Science


How to make hydrogen airships safe.

1> Suspend the gondola / cargo etc. 200 m below the lift balloons to keep them away from explosions.

2> Carry a cargo-sized parachute, perhaps with powered deployment assistance, on the gondola.

That way, if it goes bang, you don’t get fried, and you get to coast to the ground. Helps if you fly high.

May 25 2008 08:46 pm | Science | No Comments »

Genetic algorithms for making hardware do things (on FPGAs)

Dr. Thompson peered inside his perfect offspring to gain insight into its methods, but what he found inside was baffling. The plucky chip was utilizing only thirty-seven of its one hundred logic gates, and most of them were arranged in a curious collection of feedback loops. Five individual logic cells were functionally disconnected from the rest– with no pathways that would allow them to influence the output– yet when the researcher disabled any one of them the chip lost its ability to discriminate the tones. Furthermore, the final program did not work reliably when it was loaded onto other FPGAs of the same type.

It seems that evolution had not merely selected the best code for the task, it had also advocated those programs which took advantage of the electromagnetic quirks of that specific microchip environment. The five separate logic cells were clearly crucial to the chip’s operation, but they were interacting with the main circuitry through some unorthodox method– most likely via the subtle magnetic fields that are created when electrons flow through circuitry, an effect known as magnetic flux. There was also evidence that the circuit was not relying solely on the transistors’ absolute ON and OFF positions like a typical chip; it was capitalizing upon analogue shades of gray along with the digital black and white.

http://www.damninteresting.com/?p=870

May 24 2008 10:12 pm | Science | No Comments »

well that’s damn clever

http://en.wikipedia.org/wiki/Geohash

what’s nice is that you can trim off the end of the hash, to get a less and less precise hash, because it basically represents a quadtree in hash from.

Brilliantly simple, and well executed.

May 23 2008 02:37 pm | Science and The Global Picture | 1 Comment »

the trouble with ephemeralization

http://www.alternet.org/story/84190/ - interesting piece about somebody’s observations on Ikea and how it affected her family business, although she still buys it.

Bucky called this ephemeralization - stuff gets cheaper and lighter and less complex. Really surprising to see how it is playing out.

May 04 2008 03:02 pm | Science | 2 Comments »

Pigs as insecticide

http://engineering.curiouscatblog.net/2008/04/20/pigs-instead-of-pesticides/

Apples drop from trees due to infestation.

Pigs eat apples, preventing reproduction.

You think monoculture farming might be really, really stupid? Like, clearly this is pretty much how the apple trees evolved to do this - drop the fruit when there’s a problem, and something will eat it. Fence off the apples, keep out the mammals, and the natural defense mechanisms are broken.

Monoculture is stupid. Throw rocks at it.

Apr 25 2008 08:42 am | Science | No Comments »

The Shrimpy Goby

http://engineering.curiouscatblog.net/2008/04/10/mutualism-inter-species-cooperation/

Very, very cute inter-species cooperation story.

Apr 25 2008 08:36 am | Science | No Comments »

very interesting wired piece on sort-of transhumanist software

Twenty years ago, Wozniak realized that computers could easily calculate the moment of forgetting if he could discover the right algorithm. SuperMemo is the result of his research. It predicts the future state of a person’s memory and schedules information reviews at the optimal time. The effect is striking. Users can seal huge quantities of vocabulary into their brains. But for Wozniak, 46, helping people learn a foreign language fast is just the tiniest part of his goal. As we plan the days, weeks, even years of our lives, he would have us rely not merely on our traditional sources of self-knowledge — introspection, intuition, and conscious thought — but also on something new: predictions about ourselves encoded in machines.

http://www.wired.com/medtech/health/magazine/16-05/ff_wozniak

Apr 22 2008 03:33 pm | Science | 1 Comment »

“compare by hash considered harmful” considered harmful

Once a hash collision has been found and a demonstrably buggy test program created using the colliding inputs, how will you fix the bug? Usually, the response to a test program that demonstrates a bug in the system is to fix the bug. In this case, the underlying algorithm is the bug.

What’s the error?

SHA1(X) != SHA1(Y) proves X and Y are different

SHA1(X) == SHA1(Y) does not prove that X and Y are the same, although it very strongly suggests it

So, yeah, the piece is talking about hash abuse by programmers making a simple logical error in how they use hash functions. Options?

1> Compare everything that hashes, bit by bit, and in the event of a collision
A> Call Guinness
B> Store both copies with a version tag

2> Accept the 1 in 2^128 error risks which… well… etc. etc. you know the story.

http://suse.groenbaek.net/openlife/2008/04/13/non-commercial-use-of-patents-is-not-illegal/http://www.usenix.org/events/hotos03/tech/full_papers/henson/henson_html/node8.html

Apr 14 2008 07:08 pm | Science | 1 Comment »

The Perfect Game.

So I think I’ve found a perfect game. I haven’t played it yet, and there are some details to be figured out, but I think it’s pretty clearly potentially a good abstraction for a certain class of mathematical and political issues, and also looks like it could be quite entertaining if you’ve got a taste for Poker or Diplomacy.

I’ll start with the simplest form I can think of.

1> There are an odd number of players

2> Each player has a vote, which can be YES or NO

3> On each round, a vote is taken, and the majority wins

4> If there are any players on the minority side, they lose

So that’s the essence. Now add some scoring.

5> At the start of each round, players put some tokens into a pot

6> The winners split the pot between them

This creates an incentive for narrow wins - in a nine player game, a 5-4 win pays out 7.2 times what an 8-1 win pays out.

One annoying problem is that this method of keeping score is impossible with beans or tokens or whatever. One simpler approach is that each loser must give one token to each winner, which also has the right properties, although I think the different risk/reward landscape would reward a different style of play to the version in which the cost of a round is the same whether you lose narrowly or by a broad margin. Somewhere in the middle is a version where on each round, the winners gain a point and the losers lose one, but then the number of points in the game is not conserved - also a problem.

I look forward to suggestions on how to resolve this scoring problem. Also note that bribery (vote with us and we’ll give you two tokens) should probably be permitted, although on an informal basis, with such promises having no in-game mechanic.

Then there are two modes of play.

7> Players make their vote by some secret mechanism, such as placing a tile face down in front of them, and all communication is face to face at the table

or

8> Players mark their votes electronically, allowing for dialog

9> There may also be issues regarding timing, such as when players decide to wait for one person to vote, and then all change their votes. These problems can be resolved at the tabletop quite easily, but may be a little less tractable in the electronic form

This game is inspired by the Byzantine Generals Problem.

I’m open to suggestions on names - I was initially thinking Byzantium, but that’s a little boring.

Oh, and for play over (for example) email, I suggest the following scheme.

1> players post a hash to a mailing list, like so

dbef29982b1ce8213ed74f626d3bb9e1

Once all hashes have been posted (possibly including multiple votes from a single player as they change their minds or bluff)

2> Players then post the string which evaluates to that hash “YES: this is randomness 2u23i4ou23p4″

3> Strings can be checked easily

MD5 (”YES: this is randomness 2u23i4ou23p4″) = dbef29982b1ce8213ed74f626d3bb9e1

It’s interesting how much of the play in the game is protocol dependent - can players revoke votes? can they see when other players have voted? how are points allocated for wins and losses? what are the victory conditions?

And, finally, what about the pathological case, where all players are on one side?

Apr 13 2008 03:44 pm | Science | No Comments »

Male contraceptive injection - apparently works pretty damn well

It’s an injection that has a similar effect to a vasectomy, it lasts three months with limited study on 1756 men showing it can perform as an effective contraception in 95% of males. The ANZAC research institute may be on to something big here though it looks like it might not be going anywhere with some sponsors backing out and the government not showing much of an interest in the project.

One of the projects researchers, Professor Liu, says major pharmaceutical companies have dropped out and that the government itself may not be fully aware of the demand for a male contraceptive such as this one. “I think there’s a real need for it because I see a lot of couples where a female doesn’t want to use or is intolerant to a female contraceptive pill or males who want to delay vasectomy” he said adding that “My perception is that people want it but it seems to be that somehow that isn’t being heard at a government level”.

The contraceptive works by suppressing sperm output, it is largely composed of androgen and progestin and lasts for up to three months. Theoretically and with limited research there doesn’t seem to be any adverse side effects, it works on 95% of males and will really help improve the lives of many people.

Contraceptive for men a shot in the arm

Whether or not you’re an Australian citizen I highly recommend you contact the Queensland government on this issue. While the contraceptive is being developed in Australia its results could really help the entire planet.

http://botlbrush.com/blog/?p=133

Obviously 95% isn’t enough, but why isn’t this front page news?

Apr 13 2008 11:48 am | Science | 1 Comment »

Next Page »