Log in

Bad Ideas
[Most Recent Entries] [Calendar View] [Friends]

Below are the 10 most recent journal entries recorded in Bad Ideas' LiveJournal:

Tuesday, May 24th, 2005
12:26 am
The Farm: Part Three
So far, we have discussed our motivation for a new agricultural system, introduced the idea of an underground farm, and discussed some potential problems and their solutions. This time, we will describe some unique advantages of the underground system. We will also go over our basic plan for funding.

As stated last time, we want to use an automated system for planting, maintenance and harvest. An additional advantage of using this system is that a computer could easily keep track of each plant. If we were to place an RFID tag on each plant when it is young, the computer could then identify which plants are growing better than others and alert a human if there is a problem in a particular area. Recent advances in computer vision techniques would also allow the robot to identify any plants which were significantly different from the norm, such as those which have a disease or have a parasite living in them. The plant could then be killed and removed automatically. The RFID tag and associated history of the affected plant would also allow the computer to identify any factors which were different in that plant for further analysis. Additionally, the robot could identify when each individual plant is at its best for harvest and collect its fruit or grain at the optimal time, rather than taking the expected time that the field as a whole is ready for harvest and collecting it all at once.

The underground environment provides other advantages in addition to additional surface area on which to grow crops. For one thing, it is semi-closed. We've already stated one advantage of this in that planting, maintenance and harvesting can all be automated, since unwanted objects from the outside are much less likely to come in. One other advantage of this in combination with the RFID tagging and individual monitoring of each plant is that we can ensure that no "volunteer" plants grow in our field. If you look at the side of a box of Corn Flakes, it has a warning stating that "the corn in this product may contain traces of soybeans". This is an issue because of people with a soy allergy. Because we can ensure that our fields are wholly one plant, we can eliminate this concern.

The opposite is true as well: things from inside are less likely to exit unexpectedly. This has huge advantages for research purposes. An agriculture researcher studying a new breed of plant, a new fertilizer or pesticide, or insect-plant interactions can reliably assure the public that none of the experimental subjects will leave the testing area. And because all operations inside the cave are done remotely, we may use techniques that would otherwise cause concern for the safety of the workers involved, such as the deployment of radio isotopes to encourage even more rapid growth. And in the unlikely event of an infestation of pests, we would also be free to use more powerful pesticides like DDT, since it is unlikely that they would get out into the environment at large.

No doubt this is an expensive undertaking. Fortunately, it falls under many categories, so funding shouldn't be an issue. We plan to apply for grants in agriculture, electrical engineering, materials science, robotics and computer science. We also hope to get funding from groups promoting the advancement of agricultural techniques in third world countries or sensible agribuisiness and from groups which are interested in applications of RFID technology.

Sidenote: We came up with this idea in the fall of 2003, so it's actually been brewing for a while. It took us a long time to flesh some of these ideas out. Last fall, I was told that this is actually a topic of current interest among some reputable people. I wasn't able to find any information online about it, but if anyone comes across anything, I'd love to know about it.
Monday, May 23rd, 2005
12:04 pm
The Farm Part 2
The Farm: Part Two

Last time, we discussed the motivation for a change in current agriculture practices, introduced the idea of an undeground farm and brought up some of the potential benefits of the concept. Today, we will talk about some problems with the idea and put forward our solutions.

The most obvious problem with the idea of an undergound farm is the lack of sunlight. But there is already a tried and tested solution to this problem. For several decades, people have been growing some plants in artificial lighting for various reasons. In fact, this is preferable in some ways to natural lighting, since you can control the length of the "day" that the plants experience. By adjusting the amount of day and night and the frequency at which they change, you can optimize the growth of the plants. This is common practice for some seedlings as well as certain expensive plants such as orchids. The current practice is to use ultraviolet fluorescent lighting, but we believe that there is a better solution. Recent advances in semiconductor technology have allowed much higher frequency LED's for much lower cost than were possible as recently as five years ago. LED's provide much more light with much less heat waste than incandescent bulbs or fluorescents. Another important factor giving LED's the edge in this regime is that they have much longer lifetimes than other light sources. A fluorescent bulb running for long periods (10+ hours at a time) is expected to last on the order 14000 hours. This is less than three years, assuming a 16 hour day. By comparison, LED's are generally expected to provide on the order 100000 hours of operation, regardless of how long they are on at a stretch. This is over 17 years if we assume the same 16 hours per day. Thus, even if we only consider the cost of the unit itself as a factor in which is better, the LED's may be allowed to cost seven times as much as the fluorescents for the same amount of light. If we were to take into account the cost of installation, the extra energy required because of lower efficiency and the cost of the downtime caused by burned out lights, the LED's are even more certainly the better option.

A second problem is planting, care and harvesting of the crops. In current surface farms, planting is accomplished with tractors pulling trailers, crop maintenance is done either using the same tractor and trailer method or with fly-over dusting of crops with fertilizer and pesticides, and harvesting is done with a combine. But we can't use the same methods for an underground farm. In order to make the system as efficient as possible, minimize the cost of removing the soil to create the cave and make sure that the structure holding up the cave ceiling is under as reliable as possible, we want to make the cavern as small as possible while still leaving room for the crops to grow to their maximum height. This completely eliminates any possibility of fly-over dusting, and makes tractors and combines impractical. A person might fit while the crops are young, but as the plants grew to full size, it would become increasingly difficult for a human to enter the field without damaging the plants. Furthermore, we would like to avoid harvesting crops by hand, since it is extremely inefficient when compared with mechanized harvest. But again, we have a 21st century technological solution: use a robot. Using a track grid running between the upper and lower crops, a remotely controlled robot could easily go anywhere in the cave. Using different attachments, the same robot could be used for both planting and harvesting, as well as for deploying fertilizer and pesticides. Because there is no danger of stray people, animals or other objects inside the cave, and since it would be running on a track instead of with wheels on the ground, this process could be done entirely automatically, needing human intervention only when something unexpected happens.

Another problem we face is water management. Since they are underground, they will not recieve enough water on a normal basis from rain. At the same time, flooding could be a problem, since if a large amount of water were to get in, it would not evaporate as readily in the more closed underground environment. Providing water actually turns out not to be as much of a problem as it might appear. If we collect water on the surface and pipe it down to the underground, it should provide sufficient amounts of water when not in times of drought. The source of this water could be any of the traditional sources: rain, river, lake/reservoir. It doesn't matter. Once it's in the cave, we can use pipes with spray nozzles running throughout the cave to distribute. This is not a new concept (it's already used in modern farms to make sure water is distributed evenly), however because the farm is underground, there is an additional gravitational potential on the water. This allows us to save power by not having to pump water down the pipes. We let gravity do most of the work for us, only supplementing energy with a pump if necessary.

Drainage turns out to actually be helped by being underground. A major concern when designing a drainage system for a farm is that it needs to have an outlet deep enough that it is below the normal water level in the area, and so won't just come back up and flood again. Normally, this requires digging down to a depth of at least a few meters. But since our farm is below this depth already, we hardly need to dig at all to get to a suitable depth. Thus, the drainge system can just be the usual method of collection bins covered with a few inches of soil, but with the drain at the same level, rather than digging down further.

These systems require power. We would like to make the entire setup as efficient as possible, so rather than simply buying electricity from a power company, we would like to produce our own electricity as much as possible. Fortunately, modern technology once again provides a solution: methanol fuel cells. Normally, the stems of the plants are unused and left to die and rot after harvest. In addition to being wasteful, this has the disadvantage that it creates a layer of groundcover which must be overcome. Traditionally, this is just cleared away, but more recently, it has been plowed under before seeding time. Rather than do either of these things, we would like to collect this waste plant matter and ferment it to produce methanol. The methanol can be used in a fuel cell to provide electricity, which would be used to power the lights, robot and water system. While thermodynamics tells us that this cannot provide all the energy required, we believe that it would provide a significant fraction of what we need, further reducing our cost of operation.

Next time: some unique advantages.
Sunday, October 17th, 2004
3:33 pm
The Farm: Part One
Note: this idea is really long, so we're splitting it up over multiple days.

A major problem in the world today is lack of arable land for farming. We don't see the effects of this problem very much in the US because we have large areas of the country which have low population densities and good conditions for crop growing and because we have modern agricultural techniques which allow us to get much more out of the land that we have than any premodern society ever could have dreamed of. But in third world countries where much of the population must make food production its primary concern, the issue of creating enough food to feed the people is a major issue. But this isn't a geopolitical forum, so I'll leave the motivations for increased production behind and move on.

Whatever the motivation, any increase in the amount of usable land for crop growth is generally seen as a good thing. Traditionally, the main way of achieving this is to irrigate or otherwise modify nonarable land to make it more suitable. But this has problems of its own. For one thing, any spread of farmland displaces whatever was in that area before. Aside from displacing local flora and fauna, staple crops (such as grains) tend to be poor choices for preventing erosion, which may eventually cause the land to become unarable again.

So instead of trying to increase the amount of usable land surface area for farming, we propose a solution which allows you to get more out of the area you have. This is not an original concept in itself. Modern agriculture has various techniques for achieving this, such as sophisticated crop rotations, multiple crops per year and fertilizers and pesticides to encourage plant growth. But there is a limit to how much these techniques can do, and so more must be done.

So, how to get even more out of a given amount of land? The solution is simple: stacking. But trying to grow crops above ground is absurd. You would need to have some way to keep enough soil to grow your crops kept far enough off the ground that you could grow your original plants below. This would involve the raising of thousands of tons of soil to a height of at least 2-3 meters per field. That's a lot of energy usage. Furthermore, you would need to have a structure capable of holding up that much soil, and you would need to somehow prevent the soil from blowing away. This problem is worsened because the structure and soil being raised up would be more susceptible to storm damage. We believe our solution addresses all of these concerns.

What is our solution? Build underground, of course! An artificial cavern several meters below the surface and extending for a height of a further 2-3 meters would triple the surface area usable for growing crops! A farmer would be able to grow a given crop on the surface, the same crop (or possibly a rotation of that crop) on the floor of the cave, and then a vine or other plant which is willing to grow downwards on the ceiling of the cave. In addition to having all of the advantages of any stacking solution, a cave at that depth would be somewhat immune to seasonal temperature fluctuations, allowing a longer growing season and allowing a wider band of regions to grow tropical and subtropical plants. And creating stable underground caverns is not as difficult as it might seem. People have been doing it for thousands of years in mines.

Next: specifics!
Thursday, October 14th, 2004
1:26 pm
For those who have been waiting patiently for an update from us (bless you all!), have no fear! We have a draft of an idea in the pipeline. Thanks for having some patience with us, and you can expect a real update in the next day or so.
Tuesday, June 29th, 2004
4:34 pm
The assumption that Yiddish is derived from German is as inaccurate as the frequently heard statement that man is descended from monkeys. Actually, modern Yiddish and modern German have a common ancestor in the dialects of medieval Germany just as present-day man and ape may be said to be descended from a common pre-human, pre-simian ancestor.

When Jews settled in Germany in the 9th and subsequent centuries, they introduced into their newly acquired speech, from the very outset, numerous words of Hebrew and Romance derivation, connected with the specific Jewish way of life. In the German element of the Jewish speech, peculiarities began to crop up, too, since the Jews formed a cultural group of their own.

As time went on, the differences between Jewish and non-Jewish speech in Germany became increasingly marked. When many Jews migrated eastward, first within German lands and then into Slavic-speaking countries, Yiddish developed along independent lines into a separate language. Meanwhile, the German of the Middle Ages was changing into modern German. (source: Weinreich, College Yiddish)

Yiddish, although it follows many Germanic grammatical conventions, and most of its vocabulary is descended from the same linguistic ancestor as modern German, is written in phonetic Hebrew, the ancient language of the Jewish people, and not in the Roman alphabet that the Germanic people adopted from the Roman Empire. The slow integration over hundreds of years of Hebrew, Aramaic, Romance, and later Slavic languages that led to Yiddish is of great historical interest, though our understanding of exactly how this happened is very much incomplete. To gain better insight into how such languages have evolved, we propose an experiment. We immerse native Japanese speakers in rural Germany, and create a new language composed of mostly German words and governed by German grammar while still retaining the original Japenese in about 5% of the new vocabulary. The new language, written entirely in phonetic Japanese, shall be called Gerpanese. If the experiment is successful, about 1000 years later, German and Gerpanese will have taken different evolutionary paths. Gerpanese will reflect the culture of the Japanese, and the new Modern German will have diverged from the German content present in the Modern Gerpanese. Also, Columbia will have expanded its offerings to include a doctoral program in Gerpanese..
Sunday, June 20th, 2004
10:51 pm
The Worst Sorting Algorithm
I've been busy this week and skarfin is out of town, so you get a short one this week. On the plus side though, it's the idea that started us going on this crazy thing: the worst sorting algorithm imaginable (without getting silly like unsorting things on purpose).

badsort( array A )
for i=0 to sizeof( A )
do if A[i] > A[i+1]
then randomize( A )
0 -> i
return A

If someone cares to figure out the expected running time of this algorithm, I'd love to hear it.

Current Mood: exhausted
Sunday, June 13th, 2004
11:28 pm
The Traffic Light Challenge
Okay folks, here's a fun game to play when driving in the city on a one-way up/downtown street. First a little background. I'll use Amsterdam Avenue as an example. As many of you probably know, Amsterdam Avenue (and 10th as it's known in Midtown) is a one-way street traveling uptown until about 110th where it turns into the two-way street which we have around Columbia. The NYC traffic light folks have figured out that the traffic moves nicely if they delay the change from red to green of each traffic light by about the time it takes for a car to travel one short city block at 25-30 mph. So as you drive uptown on Amsterdam, each light will change from red to green as you approach it, or if all the lights in sight are green, you can catch up to a red by traveling faster than 30 mph. The beauty of this system is that in even moderate traffic, you often never have to stop for a red light. The downside is that it limits your speed, especially at night when the traffic is lighter.

So here's the driving challenge: When traffic is light, go to a one-way avenue such as Amesterdam, and catch up to a red. Then, aim to arrive at the edge of the intersection the moment the light turns from red to green. Proceed through the intersection at constant speed and maintain it until the next traffic light, which should be red up until the very moment you arrive at the intersection. Continue until the road ends or the light pattern changes. The object is to maintain a constant speed while never slowing down and entering the intersection as close as possible to the instant the light turns green, but never while it's still red. In fact, if you're good, you should be able to figure out exactly what speed the lights are timed for. Driving in heavier traffic may require adjusting your speed to avoid accidents with other vehicles and road obstacles.

Once you've mastered this skill, your next challenge is to do it during the daytime in moderate traffic. You should be sure to weave through the traffic and still arrive at the next intersection at precisely the moment the light changes for red to green. As you will see from the scoring system outline below, driving in traffic increases the number of points you can earn.

Entering the intersection when the light is red: 0 points
Entering the intersection when the light is green: The number of milliseconds that the light has been green subtracted from 1000. Minimum score for crossing an interscetion when the light is green: 1. 5 points for each mile per hour of speed.
Additional Points: 50 points for each double parked vehicle on the current block. 50 additional points for each axle above 2 on a given double parked vehicle. 50 points if a taxi cab is stopped to load or unload a passenger on the current block. 50 points per bus on the current block. 10 points for each other vehicle on the current block.

Note: The author would like to inform the user that this really is a bad idea, and is extremely dangerous to actually try, so don't do it. The author is not responsibile for consequences of playing this game, and highly discourages it, unless it's in GTA.
Monday, June 7th, 2004
5:03 pm
The hobbies of cars and computers are often compared as similar. There are body mods for cars, and case mods for computers. One can overclock a computer to improve its performance beyond spec, and one can bolt a turbocharger onto a car to make it accelerate faster. You can add a hard drive to a computer to allow it to store more data, and a trailer to a car to allow it to carry more stuff. More ram is similar to a larger gas tank. Etc.

Along the same lines, price comparisons are similar. The lowest end hobbyist computer might cost on the order $500. The highest end might go as high as $5000. For that extra factor of ten in cost, you get more processing power, much more disk space, etc. The lowest end hobbyist car might cost around $10000. The highest end around $100000. Again, a factor of ten in price, which gets you a faster car, or a car that can store much more stuff, or whatever.

However, there is one realm for which this comparison has not yet been extended by hobbyists: a supercomputer. The third most powerful computer in the world (the "BigMac") is built from 1100 dual CPU G5 nodes. This costs on the order $3 million. Each node has 4gb of ram, and 160gb of disk, for a total of 4400gb of ram and 176tb of disk over the entire system.

While no equivalent of this exists in the automobile world, we believe one should, so let's consider what this requires. A computer that cost roughly $5000 (the "high end hobbyist" price I listed above) might have 2gb of ram, 1tb of disk and be about as powerful as one of the nodes in the BigMac. Now, consider trying to make a car that is as to the "high end hobbyist" car the same as the BigMac is to a computer that a hardcore hobbyist computer afficionado might have.

A good example of a $100k hobbyist car would be the Acura/Honda NSX, which costs very nearly $100k new (more or less, depending on the options chosen). The NSX has a 3.2 liter V-6 engine which generates nearly 300 hp and about 225 lb-ft of torque. Its fuel tank is 18.5 US gallons. It can carry a total of 53.9 cubic feet of cargo (including driver and passengers). Working from the redline and gearing ratios given by Acura, this would give a top speed of about 140mph (stock).

Now, let's consider our hypothetical "supercar" equivalent. It needs to have about 2000 times as large a fuel capacity (equivalent to ram), carry 176 times as much cargo (equivalent to disk), and be about 1100 times faster. So our supercar needs to be able to carry nearly 9500 cubic feet of cargo (this is about twice what a semi can carry), carry 37000 gallons of fuel (about 300 times what a semi carries), and be able to go 154000 miles per hour (about 82% the speed of light). It should cost a mere $60 million (a bargain for that speed) and it should weigh about 1734 tons.
Sunday, May 30th, 2004
3:37 pm
The Proper Method of Heating Sake
On your way back from the drivethrough liquor store after purchasing a bottle of fine sake, a common problem is heating the liquid to an acceptable temperature for consumption. Fortunately, there is an adequate heat source close at hand: your engine. So rather than wait to get home to drink your sake, why not drink it on the way?

This requires some preparation. First, stop at your local spyware store (or hardware store, if you live in an area where the secret agent population is low). You will require two large mirrors, metal tubing, bolts to hold everything together, two long pieces of rubber tubing, a funnel that will fit inside the rubber tubing and duct tape. You will also need a ceramic vessel (to hold the sake). In addition, you will need a tool to make holes in the sheet metal of your car.

You will be attaching the mirrors so that you can see the road in front of the car in the mirrors. Attach the mirrors at 45° angles to the tubing (like a periscope), so that when you look into one mirror, you see out the other one in the same direction. Punch holes in appropriate locations on the side and inside your car so that you can attach your tubing/mirrors to the car and see forward out the car using the mirrors.

Now, pop the hood of the car, and attach the ceramic container to the engine block using duct tape. Put the rubber tubing down to the bottom of the container and tape it in place. Run the tubing to the drivers side window, taping as appropriate to keep it from moving around. You need to keep the tubing outside the car to avoid breaking laws regarding open alcohol containers inside the passenger compartment of a motor vehicle. Now run the other piece of tubing back to the ceramic container, taping as appropriate. This piece of tubing should have its top as high as possible while still allowing the driver access to the end. Attach the funnel to the end of this piece near the window.

Now get your sake at the drivethrough window. Do not shut your engine off, as this will inhibit the heating process. Fill the ceramic container using the tubing with the funnel. Allow a few minutes for the sake to heat (this will take more or less time depending on how much sake you use). Once the sake has reached a sufficient temperature, suck it out of the ceramic container using the other tube.
Sunday, May 23rd, 2004
9:23 pm
Hello, and welcome to the Bad Ideas Community on Live Journal. As some of you may already know, Alex and I have spent the past year thinking up many bad, some even truly terrible, ideas. We thought that now is as good a time as ever to bring to you a forum for sharing and exchanging bad ideas. We hope that you will enjoy this community, and thank you for joining.

About LiveJournal.com