Poaching

[pakaran]

Originally Posted by Prime95

I kind of liked the idea of masking the lower 4 digits of the smallest 50 exponents in the public report. It would be up to Scott Kurowski to decide as he's the one with access to the Primenet server. If Malfoy's poaching continues we will have to ask him to do this. Yes, you can get the information in other ways but it is a little bit harder.

Would you also change your own page, which in a sense only masks the lower 2 digits?

I think the goal is to make sure 'poachers' have to make some effort to figure out the 'missing' exponent - recall that at least one GIMPS user found a prime's exponent before the official announcement by process of elimination, and had far less info to go by than just all but the lower 4 digits of the exponent.

Regards,
Nathan Russell

[wfzelle]

Originally Posted by Deamiter

No, it should not be treated flippantly. But how much fame will the user who blindly runs George's program really recieve? Perhaps the discovery of a 10 Million digit prime will help a mathematician do something even greater. That's why I work for Prime95. Not to garnish fame for myself, or even for Prime95. Be a part of science to help advance our world, not to help advance yourself! Should we value our human nature above what we might be able to learn through working together?[/b]

I consider the chance of 'winning' a distributed project to be a positive thing for the community. Our human nature is that we want to win things and excel, what better way to do so than in pursuit of knowledge that benefits society? Basic stats tend to bring out the worst in people (cheating, poaching), but honoring a member of the community that was lucky enough to find a prime is fine. We cannot honor every participant, instead, a few men/women get selected. They know that they were just lucky, this is far different from scientific 'greed' where people take credit for the work of others and often feel superior to their colleagues. The process of selecting a winner by chance instills humility. The winner is not better than us, he is one of us.

[eepiccolo]

Originally Posted by Deamiter

Perhaps the discovery of a 10 Million digit prime will help a mathematician do something even greater. That's why I work for Prime95. Not to garnish fame for myself, or even for Prime95. Be a part of science to help advance our world, not to help advance yourself! Should we value our human nature above what we might be able to learn through working together?

I would just like to state, Deamiter, that this is basically along the lines of why I participate in GIMPS also, not just for fame. I was simply making a point, not necessarily from my own point of view. It may not seem right, but people do join just for a chance at fame.
And besides, my point still stands that if a number is poached, time has been taken away from the project as a whole, due to the poached wasting time.

[nomadicus]

Originally Posted by eepiccolo

And besides, my point still stands that if a number is poached, time has been taken away from the project as a whole, due to the poached wasting time.

eepiccolo - I like the point you made. It really rings true.
The sixth entry of this topic, by prime95, outlines a good solution. I don't think you can ever eliminate 100% of the waste. Nullifying the poacher is a good objective and perhaps the most reaonsable deterent.
The main reason I joined GIMPS: I have a fasination with primes especially palindromic (which Mersennes are Base2) and I didn't like the idea of my PC sitting idle 98% of the time. Fame is good too, don't get me wrong -- the prize money even better. The poacher effectively nullifies my contribution by duplicating my (or another person's) effort and slows down the time frame when new Mersenne primes are found.

[Joe O]

Originally Posted by Prime95

(BTW Malfoy did not give an email address for his user id).

So this forum is better protected than PrimeNet?

[Xyzzy]

Originally Posted by Joe O

So this forum is better protected than PrimeNet?

This forum has 3 possible levels of "security"...

http://www.phpbb.com/guide/phpBB_Users_Guide.html#section3_2_2_1

Look under "Enable account activation"...

I just happened to choose the middle option... I suppose the most lenient option would have worked as well...

[ADBjester]

Originally Posted by Prime95

Today GIMPS says you can have the exponent as long as you report in every 60 days and are making some progress. Unfortunately, some users have exponents scheduled to complete many years from now. This annoys users that like to see a nice orderly progress of milestones.

George, worse yet, a user who has checked out TWENTY exponents on a given machine, and reports even best-case progress on ONE of them.. updates the other exponents as "progress being made" as well.

If you want to solve the problem of "left behind" exponents, then exponent hoarding (even inadvertent) must end. If the client doesn't report an actual iteration on the exponent in question, it ought not be marked as "active". Then the longest an exponent can stay out in "limbo" while someone crunches other numbers is 60 days.

[ET_]

a user who has checked out TWENTY exponents on a given machine, and reports even best-case progress on ONE of them.. updates the other exponents as "progress being made" as well.

There are still some v.18 (and less) clients on the wild.
They tend to require exponents continuously to the server.
Most of them have probably been forgotten by the original owner, but still work. I myself got one of those beasts :-(

Luigi

[Joe O]

Looks like M21699407 was in the way of some factoring milestone!

[Thu Jan 16 12:21:57 2003 - ver 22.10]
Updating computer information on the server
Sending result to server for exponent 21306913
Sending text message to server:
UID: Team_Prime_Rib/JMO_K6III400, M21306913 no factor to 2^66, WY1: CC838DCA
Sending expected completion date for M21778201: Feb 03 2003
Sending expected completion date for M21699407: Feb 22 2003
[Tue Jan 28 22:49:50 2003 - ver 22.10]
Updating computer information on the server
Sending expected completion date for M21778201: Jan 30 2003
Sending expected completion date for M21699407: Feb 18 2003
ERROR 11: Exponent already tested.

[crash893]

Originally Posted by Kevin

couldnt you just put a extra option to trade exponents

it would be just like unreserve exponent but you put a destination username in there

This is a good idea ( I believe it has been mentioned), but there is a problem. Some people from different teams trade exponents for various reasons. So while there would be no wrong-doing, it appears to be a poach to the server.

To throw out another idea I brought up a while ago (not sure if it's came up again) was to have the server hold expiring exponents <8M, or whatever barrier we may be attempting to pass. Then George could manually pass them out to people who will get them done quickly. This would at least prevent hold-ups because a slow person got a trailing edge exponent when it expires.

[NickGlover]

Originally Posted by crash893

couldnt you just put a extra option to trade exponents

it would be just like unreserve exponent but you put a destination username in there

Putting two ideas together here, it doesn't seem like trading of exponents needs to be "explicitly" implemented. If we used George's idea (from the "Replacing Entropia's server" thread) where each checked out exponent had its own key generated and written to worktodo.ini, then we can allow the exponent to transfer to a different user if they do a check-in with the correct key. The server can assume that the original owner approves of this transfer because it is unlikely that the new owner could have gotten the key without the original owner telling him.