Who gets credit for first 10M prime if 2 are found,same time

[cheesehead]

It's running the final Technical Beta Special Edition of Windows 98, which expires on April 1, 2001 (which, according to Microsoft [but not the U.S. Congress], was the first day of Daylight Saving Time that year ).

At the end of beta test Microsoft sent me the first retail W98 release, but I lost the CD somewhere in this apartment before I'd installed it, and I'm determined not to give M$ any more of my money.

[cheesehead]

But ... uhm ... maybe it's better to just agree that my computer is stuck in a time wrinkle. It's not nice to fool with Mother Nature. :)

[kwstone]

..or Micro$oft

[nomadicus]

Alright. I'll take a shot at this and see if some discussion can be had.

To me, it's all for nothing. How many years has it been between M39 and the present? You think two 10M primes are going to magically pop out at the same time within GIMPS? What are you smoking?

I think the more interesting senerio is someone within GIMPS and someone outside discover a 10M prime at the same time (still, extremely unlikely). What happens then?

For a senerio of two 10M primes found within GIMPS. . .
Within GIMPS, I would think the 10M would be the same number. Yes???
If the two submitted numbers are the same, most likely someone poached it and I wouldn't give the prize to the poacher. (Oh no! I brought up poaching again! My appologies to all that thought the poaching topic was dead :( ).

john

[eepiccolo]

Originally Posted by nomadicus

To me, it's all for nothing. How many years has it been between M39 and the present? You think two 10M primes are going to magically pop out at the same time within GIMPS? What are you smoking?

Nothing. Just wanted to make interesting discussion.

[Maybeso]

Originally Posted by eepiccolo

Hmmmm..... Earliest verifiable discovery date? I'd like to see how a discovery date from Prime95 is verifiable, if that discovery date is when the LL test finishes, considering what has been said already about the time a computer is set to. So does that mean it goes back to who reported it first?

Well, for starters, doesn't primenet store several timestamps during the lifecycle of an exponent? The discovery date can't be before the exponent was issued. I would think that by looking at all of the dates associated with the exponent in question, and comparing them to the dates for other exponents issued to the same user, a pattern could be established for:[list]If the user usually accepted exponents from the server, or grabbed exponents beyond the servers range.
If the users progress was predictable or sporatic.
If progress was predictable, the apparent speed of the users computer(s).[/list:u]Also, by running a few iterations of the exponent on the users computer, it could be roughly estimated when they started the LL test; and that could be compared for consistancy.
And it would be reasonable to expect an even distribution of dates between the results file date and the present, for other files on the computer.
If the user refused to supply this evidence, the prize would go to the other submitter. That burden of proof stuff.

Sorry this isn't more heated, :( should I inject a little sarcasm? A few demeaning rhetorical questions?

[kwstone]

Originally Posted by nomadicus

To me, it's all for nothing. How many years has it been between M39 and the present? You think two 10M primes are going to magically pop out at the same time within GIMPS? What are you smoking?

I think the original question is legitimate. Right now big chunks of Europe are going on 3 or 4 weeks summer vacation. It's not beyond the bounds of possibility that one of those guys comes back in September and finds his box has been beeping for the past fortnight. So he logs on and finds that a different 10M prime has been discovered by some hardworking individual in China, who already reported it. In this case, he's going to be pretty cheesed off. He may legitimately and truthfully claim that his log shows that his machine found the number first.

However, I'm still for giving the prize to the guy in China. <obligatory heat> And not just because all those Europeans are a bunch of effete, freeloading, outdated colonialists who think they are the only people who can do math. </obligatory heat>

I think the operative word here is not "verifiable" but "discovery". IMHO, if a machine calculates the largest prime number and sits there beeping for a hundred years but no one reads it off the screen, or finds it in a mass of printout, then that number has not been "discovered". A machine cannot discover in any true sense. Discovery requires a human to communicate the knowledge to other humans. Therefore reporting is the acid test.

I quote from Chris Caldwell's Prime Pages about an earlier dicovery http://www.utm.edu/research/primes/notes/by_year.html

"Surprisingly Hurwitz knew about M4423 seconds before M4253 (because of the way the output was stacked). John Selfridge asked "Does a machine result need to be observed by a human before it can be said to be 'discovered'?" To which Hurwitz replied, "forgetting about whether the computer knew, what if the computer operator who piled up the output looked?" In the table below I decided that Hurwitz discovered the prime when he read the output, so M4253 was never the largest known prime."

[nomadicus]

kwstone,
Let's say I buy your argument under those conditions.

But let's say the person is on a full time internet connection and is running prime95 getting/reporting info to primenet on a regular basis.

The individual can be gone for several weeks but the full time connection and automated reporting to primenet would make it discovered (IMHO) without that person looking at it.

I don't know. I guess someone, somewhere will have to read it off of primenet to considered "discovered." Hmmm. . . .

[trif]

IIRC, M39 sat on the server for a day or more because the code that was supposed to mail George in case of a discovery was borked. Somebody finally noticed it in the Primenet reports and queried the mailing list. So if the discover had been away and not noticed that his machine was beeping, and the discovery sat unnoticed on Primenet for a day or two, does that mean the discovery date should be the date that a human finally noticed?

[kwstone]

It's kind of like that old Zen Buddhist koan.

If a man speaks in a forest, and there is no woman there to hear him, is he still wrong?

(I may have got the original mixed up a bit, but it was something like that)