Maybe If They Hadn't Hyped It I'd Be Kinder

I have mixed feelings when I recall my first bicycle. One might expect that I would have dropped cycling not long after taking it up; that first bicycle had no seat.

This segues to today's topic, the Segway. With such an introduction I hope you see where I'm going here with a review for a product I don't even need to use to know it's dumb.

Every morning a Segway whines down the alley outside my bedroom. When I ride home through the park in the afternoon, I often see a herd of tourists on Segways as it is apparently part of some sightseeing tour. I've seen them other places too with surprising frequency. Surprising because they are just so dumb.

As a nerdy guy, I can definitely appreciate the idea and I could see how it'd be fun to build such a thing and play with it. However, as a mode of transportation, it is extremely restricted. It pretty much is in the same transportation category as the unicycle.

This analysis is pretty easy for me because I know of another type of vehicle that has *all* of the advantages of the Segway plus many more and none of the disadvantages. Of course I refer to the bicycle. I've seen enough videos of people learning to ride the Segways to know that it's not quite perfectly intuitive. With a bicycle, you can at least learn when you're young and unbreakable but I've never seen a Segway for half-sized people. Accepting that both vehicles are reliant on a decent sense of balance and taking riding skill as given, we can move on to other issues.

I would be very nervous about how the Segway handles inevitably treacherous road surfaces. Clearly the thing is going to perform very poorly off road, especially in single track, but I find the worst riding surfaces are on bad pavement. If you jar one wheel on a bike, you'll probably jar the other one, but nothing perturbs your stability.

Then there is the long list of practical issues. Can a Segway carry cargo? Where? A kid? A trailer? You get the idea. Bikes are a pain to transport on buses and worse on airplanes, but it's at least possible. You can mail a bicycle relatively easily. A Segway weighs 48kg, about what my expedition bike weighs when loaded for winter trekking. I've never seen one locked up anywhere. Is that even feasible? It seems like a huge liability to leave it locked up all day somewhere.

Of course the real advantage that the Segway has is that you can arrive at your destination free from the symptoms of physical exertion, you know, like being in shape. With bicycles you have to expend a modest amount of energy while the Segway magically transports you. I should point out here that on a bicycle I have kept up with the Segway's rated top speed of 20kph for about 10 times the thing's 38km range (yes, in one day). That may be slightly extraordinary, but if the thing really only has a top speed of 20kph, I'm confident I can beat it anywhere, anytime on a bicycle without cracking a sweat. The reason that powered propulsion is not an advantage anyway, is that if you want to use your body to provide the motive power for your Segway you can not. On the other hand, if you're a lazy cyclist and want some or all of the motive power provided by your utility company, you can get an electric bike. Or a moped. Or at $5k-$6k, just buy a motorcycle.

Maybe Segway fans have something against bike seats. Certainly the seat interface is one of the main complaints about bicycles. I don't know about you, but standing board stiff while directly facing a persistent wind doesn't sound too comfortable to me. And if it did, I could just do what I do sometimes when I see the Segway tourists in the park. I just stand up on the pedals and lock my knees in an awkward-looking position, but one that is unmistakably that of a Segway pilot.

Ideas For Navigation Systems

I know roads. I am acutely sensitive to many things about them that normal people are not. Riding my bicycle every day gives me a chance to reflect on the dreadful condition of the streets of my neighborhood. It's not merely inconvenient for me, unless it is not too much of an understatement to say that being dead is inconvenient. There's one place, for example, where a freeway off-ramp merges into a high speed surface street going down a hill. About 18" from the edge of the road, about exactly where a cyclist would be is a parallel crack about 2 inches wide running the entire length. If you get your wheel caught in that in the wrong way at the wrong time - life ends. I play this game every day. I take road surface topology extremely seriously.

On my route I know every crack, bump, dip, pothole, storm grate, man hole cover, uneven pavement, curb, cateye, stripe, puddle, patch of gravel, oilslick, and gutter. I do a fair job of avoiding hazards, but sometimes I get caught. About a month ago I had my front wheel fall into a deep crack in the road and it twisted the bike and dropped me on the ground. Strangely I was actually going up hill at the time and it wasn't a serious fall.

Thinking of my large internal database of road surface defects, I came up with an interesting idea for on board GPS navigation systems. Although bad roads are a life and death issue with cyclists, even driving a car bad roads are a severe annoyance. Why not have the GPS systems help you drive smarter with respect to road surface issues?

Basically the way it would work is that a GPS system would need to be communicating with a central server. My friend, Seth, works for a company that has just such a product designed primarily to report back on traffic conditions. No matter what, this technology can't be too terribly prohibitive these days and soon everyone will have it trivially in their cell phone. Thanks to the accelerometers found in certain modern mass produced items like hard drives (they tell the drive to park the head when it's moved suddenly to prevent disk corruption in portable devices) this kind of technology is cheap and easy.

Basically your car would just drive along and as you hit pot holes and drainage dips, etc., your GPS would report that back to the central server which could be compiling a rather accurate profile. Eventually, you could ask your GPS computer to plan the best route for you that involved the quickest way *or* the smoothest. This concept need not stop with bumps but could also apply to curves allowing you to seek out the least curvy road or the opposite if you're testing your new motorcycle tires.

Another interesting similar idea is that you can rate your drive when you arrive. In other words, did you feel that the drive was decent? The computer could then start compiling profiles of areas that can be inferred to cause dissatisfaction. Then the computer can be instructed to produce the "best" drive. Collaborative filtering could be used so that the curve-favoring people got the same advice as each other, but the opposite advice of the people tending to not enjoy curves.

I'm also working on installing a roof rack on my car. I remember once driving with bikes on the roof of a car and hitting one of those swinging bars at a parking garage. Now I'm pretty paranoid about forgetting about my bikes and driving into something stupid. This is another area that the GPS computers could help with. You could tell it specific points that you know about (which it could share with others) and it could then remind you about the bikes. So for example if your garage is too short for the bikes, when you put them on the car, you can tell the computer that and as you approach it, it will remind you. Easy to do. Slightly harder is hooking that up with a database of all fast food restaurants so the computer can say, "I'm sorry Dave, I can't let you go through the drive through."

Captchas and the Matrix

If any have ever doubted the potency of my geekiness, my latest hobby should make it clear. I've never been a TV watcher, but I don't think that's the fault of flat moving image technology. The reason I don't watch TV is because it's typically so depressingly moronic. The internet to the rescue!

I have recently discovered that there is a lot of stuff that is quite worthwhile in the eye of this beholder. My latest trove of interesting things to watch are computer science lectures. Google Tech Talks are all on-line as well as many university lectures. Since one of the greatest things about my job is the chance to get exposure to some of the most brilliant people in computer science (i.e. some of the most brilliant people), it's a real pleasure to do that at my liesure.

Last night while researching HCI security I came across a presentation by Luis Von Ahn. Wow. This guy is amazing. He seems to be one of the key people behind (and perhaps even the outright inventor of) captchas. Whatever you may think of that from a day to day standpoint, it can't be argued that it's not damn clever. But that's just the tip of the iceberg for this guy.

His next brilliant project is espgame.org. The principles behind this show incredible insight and big picture thinking. I like that! Then today I read about a new thing this guy has which is captchas that actually are useful. The idea is that the captchas are based on scanned documents that the OCR couldn't read. Basically by typing in the captcha you are helping to enter the text of works of literature. Ok, I don't know exactly what you're labor is being applied toward, but the concept is nothing short of genius.

Basically, this guy thinks big and I loved his stats about how in 2004 humans spent 9 *billion* hours *playing solitaire*! He asks the question, what if that energy and initiative (such that it is) could be applied to something that wasn't completely useless. It's a valid question.

He also is interested in what humans are good at that is useful which computers are not so good at. These kinds of problems are interesting to me too. He wryly envisions a possible future where machines become more intelligent and capable than humans just like in the movie the "Matrix", but unlike the movie, the computers wouldn't keep humans around as an energy source; they would, quite plausibly, keep us around for the clever tricks our biologically inspired hardware can do that machines aren't good at.