Slashdot Mirror
Young Irish Scientists Win Award for Linux Project
Trevor Johnston writes "Last week, three Irish students used a Linux box together with an old Basic programming language running on an Amiga emulator to display graphically the output of their own computer learning program. For their troubles they were awarded 'Best Use of Information Technology' at Ireland's Young Scientist Exhibition."
The award they were given has almost nothing to do with the tech. they presented. From their description, the whole contest sounds somewhat bland and menial.
The only interesting thing is the fact that they wrote it for Amiga, though they really didnt have to use it. They said that the only reason Amiga was used was because the programmer had experience in that area, not because Amiga was much better for what they were doing.
Also, the Linux program was written using C++.
"A great deal of intelligence can be invested in ignorance when the need for illusion is deep." --Saul Belloe
To save a lot of /. readers the trouble, I'll now post a summary of the the next 42 messages:
I was doing that with linux before I was 4 years old! (See related story)
As a teen myself, I have been to these kinds of Science Compititions (Timmins, my hometown, hosted the Canadian wide science compitition for high-school students last year). I find it disgusting when some kids actually fool their way up to the top by actually using other people's innovations. I once seen this kid that used a 3d modeller in windows, and recreated his hometown, claiming it to be some sort of "Virtual Map", that's about it. Nothing new there, sure, it looks cool, and the judges will gawk and of course they win the prize, ech, disgraceful to the informed community at large.
Anyhow, I really like the fact that these kids actually took their time and interest to develop something that although may seem a little redundant, they actually learn something useful and bring some perticular incite to others.
Keep it Up! :)
So these students have developed a rudimentary adaptive processing algorithm ... great! As a participant in science fairs all through middle and high school, I can say two things about them right away -- on the whole, little original actually gets done, but the mere experience of performing experimental research in what passes for a peer-review environment has produced quite a few outstanding student (and later, professional) scientists. Moreover, at each science fair, I invariably made an assortment of contacts that later proved important to my future career as a student and researcher.
So, anyway, don't knock them for reproducing ancient and well-studied algorithms. Instead, encourage them for trying to develop something at all, and hope they continue to do so.
Quantum mechanics: the dreams that stuff is made of.
It has two methods with which to complete the task and by running a certain amount of games, the computer can work out which is the better method to use.
This is actually the 2 armed bandit problem: you have a 2 armed jack pot machine, and everytime you pull an arm, it gives a certain payoff, but the payoff is probalistic. So do you keep pulling the arm that has the higher average payoff? Or do you try the other arm? It's a tradeoff between exploitation and exploration, and the solution is surprisingly mathematically involved (see Gasoml by Goldberg).
John Holland showed in the 70s that the genetic algorithm is a near optimal solution to this problem. So in a simple way, these students have rediscovered the essential issues of genetic algorithms.
I write code piss drunk all the time. (read: I've finished the sixer of stout and I've moved on to rum.) The problem seems to be the next morning, when I have to look at the uncommented lump of spaghetti who's purpose I only half-remember.
As a whole, 1/8 of the stuff I produce greatly exceeds my sober level; it is perfectly brilliant. 1/4 of it is marginally competent, my norm. The rest of it is not fit for the light of the CRT, and may be only slightly more intellegible than Vogon poetry fed through Babelfish.
Brilliant or not, I usually just 'rm -rf *' it; I never have the same grasp of it again.
.sig: Now legally binding!
Ten years ago, you'd have needed to be in a university to get the resources to tinker with that kind of project. Now it's in high school science projects. That's got to be progress...
I think this story is a fine example of what a benefit Open Source can be for Computer Science education.
Lemme guess, you did that to increase your hacker purity test score?
I did, as well. Actually, you get another point for attempting to increase the score.
Ah, those high school days...
Of course, The Diamond Age also featured evolutionary design - the nanites, remember?
I'm surprised that no one else has commented on this yet. Here we go.
;)
Trying to use Prolog for this would have been a bad, bad idea. Prolog is for logic programming, which is great for theorem provers and any other such problem which depends on logical relations between a bunch of data, but entirely unappropriate for a graphical real-time game. Especially considering that the amount of AI involved in this problem was very small. Especially considering that graphical games require languages appropriate for graphical programming, and Prolog was definitely not designed with graphics in mind.
Now Lisp, on the other hand, would have been appropriate. There is a good reason why it is the language of choice for AI work, after all: it's easy to model most problems using Lisp. In fact, I'm developing a General Game Data Model (GGDM) for Scheme (on RiceU's DrScheme environment, which provides just about everything you can ask for - easy graphical programming, a simple network model, good multithreading support and a lot more), a simple extensible class library indented to turn all of these kind of problems into a simple matter of defining a few objects with behaviour defined on the fly. Anyone interested can email me.
As for C++... well, I can only guess what a mess the code wound up looking like. Maybe the boys care to open the source?
To the editors: your English is as bad as your Perl. Please go back to grade school.
And now for the brickbats to some of the posters I've seen here:
Again, my hat's off to these people, and I hope to see more from them in the future!
Strike while the irony is hot! -- The Freethinker