What projects would you like to see?

[Nexus]

If someone started a new DC project tomorrow, what would you like it to be?

[Welnic]

df, phase III

[PY 222]

A project that have a wonderful staff, great stats and something worth while that we can really be proud of at the end of the day.

[FoBoT]

i have one idea, but i don't think the # of cylces available is adequate



calculate the size of my butt

[CaptainMooseInc]

I'd like to see a DC project that tackles one of the 6 mathmatical problems that are still theories and never have been solved. I say 6 because 1 of the original 7 had a potential solution found for it by some russian mathematician. It still isn't confirmed yet. I'd just like to see all 6 figured out before the end of the first half of this century. It took a few years to crack that one encryption thing, it might take longer to figure out the math problems. With the computing power we have now though, I think it could be done.

[Nexus]

I'm not so sure those problems are the kind of things that can be tackled by brute force. If any of them can be reduced to checking a finite number of cases then yes, go for it.

What I would like to see is a DC project that designs a fusion reactor. For those who don't know, fusion is the power source that the sun uses. It involves hot hydrogen atoms colliding and forming heavier elements, during which energy is released. These reactions can be got going for a little while in the lab (~30 seconds from memory) but the trouble is getting the hot material to stay away from the sides of the container. It's that bloody hot that if it comes into contact with the walls your container will be damaged. So you've got to come up with design for a magnetic containment field that keeps the plasma in the middle. That's exactly the sort of thing DC can do well as it's basically checking and optimizing various designs.

[Death]

http://wiki.seventeenorbust.com

17oB! v.3

[Thor]

@Nexus

That would definitly be a project worth crunching for!

maybe it starts one day...


Thor

[Fozzie]

I don't think the DC community is ready for a brute force attack on that subject.

Fusion sounds cool though.

[jonnyw]

Originally posted by CaptainMooseInc
I'd like to see a DC project that tackles one of the 6 mathmatical problems that are still theories and never have been solved. I say 6 because 1 of the original 7 had a potential solution found for it by some russian mathematician. It still isn't confirmed yet. I'd just like to see all 6 figured out before the end of the first half of this century. It took a few years to crack that one encryption thing, it might take longer to figure out the math problems. With the computing power we have now though, I think it could be done.

fancy giving us a description of one or more of these problems

[Helix_Von_Smelix]

Originally posted by Fozzie
I don't think the DC community is ready for a brute force attack on that subject.

Fusion sounds cool though.

Fusion...........cool. Now that is funny. Nice one Fozzie!!!!

[Nexus]

Originally posted by jonnyw
fancy giving us a description of one or more of these problems

Millenium Problems

1: Birch and Swinnerton-Dyer Conjecture

This conjecture asserts that the existence and number of integer solutions to algebraic equations like x^2+y^2=z^2 is related to the behaviour of the Riemann Zeta function near 1.

I don't think Distributed Computing can prove this one, just show it is true for an increasing number of cases.

2: Hodge Conjecture

One way of studying complicated shapes in mathematics is to treat them as numerous simpler shapes glued together. Unfortunately, some of the simple shapes involved don't really have a sensible interpretation (that is, you won't find anything like them in our universe). Hodge's Conjecture states that a useful class of these shapes are actually examples of something even simpler.

Well, as I don't really understand the problem too well I can't tell whether it's something that DC can tackle. I would guess though that it's not applicable.

3: Navier-Stokes Equations

Physicists think that systems of turbulence like water currents and turbulent air obey equations called the Navier-Stokes equations. If solutions can be found for these monstrous equations our understanding of turbulence would be increased dramatically.

This is one I think DC can deal with. At least, it should be possible to find approximate numerical solutions, to any desired degree of accuracy. It would take a lot of crunching power though.

4. P=?NP

This is a great unsolved problem in computer science.
P is the set of all problems that can be solved in polynomial time. NP is the set of all problems for which a given possible solution can be checked in polynomial time, but that finding a solution from scratch could possibly take longer than polynomial time.
If it were found that P=NP then many "hard" problems could be shown to be actually fairly "easy".

No way in the world can DC take this one on.

5. Poincare Conjecture

This is the one Grigoriy Perelman has probably solved.

6. Riemann Hypothesis

ZetaGrid is already tackling this one. It will never prove it (though it may well disprove it by finding a zero in the wrong place). Look at the ZetaGrid homepage for more info.

7. Yang-Mills theorem

A theory called the Yang-Mills theory has been found that describes the behaviour of Quantum particles remarkably accurately. Mathematically speaking, this theory is standing on very shaky ground. It doesn't appear to be based on anything much. Finding a foundation for it would be a big step in our understanding of the quantum world.

DC is hopeless when it comes to developing mathematical theories. No go here.

[jonnyw]

Interesting stuff.

Thanks for that

[jasong]

i don't know if this counts, but...

I read in Discover magazine that the number of possible Unified Field Theories is approximately 10 to the 100th power. They didn't explain their reasoning, but I'm guessing they took all the major questions and all the possible answers and did a little multiplying.

I'm not sure if this could be a DC problem, but a program made to come up with a way to divvy out funds to approximately minimize total time and money to attempt to find the answer would be very nice. Of course, quantifying many of the factors involved might be a problem.

[jaat]

For those who haven't seen Ramsey numbers before, here's a introduction through a puzzle:

Show that in a group of 6 people, there are always 3 mutual strangers or 3 mutual friends.

(Either you are stranger to a person or a friend. Also, if A is a friend of B then B is a friend of A)
The above claim just says R(3,3)<=6. You can show the other inequality as well.

R(4,4) =18.
So, among 18 people there is a always 4 mutual strangers or 4 mutual friends. Also, you can set up relationships among 17 people so that there isn't any group of 4 mutual strangers or 4 mutual friends.

R(5,5) is not known.

I am actually surprised to see no DC project around to calculate various Ramsey numbers as they are within grasp(at least the first few) and it would be very interesting for mathematicians as well who have been trying to pin these down for quite some time.
A latest survey is
http://www.combinatorics.org/Surveys/ds1.pdf (Table in the first few pages is interesting. I would rather not concentrate on stuff after that table.)

For those interested to know more about ramsey numbers:
http://mathworld.wolfram.com/RamseyNumber.html


I think the best bet and the most interesting number is R(5,5). Presently, it is known that 43<= R(5,5)<= 49.

Also, Paul Erdos once remarked that if aliens come and say they will destroy earth unless we tell them the exact value of R(5,5), then we should get down immediately and deploy all computers in the world to calculate it but if they ask for R(6,6) then we should figure out ways how to attack them.

A simple attack would just go through all 2 colorings for complete graphs of size 42 to 49 and check for monochromatic cliques. Surely, smarter ways could be found but it seems that calculating R(5,5) is custom made for DC.

Jaat

[Nexus]

Good first post, jaat. I agree. Ramsey numbers are just the sort of thing Distributed Computing would be good at figuring out.

[magnav0x]

Nexus a fusion reactor is in the works as we speak, but different countries are having different views on where it should be placed, which has stalled it for the past year....most blame the US, Japan, and Korea on the stall. They mentioned using a 100-million-degree gas (plasma) suspended in an doughnut-shaped magnetic field for the reactions to take place. 500MW of fusion power in 6 minutes would definately make any other energy plant obsolete

[gopher_yarrowzoo]

500MW gonna need a big capacitor to store that much danr and erm 100 Million Deg Plasma - hmm can we say instant combustion to anything organic that touches it... *sniff*

darn mathamatical theorms prove x^y can =y^x etc..