[QUOTE=chalsall;387625]Another idea (proposed many times) would be that any credit given for a "poached" assignment be given to the original assignee. Including (the /very/ unlikely) newly found Mersenne Prime.

This might satisfy slow "credit whores", and those who simply want to see the waves advance.[/QUOTE]

That's not a bad compromise... Basically someone who just can't stand having that one outstanding assignment that keeps some milestone from being reached, yeah, they can go ahead and "donate" their own computer time to that other person who actually has the assignment, giving that original assignee the credit.

The only thing is, it kind of sucks if you poach someone's double-check... I mean, if you poach a first time, then at least when the original computer checks in, then hey, it's a double-check, which is cool and all. But if you poach a double-check, then we'd assume the original assignee would check it in at some point which would usually be a useless triple-check. It would have been better if the poacher just got a regular double-check assignment and let that straggler come in on it's own.

This all assumes work is being done on the assignment and the recycling rules don't have anything that lets things fall through the cracks. I'm with George though in thinking that once the grandfathered work finally peters out, it'll improve things in general.

How about a sliding scale corresponding to the length of time it's been reserved? That is, if you poach one that's been reserved for 10 days, most of the credit goes to the original reserver. If it's been 300 days, the poacher gets most of the credit, but some goes to the original person. You don't want to give all credit to the original reserver so as to discourage people from reserving and then doing nothing, but you also don't want to have people snatching recently reserved exponents.

[QUOTE=Primeinator;387643]I've never had this combination... and it sounds utterly life-changing :w00t:[/QUOTE]

It is. If you haven't, do.

I agree that the "Assignments below X tested once" milestone is important for showing progress, and this is the reason we even have the milestones to begin with.


Regarding the poaching stuff: Why does a rational person poach assignments? Nothing we do is going to change the behaviour of irrational folk anyway, so let us ignore those.

A rational person will poach an assignment to steal the chance of finding the prime or to make progress on a milestone. Credits are irrelevant because short of knowing the residue beforehand, you had to put in the GHz-Days to get the result anyway.

Giving all credit to the poach[B]ed[/B] is a good start. Denying that the poach[B]ee[/B] ever did anything will be a deterrent to anyone trying to steal the prime discovery. "I found it! I found it!" "Nope. Our records do not show that you did." It will also be a deterrent to anyone who is concerned with getting as many credits as possible and does not wish to give away "free CPU work".

Anyone still not intimidated by these downsides will probably be submitting Anonymously anyway, so the credit penalty will likely not scare anyone.

The only rational people left, then, are the ones purely seeking milestone progress. The rules we put in are going to kick in fully in three months and their effect is intended to say that all 1,500 (is that correct) of the assignments most relevant to a particular milestone will be finished within a short time (2 and 6 months I believe for DC and LL respectively).

The only way to deter anyone who is STILL too impatient is to refuse the submission of poached work (a BIG no-no) altogether, or as George suggested, to hide the details of the front X assignments so that nobody knows which 33 (for example) assignments are holding up the queue. I want to say that this is going to be unnecessary with the new rules at full power but perhaps it is worth withholding data from the public.

Let us all wait now till the assignments under the old rules expire, which will be soon anyway. The revolt that lead to the new rules was started because of more than two years of backlog piled up in huge amounts. The current amount of backlog is only a tiny fraction of that. Time is on 'our' side now, so the best thing to do is wait and spent our resources elsewhere. I don't think we need any new rules now.

I agree with you. From Feb/2015 on, the "excuse" for poaching will be virtually none, providing the rules will be really enforced. For me that´s a very important point. We definitely need a strict enforcement of the agreed rules. And, if and when they are enforced, I reckon George´s argument about math wil no longer really matter. The assignment will be finished sooner than later, if not by Mr X (the poacher, whose result was refused), by Mr. Y (the one who gets the reassigned exponent in a reasonable time, due to the new rules). That is to say that I insist that poached assignments should not be accepted by the server, once the grace period is over and the new rules are in full swing.

@Madpoo: Btw this is a good option for another milestone: "Number of Grandfathered assignments left" :smile: Or to be more diplomatic: "Number of exponents assigned prior to 2014-03-01 (or March 1st, 2014)".


[QUOTE=lycorn;387670]I agree with you. From Feb/2015 on, the "excuse" for poaching will be virtually none, providing the rules will be really enforced. [/QUOTE]

Actually the last grandfathered exponent will last much longer. According to George's code an exponent assigned 2014-02-28 and say 90% done will live for 365 + (90-10)*3.33 = 631 days (Nov 21st 2015) (and one 90% done will probably fall under the unspecified grace period because it is almost done).

[QUOTE=tha;387665]Let us all wait now till the assignments under the old rules expire, which will be soon anyway. The revolt that lead to the new rules was started because of more than two years of backlog piled up in huge amounts.[/QUOTE]
About my original post about the 3 exponents in question:
I mentioned monitoring them to see if they really will complete in a reasonable amount of time, before acting. Most seem to agree that waiting and watching these is good.
A prior poster pointed out that all three [I]should[/I] be finished before the end of the year. That is the time frame that I mentioned.
These are first timers, thus any work turned in by someone other than Anon and the other could be viewed as an early DC, with no loss to those that hold the assignment.
I suggested a particular person of good repute and good machines as the designated actor. Firstly, they could watch the progress of the original assignee and hold off their report. Secondly, they are not a credit whore, no issue there. Thirdly, if on the tiny-tiny chance that one of the numbers is the next MP, they could communicate that to George, privately, letting the assignee be the one to report it to PrimeNet. I feel they are such character that they would 'do the right thing' in the circumstance. If one person is semi-sanctioned to do this, then everyone else should not be tempted and no undue worry or extra effort will occur.
The only reason I made the suggestion was to make it all nice and tidy at the end of the year.

The 10M digit DC milestone is nice, but that and the future ones will be taken care of soon enough by the new assignment rules.

[QUOTE=Uncwilly;387676]I suggested a particular person of good repute and good machines as the designated actor. ... If one person is semi-sanctioned to do this, then everyone else should not be tempted and no undue worry or extra effort will occur.[/QUOTE]

I don't know if you were thinking of me as the "designated actor", but if no one complains, I'd like to take on this role. My rational is:

1. 51907363 should have already been recycled according to the above SQL.

1.1. Beren has only progressed 0.3% since 2014-05-01.

2. 52957519 and 52983583 should be recycled around 2014-12-10, and they have only progressed ~10% since 2014-05-17.

Objections? I could process all three in just over three days on machines which have 100% success rate DC'ing.

[QUOTE=chalsall;387721]1. 51907363 should have already been recycled according to the above SQL.[/QUOTE]

I do not mind or complain, just want to point out it should not be recycled until Nov 19th or 20th:
2013-05-27 + 365 + (63.1-10)*3.33 ~ 2013-05-27 + 541 = 2014-11-19

[QUOTE=ATH;387723]I do not mind or complain, just want to point out it should not be recycled until Nov 19th or 20th:
2013-05-27 + 365 + (63.1-10)*3.33 ~ 2013-05-27 + 541 = 2014-11-19[/QUOTE]

Whoops -- you're correct. I erroniously calculated based on the assignment date being 2013-05-07 rather than 2013-05-27.

But, regardless, considering it's current rate of progress, it is definitely going to be recycled before it's completed.