me from this awful use of technology for its own sake. In addition to making us do work that someone else now does for us, every user would have to learn and deal with the peccadilloes of a complex technical implementation (see chapter 1). For example, you couldn’t put your purse or coat or packages from other stores in the cart because that’d throw off the weight sensor. If you’d changed your mind after scanning something into the cart, you’d have to scan it again before putting it back on the shelf, and make sure to press the credit button. Toddlers would no doubt be fascinated by the laser scanner, (“Lookit! OUCH! My eye!”) and so would school-age children (“Cool! A real ray gun! Lemme zap that old lady over there. POW!”), not to mention teenagers (I’ll leave that to your imagination).  You’d have to read directions, no doubt written in a language that occasionally resembled English, just to buy a candy bar. And of course each store would use a different type of cart, so you’d have to learn them all. I can think of no way short of an all-out first and second strike nuclear exchange to do greater damage to the world retailing industry.

A grocery delivery service such as Peapod (www.peapod.com), where you order groceries online and they get delivered to your door, would be a far better solution to this problem. It not only saves the checkout time, but also the shopping time and the transportation time, probably an hour or more each week, and reduces road traffic and energy usage overall. It works profitably in rural Massachusetts where I live, so it ought to make even more money in densely populated Singapore. If you wanted to preserve the in-store shopping cart experience, then radio-frequency ID tags on each package would allow you to tally up a whole cartful simultaneously by pushing one button at checkout instead of individually scanning each item[2]. But the cart-based laser scanner for individual items is a giant step backwards.

I placed the Singapore team in the middle of the pack because they had worked harder than anyone else. The other judges, less jaded than I am, or perhaps dedicated more to programming than to usability, placed them higher, and they actually tied for third place and won some money. If I’d known that they were that close to a prize, I’d have whacked them harder to keep them out. In case you’re wondering, I told them all of this when the contest was over.

Fortunately the contest contained one absolute gem. Tu Nguyen’s bilingual data entry system for waiters in his family’s Vietnamese restaurant blew me away and was the judges’ overwhelming choice for the $25,000 first prize. The chef, his father, spoke little English, and bilingual Vietnamese waiters are hard to find in Omaha, Nebraska. If Tu was ever going to escape the role of go-between, he had to find some way to make the restaurant work without him. So he put the restaurant’s menu onto a Pocket PC, essentially a paperback-book sized computer with a wireless network (See Figure 6-3.) They’re cheap enough ($200 and falling) so that each waiter could carry one. Waiters would enter the customers’ orders in English and send them over the wireless network to an ordinary PC running Tu’s server program. This contained a translation table that converted them into Vietnamese and printed them in the kitchen.

In addition to changing languages without human translation, orders got to the chef more quickly because waiters didn’t have to walk to the kitchen or queue for a single entry terminal. The wastage rate and flow disruption due to misunderstood orders plummeted, and the angry customer saying, “This isn’t what I ordered, you [racial epithet]” became a thing of the past. Faster order processing led to faster table turnover in the small restaurant, increasing their sales per square foot. It became possible to reliably tailor a dish to the diner’s taste, adding extra broccoli or omitting the cilantro, which increased customer satisfaction, and hence return visits and word-of-mouth publicity. It also improved their reviews in the local media, as professional reviewers always ask for special orders as part of their investigative process.

Rather than using technology for its own sake, Tu focused with laser precision on the business problem he was trying to solve. This small, simple, beautifully tailored system dramatically lowered the friction of almost every aspect of running this restaurant. Replacing the translation table would allow it to work from any language to any other language, say, entering orders in Spanish and having them come out in Japanese.  It’s under consideration for use in Omaha’s new sports stadium, and I imagine that Tu will have his choice of job offers on graduation, if he doesn’t run with it himself. Chalk up another immigrant success story for the Land Of Opportunity.

I tried very hard to explain to all the teams that had misunderstood. Computing isn’t a technology field any more, it’s a people field. Of all the things that I did at Tech Ed, I think, I hope, that my influence on these impressionable students, rewarding careful problem definition and punishing technology for its own sake, might just be the biggest contribution I could ever make to the industry – apart, of course, from the book you are now reading.