Auto-renamer for the Audiogalaxy Satellite so your files will be named however you want. Actually it just needs to be tested with the new UI code.
PHP personal portal. And weblog (to supplant Blogger, once I decide to spend the $$$). And whatever else I want it to be.
An mp3 databasing scheme in partes tres:
Scanning my local disks for mp3s and collecting the ID3 tag information, and then inputing that into a MySQL database. Looks like it will be in C++ with the MySQL++ API and the id3lib ID3 tag manipulation library
Creating a web-interface (in php, of course) to said mp3 database with the ability to search on various terms and download mp3s from the library.
No third one. I just wanted to say partes tres ("Omnia Gallia in tres partes divisa est....").
PJ Harvey in Concert. Granted, it's Houston, and it's a Monday night, but oh well.
Why don't they have the timer on windshield wipers connected to the speedometer? That way, when you go faster, the wipers go faster. When you slow or stop, the wipers go slowly. Seems like a no-brainer to me.
Then what's with the word speedometer? I mean, c'mon. It looks like some invented 50s product, the amazing Speed-o-meter! It's such a silly word.
I need to start carrying around flyers for Austin Light Rail in my car. Next time I am in a traffic jam (like this evening), I will walk up and down the stopped lanes of traffic handing them out. The only problem is that I won't have anyone there to inch my car along.
It has become apparent to me in the last few years that there is a significant life lesson I never learned. It was never indicated to me that there even was a lesson to learn. I was never taught how to deal with ambiguity and uncertainty. Perhaps it was hinted at as part of other studies, but there was never an explicit focus on questions like "What if you're wrong?" When it comes to teaching "life lessons," we too often assume perfect information. In daily interactions with other people, there is so much ambiguity. So much is unsaid or unclearly said. And yet we generally act based on this superficial knowledge we've gathered; we take people's statements at face value rather than looking beyond them. And even when we make inferences based on them, they are either very tentative and tenuous in our own minds, or have the power of fact, with no middle ground. Furthermore, we are a very trustworthy society. This may sound counter-intuitive at first, but consider the information you gain daily. Newspapers, websites, personal conversations.... How often do you question what you read? Some sources (*cough* Slashdot comments) are usually taken with many grains of salt, but in general, you trust what you hear and see. There is nothing necessarily wrong with that, but there is a tendency to think that only what is stated explicitly is all that is there. Furthermore, these sources will be wrong from time to time. In statistics, there is the ready acknowledgement that gathered data might not reflect reality. Somehow that lesson hasn't escaped into the real world. There needs to be an structured way of teaching people how to accept and deal with ambiguity.
I want an email client that changes my message settings based on rules. For example, when I post to a newsgroup, I want to make my email address ketan@REMOVEketan.org or something like that. I want to be able to easily switch sig files so I can have one for work and others for personal mail. That way possibly offensive quotes ("Pickup Line #11: Chick do now.") don't get sent to, say, a prospective employer. I think every significant piece of software should have a scripting engine so I can make it do things the way it should have done them in the first place. Python would be a good start. Emacs had the right idea (although they might have considered practicality when picking Lisp as their language of choice).
Virtue is its own reward. That's what they say. I can't say I disagree. It's almost a tautology. If you had some incentive to do the right thing, then you're doing it for the incentive, not for the rightness of it. The problem isn't knowing what the right thing is, as some people have said. Nor is it in actually doing it. Both those things are easy. The hard part is living with the consequences. Whenever you're tempted to do The Wrong Thing, you have a million reasons to yield. But you only have one reason to do The Right Thing, and that is because it is right. And you'll never forget all those reasons you turned your back on. It's easy to do anything. The hard part is living with the consequences.
It occurred to me to purchase theantitruth.com or theprodrug.com and put up a parody of theantidrug.com website. But that would be a lot of work, and do I really need John Ashcroft's goons coming after me? I have enough trouble finding a job as it is.
But yes, those domain names are free, at least they were as of yesterday. And to think that just yesterday I read that the campaign had purchased alternate versions of the domain name to stave of such parodists. Clearly they didn't think very hard, as those two were the first I thought of. I'd say they need me on the payroll, but even I don't need money that bad. I'd probably sell drugs first. Oh, wait....
More on people missing the point on copyrights and patents... It's an incentive, right? So let's see how that works. Suppose you're a potential author/inventor. You are considering taking time off your job to write a novel, or tinker full time on that funny-looking scooter in your garage.
Now suppose this could go on in any of four hypothetical worlds. In the "information wants to be free" world, the minute you show your book or your toy to someone else, they can run out and sell your book without a dime to you, or manufacture an identical scooter. At this point, unless you're really invested, you're just going to give up.
Now suppose you have 5 years of exclusivity, in the "as short as possible" world. That might just be long enough for you to make the cash to justify your year off. So you'll go for it. But what about Giant PharmaCo? They want to make a new drug that will cost them about $1 billion to develop. They figure they can make about $400 million in revenue per year, of which $100 million is profit. If they can only go for five years, they make back only half of their original investment. Clearly that's a losing proposition.
So suppose the term is extended to 20 years, in what I call "a period long enough to sound right in my contrived example" world. Now you, the individual creator, are doing quite nicely. And Giant PharmaCo makes quite the tidy profit off their new drug, so that 20 years later, when it falls into the public domain, they're well recompensed for their investment.
Finally, there's the "intellectual property is just like any other property," wherein there's no expiration, or at least no effective expiration (a work created today has life of author + 70 years of copyright protection, which is easily a century in the general case, at which point I'll be long dead, so it benefits me not at all). Given the trajectory that most works follow, namely, the majority of their profit comes in the decade or two after initial publication, the extra 80 years of profit doesn't really affect anything. Furthermore, profit made 25 years from now is pretty unlikely to affect your decision to invent now. If your work does well, you'll have more than enough money from the 20 years. If your work does poorly, it's not going to make a difference whether you own copyright on it for a year or a century. Very few will fall between those two extremes. So as an incentive to me, the individual inventor, there's very little difference between a term of 20 years and a term of 100 years. And as it states in the Constitution (as quoted previously), this is "to promote ... progress." If it doesn't make a substantive difference to me as an inventor to extend my term 80 years, then it's not promoting progress, and so shouldn't be that way. Corporate interests might feel differently, but the whole point of copyright was to protect the weak from the strong by preventing the latter from stealing the ideas of the former.
I realize I'm mixing copyright and patents, but to an extent, they can be treated the same way in an example such as this.
So, this is the way pay-for-play music should work:
Free - access to a streamed, medium quality version of the song five (5) times every two (2) months (per song). This is key. You must be able to sample before you fork over money, and not just once. Listening to a song five times in two months seems the right balance. If you really like the song, then you're going to want to listen to it more than once every 12 days from your computer, so you'll pay for one of the other options (see below).
$0.35 - streamed access to same quality file as above, but unlimited listens.
$0.60 - streamed access to file of quality comparable to 128kbps - 192kbps mp3 on demand with unlimited listens.
$1 - download of a perfect quality, uncrippled file with which you can do anything you want short of redistributing. You can burn it to an innumerable number of CDs, listen to it forever no matter if your subscription/membership is no longer current, etc. Basically a WAV file or some other lossless sound file.
$4 - access to the song in same format as mentioned above, but with each song broken out into the individual tracks to facilitate remixes, etc. This would be a niche market, but I'm sure there are enough out there who would want this to justify it.
Naturally, in that last case, there would be restrictions on redistribution. I have no problem with that; my problem today is that the record companies aren't giving me what I want no matter how much I'm willing to pay. They're not serving me, their customer.
And of course there would be bulk options. Buy 10 songs and get 10% off; buy 20 and get 15% off. If you buy 50 or more songs, there should be a minimal additional cost option to have them instead burn the songs to a CD or DVD and mail them to you (saving both sides the bandwidth). With a dozen distribution centers around the US a la Amazon, your songs would come to you in a day or two, and they'd be exactly what you wanted. Or they could partner with Best Buy and similar retailers to do this packaging, but it would be difficult to get the more obscure music that way.
As you can tell, I miss my Audiogalaxy. I watched "The Bourne Identity" (boring, generic) yesterday. A couple of the songs during the movie interested me, but I knew there was no way I'd be able to get ahold of them short of buying a soundtrack that I knew would have no other songs on it to interest me (as I'd just heard it during the movie). If any of them were on any albums, then I would have to buy a whole CD of material when I only had verified that one song on it was something I wanted, an even worse case than the soundtrack.
Ok, so I've been thinking about this a while. I want copyright to be shorter. I do like the idea, and I do see the need and want its protection. Then it occurred to me. I just don't have to use all of it. So. Let this serve as announcement: Everything that I copyright will fall into the public domain exactly 28 years after first publication. That was the maximum allowed by the initial copyright term granted by the United States Congress in the 18th century, and it is enough. Some obvious caveats, of course. This doesn't apply to work I do for someone else, because I don't/won't hold the copyrights; they get assigned to my employer as part of the standard contract. Plus I realize I haven't published much that anyone would want to appropriate, but I'm still young. Maybe someday I will. Right now I will commit myself, though. No matter what happens in Eldred v. Ashcroft, I will only take 28 years from the commons. No more. Possibly less. I urge you to do the same. No matter how small you think your contribution to the public domain is, commit yourself. All writings. All weblog posts. All photographs. All code. All correspondence (excepting private and personal). All graphics. All video. All musical compositions and recordings. And everything else that can be copyrighted that I haven't thought of. 28 years.
I just returned from Colorado Springs. My job sent me to this conference on "community networking." i.e., non-profits, municipalities, and other non-corporate institutions looking to use the Internet and technology to improve their services and offer new ones. It was more of a meet-n-greet type of thing than a "do stuff" thing. I got to stretch my schmoozing muscles. I was by far the youngest and least-qualified person there, probably because I was a sub for my "boss."
The conference was at the Garden of the Gods Club (not a hotel). It was pretty swank. My suite was bigger than my apartment. Hell, the bathroom was almost bigger than my apartment (not really). The back balcony faced the Garden of the Gods (a rock formation) and then Pike's Peak beyond. I'll upload pictures shortly. The club was one of those exclusive ones that always makes me feel slightly uncomfortable. Conservative money. I have a visceral (anagram of "claviers") reaction to places like that. Eh.
I got window seats on both flights back. Colorado is very brown right now. There are dry streambeds all over the place. Most of them were invisible, but it was quite clear where they were because of the threads of trees snaking across an otherwise barren landscape. As the terrain changed, there were more and more canyons carved out by these rather small streams. The brown landscape was broken up by these rugged canyons which were filled with trees. It was very fractal. On the DFW-Austin flight, I managed to figure out where we were relatively early. I think I identified I-35, then Waco. I picked out Round Rock, 620, Mopac, 183, Braker, 360, and 290 as they appeared. It was cool.
Airplanes are optimal places for LANs. What they need to do is mount 9 cameras on the bottom of the plane, looking forward, backward, left, right, and in between, which would be served via streaming media on the LAN. Airplanes would make a good testbed for a resilient network model that would be smarter about security threats (like business travelers carrying Code Red on their laptops). Then there's the coolness factor of playing Doom III against a friend, or streaming a DVD so you both could watch. Yes. Gigabit ethernet on airplanes. Genius, I tell you.
Here's a couple of half-baked product ideas I've had recently:
Bike leg protectors - to keep the lower trousers leg from getting caught in a bike chain. I figure a rectangular piece of nylon with velcro. Wrap it around your lower leg over your trousers bound and covered. Maybe with optional shoe covers to protect your... shoes from mud and dirt.
Ceiling-mounted storage - Retractable storage bins mounted in your ceiling. These would have telescoping rigid supports to minimize horizontal travel. When you want it, reach up and pull it down. A ratcheting mechanism holds it in place when you release it. When you're done, pull down and it pulls back up, like window blinds. I'd use it for a notebook in my office and for a spice rack at home in the kitchen.
I know I have a couple more, but I can't remember them right now. Maybe that's for the best.
I forgot one of my other ideas. Are you ready? Really? Ok... six-door passenger sedans. No, seriously. I figure, if you have 4 kids like mute, you're currently restricted to SUVs, station wagons, and minivan (or vans, I guess) if you want to take your whole brood somewhere. Even then, it's inconvenient because you need to climb to get to the back bench. So I figure, take an ordinary 4-door car and stretch it. Maybe do one of those neat clamshell door things. I tried photoshopping a Honda Accord but I quit because it was too hard. My Accord is 16 feet long. Adding 4 feet for another bench of seats is a lot, but that's just a raw conversion. I bet they could do it with half the extra length. Maybe it would have to be a station wagon so you could sacrifice some trunk space. But definitely 6 doors.
The NY Times has been doing a continuing study on the progressive depopulation of the American Great Plains. I've also been reading about how bison herds are slowly growing, with an increasing, though niche, market. And then there is the looming problem of what to do for energy. I've combined all of those ideas into a single plan for action. First of all, let's forget about trying to stop the depopulation of the Plains. There is no reason to want people to live there if they don't want to. Second, let's encourage the growth of free range bison populations in those vast empty spaces. Then encourage the use of bison meat as a beef substitute. Third, build large numbers of wind farms and nuclear power plants. The former would be where it is, well, windy, and the latter in places that have no use for anyone. The electricity thus generated would be used to split water into oxygen and hydrogen. The oxygen would just be released into the atmosphere, while the hydrogen would be transported cross-country via either high-speed train or pipeline to the major inhabited areas for use in power (re-)generation. I'm sure that can be made as efficient as current transmission across power lines, which can lose almost half its energy in transit. The wind farms and nuke plants would run continuously. Using hydrogen as an intermediate "battery" would allow us to decouple production and consumption, and thus gain further efficiency through operation at optimum load and maximization of capital investment (like a hybrid automobile). Having the nuclear plants out in the middle of nowhere would make it considerably easier to secure them against the (exaggerated) terrorist threats on everyone's mind these days. Using high-speed rail for transport would also help to subsidize the development of transcontinental high-speed passenger rail (a pipe dream I'll save for another time) that could potentially allow you to travel from New York to LA in 12 hours. Which would also be nice for the people stuck working at the nuclear plant out in the middle of nowhere. Crazy talk, I know, but sometimes I just get to thinking about things.
With guests coming in from out of town looking for free beds, I wonder if there's any market for a co-op. i.e., anyone who has a large event such as a wedding or reunion can ask other members of the co-op to put up their guests. People who do get credits for asking others to do the same. Just a thought.
An essay doesn't begin with a statement, but with a question. In a real essay, you don't take a position and defend it. You notice a door that's ajar, and you open it and walk in to see what's inside.
I think one of the reasons that I find Malcolm Gladwell to be so interesting is that he doesn't write propaganda pieces as much as non-fiction stories. He takes you on a journey through an issue, depositing you at the end not with a specific opinion but with a sense of having learned something.
Writing is an interesting skill. I was always a serviceable writer, but never excellent. One of my goals with this weblog has been to develop that skill. I think of my longer posts as being like essays, having escaped the definition of essay I learned in high school. The first goal was to become comfortable writing. Like many other things, it comes more easily with practice. I may not be Malcolm Gladwell, but I'm better than when I started. Like Paul Graham, I also use writing as a way to discover something. Generally, what I discover are my own thoughts, but in doing so, I find evidence and perspective from other sources. Writing forces a certain rigor on my thoughts; inconsistencies and bad logic are much more apparent in writing than in the nebulous and ephemeral thoughts swirling in my head. Making yourself write your ideas down and mull over them forces you to reexamine them; in doing so, you improve yourself. You eliminate the bad ideas and add to the good ones, but, even more importantly, you train yourself better to do so, so that the next time, you waste less time on bad ideas and develop more good ideas more easily. Like Paul Graham, I feel that we don't emphasize composition enough, and that we have lost something important now that rhetoric is no longer an independent field of study. Writing essays is a way to discover a point of view while also training myself in the art of structuring and articulating a position; persuasion, while good motivation, is a secondary goal. It is enough for me that it makes me smarter.
I was having a discussion with a co-worker about the frustrating rate of change. My take on it was that change is always far too slow for our liking, but better ideas do eventually win. It's a fundamental issue with human nature. People can push really hard for a while. People can give up. But people can't, in general, push steadily on an issue for a really long time. To pick a nerdy analogy
1
, it's like a rocket vs. an ion drive. A rocket pushes really hard for a short period of time, coasts for a long time, and then decelerates hard at the other end. An ion drive, on the other hand, accelerates slowly and steadily for half the journey, flips around, and then decelerates for the second half. Over a long enough voyage (i.e., to Jupiter, not the Moon), the ion drive will get there faster. You will get a lot more by pushing steadily and undramatically. If you don't, those who oppose you can just wait for you to burn out and go back to business as usual. You risk pushing too hard and provoking a backlash. Or your violent revolution might actually succeed, but leave you worse off than before
2
. Not to say that violent revolution is always unjustified
3
, but rather that people are far too eager to choose that strategy. If you look at the people who have been most successful in their lives, the vast majority of them have been tortoises. They've been working steadily at it for a long time
4
. In the end, the most important thing is not intelligence or ability, but rather, not giving up and not stopping
5
.
1
The more pedestrian one being the tortoise and the hare, of course.
4
Notice how people still pay attention to Al Sharpton?
5
Even in showbiz. Look up some actors and actresses (especially in tv, where the same faces show up over and over) and see how much crap they did before they had a hit or you ever heard of them.
Malcolm Gladwell is my hero. Not just mine, either, according to Fast Company. You may know him from his book The Tipping Point. I liked the book, but I tend to favor his New Yorker articles. They're more diverse. Business types (like Fast Company) have focused on business-y things, but that's just an application of the general theme: think about things that people take for granted. "The Tipping Point" is one example of that theme that's easy to apply to business, but it's just one idea.
It's less important to understand or even agree with any given idea that he proposes. What's important is what's behind it. Think about how things work and why things are the way they are. Don't assume you know, because you probably don't. Apply logical thinking to things that people tend not to think about logically. Malcolm Gladwell is successful because he does these things and then writes about it in a nice, neat article. The ideas matter less than the process, and that's what I need to explore.
I may be better at this sort of thinking. It certainly comes more naturally to me; it's hard for me to imagine any other perspective. To be successful in my professional life, I need to figure out those things that I do that are unique and find a way to make it useful. That's where software and operations research come in. To do both well, you have to look beyond your surface understanding of some process. You have to trace it backwards to see where it all comes from. You have to understand how the parts fit together. You have to have an instinct for which ways are better. And you have to have a certain restlessness that makes you continually motivated to do things better. That's me. That's a valuable and satisfying niche, and I need to break into it. That's my theory, at least. I guess we'll have to wait and see if I'm right.
On the subject of milestones, I want my odometer to ding or otherwise indicate when I reach a particular mileage. Last week, Jessica's car hit 133,133 miles. We knew it was coming, but it slipped by us unnoticed. I want to be able to tell my car, "ding at 133,132" so I notice. And I guess you could use it for, I dunno, useful things, like oil changes or maintenance or whatnot.
I really liked calculus when I learned it in high school. It was almost a transcendental experience. It was the first math class that I took that explained the world. As much as I loved calculus though, I can no longer say it's the most important math. I am now convinced that a good working knowledge of probability and statistics is essential to making sense of the world. I couldn't tell you the difference between a Gaussian distribution and a Poisson distribution. Nor could I tell you how to calculate a confidence interval. I do know that random events can often be aggregated into a distribution. I do know that you can never know anything for sure from a subset of the data and how that uncertainty varies. I know that coincidence is inevitable. I know that good strategies often don't succeed the first time, and that bad strategies often succeed for a while. I know that, in the long run, just a couple percentage points make a huge difference. All of these things make me richer, safer, and happier. Not knowing those things will make you poorer, less safe, and more frustrated. It's not about making really smart decisions, just avoiding stupid ones. Unfortunately, probability and statistics tend to be taught by the dullest, driest instructors around. That's really too bad because it could be really fascinating if taught right. I know that I didn't learn all I could have because I had a boring professor. There is hope out there. There are books for people who hate statistics. You (and I) should read them. We'll be better people for it.
A while back, I had the idea of photo maps to help visualize directions. Amazon has added something like this to their A9 web search to show you a particular location. It just shows your destination and the area around it rather than a point-by-point path from A to B, but it's still pretty slick. Just goes to show you that good ideas are inevitable.
gentrifugal force, n.: the apparent force that flings lower income people out from the center of the city as a rising tide of higher income people moves into the previously-neglected-but-now-newly-revitalized area.
I think it's kind of neat that you can get free wireless Internet access in cafes and restaurants. Not all that useful, but neat. What would be much more useful to me would be to have free wireless access in waiting rooms, like doctor's offices, dentists, or at the auto shop. I would definitely steer my business toward a mechanic that gave me more to do while waiting than watching "Days of Our Lives" or reading the June 1998 issue of "Bass Fishing" magazine.
I've worked for multiple software startups. It's inevitable that I'll fantasize about starting my own one day. One thing I've been giving some thought to is equity. The common tech startup route is to grant options. That's ok, I guess. Last year, I read about the method used by the large defense contractor SAIC. The company is completely employee-owned. All shares trade on an internal exchange. That's pretty cool. I would want to "improve" on that model by allowing the price to float (SAIC's is set by the board of directors) and allowing trading more frequently than quarterly. In addition, I would have a dividend. An employee's equity-based compensation would be dependent less on the valuation of the company and more the income from the dividend. If I wanted to get really fancy, the internal exchange could support calls and puts. I haven't yet decided whether new employees should get grants of shares or options or nothing at all. Bonuses would be partly as cash and partly as equity. Employees who leave would have a grace period during which they could sell their shares, probably in the range of years to allow for stability. Naturally, there would be a dividend reinvestment program (DRIP). There might have to be exceptions for outside investors (ideally, not venture capitalists), but I still think the ownership should be time-limited as with other non-employees. It's a half-baked idea. I'm going to stop typing now.
This is something I would send Google if their submission form didn't keep accusing me of sending malformed HTTP headers (as if I would do such a thing):
I'd like to be able to ask for directions from one place to another based on what the name of the place is. I'd like to be able to say "Tell me how to get from Hollywood Video to Lowe's in this zip code." If there are multiple matches, let me pick a particular Hollywood Video store. I can currently look up the addresses of Hollywood Video and Lowe's and then get directions from one to the other, but I'd like to do it in fewer steps.
Now, let's suppose I don't know how to get the Hollywood Video in the first place. I need multi-step directions. I need to know how to get from my house to Hollywood Video, and then from Hollywood Video to Lowe's, and from Lowe's back to my house. That's 3 separate legs. Right now I have to do those as three separate queries and three separate sets of directions. It's not climbing Everest, sure, but it could be easier.
Now, to make it a little more useful, let's say I want to do this all on my way home from my job. Then I need directions from Work -> Hollywood Video -> Lowe's -> Home. It's the same thing to compute, but the interface would be different. A wizard suggests itself, but good feature request describes the problem and leaves the solution to wiser heads.
Even the above is too much thinking for my little head, though. Maybe I don't care whether I go to Hollywood Video or Lowe's first. After all, Lowe's might be right next to my work, and Hollywood Video might be right next to my home. Going to Hollywood Video first would be rather inefficient. If I could say, "I want to go from Work to Home, stopping at ANY convenient Hollywood Video and ANY convenient Lowe's along the way," and have it pick the most efficient combination of store locations and route, well, that'd be swell.
Then there's the Google Becomes Skynet version where I don't say what store I want to stop at, just that I want to rent a movie and buy some plywood on the way home. That'll take a lot more work, but, in the end, describes what I'm looking for without cluttering it with unnecessary details.
I'm trying to change my posting style to be more appropriate to the format. I have over 30 drafts of lengthier posts in various stages of completion, some of which are a year old. I tend to think of this weblog as being like a book, where you present fully-formed thoughts, but it's not. If you present a half-baked idea in book, you're in trouble. Me, I can just say, "hey, that thing I said yesterday was dumb." That's not to say that baking things longer is a bad idea, just that I want to move the slider a little bit. I shouldn't assume that I need to fully develop an idea before publishing it. In some ways, that might be counter-productive. I figure it's more valuable to see the evolution of an idea than to be presented it as a fait accompli. It's kind of like showing your work in math class. The result is useful, but the process can be just as important.
I'm not too concerned about the recent jump in gasoline prices, but it does give me an idea: prepaid gas cards. Instead of being denominated in dollars, however, these cards would be denominated in gallons of a particular grade of gasoline. Gas companies would sell, say, 100 gallon cards, with the price being set at whatever the current price of gasoline is. For example, the Exxon near my house would sell one for premium right now for $299. They wouldn't sell them constantly, nor would their be an infinite quantity of such cards. Gas stations would sell them when they thought the price of gasoline was likely to drop, while customers would buy them when they feared it would rise. In effect, it would bring gasoline futures trading to the retail level, allowing both buyers and sellers to hedge against drastic price changes.
The New Orleans disaster is clearly a tragedy of great proportions. However, we would be setting ourselves up for another tragedy in the future if simply rebuilt the city. Consider: much of New Orleans is at or below sea level. Consider: New Orleans is bounded on the north by a large lake and on the east by the Gulf of Mexico. Consider: the largest river in North America flows through New Orleans. Consider: New Orleans is frequently in the path of hurricanes. Consider: hurricanes will likely increase in power and frequency due to the warming Earth. Add all that up, and it sounds like rebuilding New Orleans is not something we would have to do just once. There's a legal doctrine called "coming to the nuisance" that is relevant. Basically, if you move next to a pig farm, you can't complain about the smell. If people insist on living in New Orleans, knowing the dangers (especially now), then they should be on their own. If it's worth rebuilding New Orleans, then it's worth doing it right, which means doing it somewhere else.
Let's say you're shopping online, and you see a product you like. Unfortunately, it's too expensive. You know that particular retailer has sales from time to time, but you're not going to remember to keep checking the price to see if it's come down to your level yet. So you pass and you move on. I want to be able to go to a store and say "I'll buy it for $40." I give them a commitment that I will buy it if the price drops to $40 (with a built-in expiration, of course). I have no guarantee that will ever happen. Nor do I know when it will happen. I just make an offer. If the merchant is willing to sell it at that price, they'll push the transaction through immediately. Or perhaps it allows them to unload inventory at a later date with less hassle and for more money, because they know they have a buyer. It's a win-win situation. At worst, I don't get the item for more than I'm willing to pay and the merchant doesn't sell it for less than they're willing to accept, which is the status quo. At best, however, I get the product for a price I'm willing to pay, while the merchant can increase volumes, find willing buyers, sell overstock/clearance inventory more efficiently, and take advantage of price discrimination to maximize profit.
We need a word that means sort of what physically and manually are used for when it comes to actions on a computer. "Physically copy the file to the server" is no good. Nor is "manually run a clean build."
What the credit card companies need to sell is a device that informs you in real time of an attempt to charge to your card. You then have some amount of time to approve the transaction. If you do not approve the transaction in time, or explicitly reject it, the payment is declined. You should also be able to whitelist some merchant so that their transactions are automatically approved. The query should happen as close to instantaneously as possible, so that you can use it to authorize transactions at any ordinary retail establishment. Alternatively, instead of a standalone device, it could be integrated with a phone, perhaps as just a text message exchange. I know Paypal is working on something like that, but Paypal is not supported nearly as much as credit cards, especially not in the physical world. Of course, as someone who as to work with credit cards, processors, and the bizarre mainframes that drive it all, I know it wouldn't be easy. But it's definitely an idea whose time has come.
Austin just got a car share service. Basically, a car share is a membership-based organization that allows its members to rent cars at a low rate, in addition to a monthly fee. The idea is not much different from a rental car on the surface, but it allows a considerably different mode of use, as rental car companies charge by the day and cater to out-of-town visitors and people with their primary vehicle in the shop. The idea behind a car share is to make it possible for some people to not have a primary vehicle at all.
We own two cars, which are idle on average for about at least 22 hours per day. That seems like a waste of our money, even with the inexpensive cars we have. Compare the cost of a car with the value you get out of it, and it's pretty expensive
1
. This isn't for everyone right now, but it's for some people, and it can drive change (haha). Consider someone who lives in a downtown area, with their job and grocery shopping in walking distance. Consider someone who has a motorcycle or scooter as their primary transportation, but still owns a car or truck for backup. Maybe you're a student who runs errands on the weekends, but stays on campus the rest of the time. Then there's the couple with a single car that occasionally needs a backup. Or, with a slightly different system, maybe you drive a tiny Toyota Yaris but occasionally need a truck.
I think this could combine well with a resurgence of rail travel. Energy costs are enormous for airlines, which trickle down to us. Then there are the delays, security issues, airline peanuts, cramped spaces, no-fly lists, as well as airports often being located well outside the city proper. Trains, on the other hand, are marvelously efficient, low-key, and far more spacious and comfortable. For medium length travel, say, the 200 miles from Austin to Dallas, the times required for driving, flying, and riding a train don't differ much. On a train, I can read, play a video game, or take a nap. Given the choice between riding for 4 hours of useful time
2
, or driving for 3 hours staring at asphalt, I'd definitely call the former a more valuable use of time.
I can do this now, but I have to deal with the hassle of a car when I get to Dallas. If the car share programs in those two cities are affiliated and near the train station, I get the advantage of inexpensive, environmentally-friendlier inter-city travel combined with the convenience of a car for local driving, without having the overhead of actually owning a vehicle. The cities of Texas may be less ideal for this sort of model, but I can imagine it being very useful in the urban Northeast, and its success there trickling back to us.
Of course, the real revolution will come when the cars can drive themselves, but that's a topic for later.
1
And that's with a slow train; imagine a medium fast train at 120 mph.
2
Take a $20,000 car that gets 25 mpg that I keep for 5 years before selling it for $8,000. I pay $1600 in sales tax and $500 destination charge. I drive 10,000 miles per year, using 400 gallons of gas at $2.75/gallon. My insurance costs me $200 every 6 months. That's $21,800 net over the 5 years and 50,000 miles, which is $0.43/mile.
Take a look at this trivial web application. It's just sort -fgu -t, -k2. I do something like that manually when I'm transcribing our shopping list, so I don't have to hunt through the list for the next item every 30 seconds. It's just a lot more efficient to go down a list. Even numbering the list is a big improvement.
I take two lessons from this. First of all, I've already had this idea. I've had the idea to make it a web-based application as well. In fact, I've had ideas built on this basic concept that would be enough to keep me working on it for years. I've elaborated the base concept so much as to make the end product fundamentally different. I've also made it big enough that there's no way I'll even start it, because what I'd be able to do in any short amount of time would fall so far short that I'd consider it useless. The above may be a trivial effort, but I have no doubt it will be useful to a number of people. It forms a nucleus for growing something more.
The other lesson is how obvious this idea would be in a slightly different context. Imagine you were programming a robot to do your grocery shopping for you. Any solution that didn't involve the robot fetching items in a similarly sequential order would announce your incompetence as a programmer. It's so obviously the right thing to do that any halfway decent programmer would do it. That to me suggests a business opportunity, or, more precisely, a whole category of them. Computing has lots of useful algorithms for efficiently sorting, searching, collecting, packing, etc. that can apply well to the real world. Most people, however, are not programmers, and even those that are can be distressingly narrow-minded about applying their knowledge to optimize their daily activities.
Now, you could easily say that I'm a software guy and so I see everything as a software problem. When all you have is a hammer, everything looks like a nail, right
1
? You would be quite right, but you'd also be missing the point. What matters is that computing concepts can effectively solve problems people have in the real world. There's something special about computing in its broad, nearly universal applicability to just about any problem that can be coherently described. Besides, having a perspective that differs is a benefit rather than a handicap when trying to come up with novel businesses.
There's an example of this that you've probably already heard of: Getting Things Done by David Allen. Now, he's not a computing guy. That's not really relevant. A key part of GTD, as I understand it
2
, is keeping track of the status of all of your tasks. Every task is either:
In process right now
Waiting for you to advance it
Waiting for some condition to happen (time, other input, etc.)
Done
Future
Those aren't quite the categories Allen uses, but it's pretty close. Along with this, you need some method for prioritizing what task to work on and how to handle the appearance of a task more important than the one you're currently working on. This sounds a lot like the process model that operating systems use to multi-task. There are a number of programs running at once with a single CPU
3
. Only one can be running at any given time. The rest are either waiting for input from the user, or sleeping in the background pending some event, like 3pm or the completion of a download. The solutions are similar because the problems are similar.
There are a million examples that I could come up with
4
. The principles are universal. That's obviously not the entire formula, else I'd be typing my letter of resignation instead of a weblog post, but it's certainly a start. What remains to be determined is which real world problems map well to which computing concepts; some of them are obvious, but I'm sure there are many surprising and subtle ones. Furthermore, though an algorithm may map well conceptually to a real world problem, turning that into a useful product or service is hardly obvious. There's a lot to mine there, and I'm certainly going to be keeping these ideas in the back of my mind.
1
Or, when all you have is a gun, everyone looks like a criminal, but that's a tangent too far even for me.
2
Caveat: I've only read the first couple of chapters. Jessica, who has read more, assures me that it's not all obvious from there (unlike many "life method" books).
3
In the general, personal computer case, though this is rapidly becoming obsolete. The general principles are the same, however.
The Internet has made cheating on papers far easier than it was in the past. You could read this NY Times article about it, if you wanted. My point is less about that there is cheating and more how one could automatically detect it. Right now, it seems like it's ad hoc, relying on the grader's intuition and checks against sources like Wikipedia.
I think that instructors should use statistical analysis of a student's work to come up with a rough model of their writing patterns, including vocabulary, phrases, sentence structure, and the like. Then they should compare the student's work against that corpus. These profiles should be shared across courses and semesters. It would make it much harder to copy other people's work because it would require copying from the same people each time. That's tough to do with an external source, and also tough to do with a classmate, as their profiles would be available for comparison as well. Or the cheat supplier would have to emulate the cheater's profile, which is not easily done. That's still in the realm of possibility, so schools would have to make sure that at least some of the work in the profile was the student's. That would be accomplished by having regular, in-class writing exercises, under supervision, hand-written, and without the aid of computers or the Internet. Those pieces would be scanned in and added to the corpus. Obviously, none of this would make cheating impossible, but it would hopefully make it harder than just doing the right thing.
It occurred to me that detecting cheating as I just described could make for a fine business. You run it over the web, where professors, instructors, teaching assistants, etc. could upload their students' work
1
. There would be a remote API that allowed custom front ends on the submissions, for example to perform scanning and OCR on hand-written work. The service would be paid for directly by the institutions using it based on the volume of material submitted. They could also get added value from paying for comparisons across institutions and against sources like Wikipedia or published works pertinent to the topic. The general approach could probably be extended to check student work other than written essays, such as computer programs.
I like this because it's a real business solving a real problem that relies on none of the trendy "Web 2.0" features. There are no tags, no social networking, no mashing up, etc. The business model is simple and brings in money from the beginning. There are many ways to extend the core idea as well as integrating with related jobs such as grading. Of course, just because I like it doesn't mean I'm going to do it. It's too much work for too low a chance of success, and I have a little girl to play with.
1
Eventually, the students could just submit their work electronically through the site itself.
Sometimes I type without looking at either the screen or the keyboard. I never internalized orienting my fingers based on the pips often found on the F and J keys, so sometimes my hands end up being off by one. Tjat cam ;ead tp cpmoca; resi;ts sp,eto,es. I should just write a thing that detects when that's happening and automatically corrects it. How many semicolons do I really need in one sentence?
Are confirmation numbers really confirmations, or are they just record keys? That is to say, how does a confirmation prove that I paid my gas bill other than to uniquely identify my payment record? If their database farts and my payment record gets lost, is my confirmation number worthless? My guess is yes. What they really need to do is cryptographically sign a receipt at the time of payment that I can hang onto. It won't prove that I paid, but it can prove that they said I did, which is good enough for me. I suspect most people haven't thought about what a confirmation number really means, otherwise they'd realize that it probably doesn't confirm much of anything.
In software, there are two processes for making software support different languages. One is "internationalization," which is where you make the software look up the various phrases and words from external sources (dictionary files, etc.) rather than having them embedded in the source code. That makes the software merely capable of supporting other languages. The other is "localization," where you actually produce the files with the text for Swiss German or whatever and package a version of the software containing them. Some people decided that the words "internationalization" and "localization" were too long. Rather than abbreviate them the way normal people would, they abbreviated them as "i18n" and "l10n," meaning "'i' then 18 letters then 'n'" and "'l' then 10 letters then 'n'" respectively. I figure I can follow that model and abbreviate my name as "k3n g9r" (and put it on my license plate?), pronounced "Ken Gee-Niner." Kieran would be "k4n," and Uma would of course be "u1a."
This 2 month paternity leave has been a huge help. I think everyone should get a chance at it. The civilized thing to do would be for every father to get 3 months and every mother 6 (even 9) months. However, it's not fair to make employer's foot the bill for something like this; it's good for us all, so we should all pay for it. My half-baked idea is to have the checks come for the government. They'd be on a sliding scale, so that someone who makes $20,000 on an annual basis gets their full pay, while someone making $60,000 gets three-quarters, and all such pay is capped at (an annually-adjusted) $50,000. Those numbers are just for illustrative purposes; actual numbers would depend on a number of factors, not least of which is real data.
Where would the money for this come from? Simple: eliminate the child tax credit. At $1000 per child, the child tax credit is worth approximately $18,000 until the child reaches majority. Average annual household income in the United States is about $45,000. In the simplest case, where the mother and father earn equal amounts, 6 months pay for the former and 3 months pay for the latter works out to about $16,000, which is roughly the same. Obviously, the present value of the child tax credit isn't $18,000, but the numbers are in the same ballpark.
One oversight in this scheme is what to do when only one parent works outside the home, especially in the case where one parent decides to stay home. There are other details I haven't worked out, but I'm confident such problems are relatively minor and could be worked out easily.
Seems like there's a market opportunity for a pickup truck taxi service. My options for moving heavy things myself are:
Have my own truck - not bloody likely
Mooch off a friend - Hi, John!
Rent a truck - rental period is longer than I want, plus I have to deal with the hassle of picking it up and dropping it off
On the other hand, I could just call someone, have them meet me at my home, drive over to the place I'm getting stuff, load it up, and drive it home. As a bonus, the taxi driver could double as a helper. Home Depot rents trucks at pretty inexpensive rates, but it's not so handy if you want something from Lowe's or some guy named Craig who has this list. I wonder what it would cost to provide such a service. I figure the driver needs to get $12/hour, and it would cost something like $0.60 per mile. Then you've got to amortize a $25,000 crew cab pickup truck over, say, 4 years. So you'd probably charge something along the lines of $40 for a one-hour job. That's basically a guess; does it at least sound right? I think there are probably enough people who'd be willing to pay that. Maybe not in Austin, though; everyone knows somebody who has a truck.