Newer milestone thread

[Mr. P-1]

Originally Posted by cheesehead

And, as I have tried to explain, though not concisely in one phrase before this: The anti-poaching "rule" is an ethical derivation from the pledge GIMPS makes plus logical consequences of that pledge.

It's already very clear that you consider poaching to be unethical. I think that's a reasonable position, but beyond that, I'm not really interested one way or the other in discussing the ethics of the matter.

If that's all you're arguing, if your claim that there is a "rule" is just another way of saying "it's unethical", then there's no substantive disagreement between us. In that case I suggest you drop the misleading "rule" language and concentrate on persuading people to adopt your ethical viewpoint.

you were trying to use an exception to the literal phrase "you will be assigned an exponent that no one else has tested" to show that the supposed "rule" did not exist (which is what you've been contending). But that exception had no bearing on the ethical derivations I had been explaining. Therefore, your introduction of the "exception" was just a distraction from the ethical argument.

I do not agree that the premiss "you will be assigned an exponent that no one else has tested" entails the conclusion "there is a rule prohibiting poaching". My main point is that your argument is invalid. As an aside, I remarked that the premiss is not completely true.

I apologize for using "dodge". I should have used "distract from".

Clearly my aside has distracted from my main point. That was not my intent.

... and, as I've been showing, a logical ethical extension of that exclusive-assignment effort is to oppose poaching. Such opposition can be considered a "rule", so there is, in effect, an anti-poaching "rule" which is the derivative of the written pledge.

As far as I can see, the logical ethical extension is at best that the GIMPS administrators ought to oppose poaching, and perhaps ought to impose a rule prohibiting it. It does not follow logically that the administrators do oppose poaching or that there is a rule prohibiting it.

But even the conclusion that they ought to do these things ignores the fact that there might be other relevant considerations. It's very clear that George declines to take a stance one way or another in this controversy. I strongly disagree that he ought to depart from that wise position.

The fact that George has so far declined to support your interpretation confirms that there is no rule in the 'administrator's fiat' sense of the word. Similar the lack of agreement in the discussion thread indicates that there is no rule in the 'social consensus' sense of the word. That only leaves a possible rule in the 'categorical imperative' sense of the word. As I said above, I've no interest in discussing ethics, but think you should drop the "rule" language if that is your only point.

My contention is that if GIMPS's operations are ethical, then as long as the assignment exists, GIMPS has a continuing ethical obligation to keep it exclusive.

Your position of abandoning exclusivity after the assignment is made makes a mockery of the pledge: "Here, we'll give you an exclusive assignment, but we won't lift a finger or say a word to keep it that way for even a moment more. We disavow any responsibility for that exclusivity the moment after you've received the assignment. If we did not make available the details (other than the ID string) of your assignment for all to see, no one could covet your assignment. But because we do make such details available to the public, we are providing would-be poachers with the information they need to swoop down and test the same exponent on a faster system. Although the latter action destroys the meaning of your assignment's exclusivity, we will make no peep of protest, nor will we dissuade any such action."

My point is that GIMPS delivers a lot more than it promises. It only promises that your assignment is exclusive at the time it is granted. In practice it remains exclusive until the assignee returns it (completed or uncompleted), or until it's deemed abandoned by reason of not having been checked in. Or until it's poached.

You think GIMPS could and should do more to discourage poaching: That's a reasonable position, but it doesn't follow that there is a rule against poaching.

There are consequences to actions such as GIMPS's publication of assignment details, but you think GIMPS has no ethical responsibility for those consequences -- is that correct?

No that's not correct. I think there is no rule against poaching.

That would all be nice in a utopian world where no one wants to interfere with someone else's work.

I do not claim that this is nice.

[Mr. P-1]

Originally Posted by ellipse

By "banning", I meant that their accounts are frozen. They won't get any credit for assignments turned in, they won't be given new assignments, and any results they turn in are rejected.

I agree that refusing results goes against the spirit of GIMPS and other DC projects. But it's the lesser of the current and proposed evils (allowing users to poach exponents or removing progress reports and countdowns from the milestones page).

Removing these data from the milestone page would be so offensive to you that you would walk away from the project in disgust. Yet you advocate that GIMPS should post false data, i.e., data which count as untested, exponents which, in fact, have been tested???

[ellipse]

Originally Posted by Mr. P-1

Removing these data from the milestone page would be so offensive to you that you would walk away from the project in disgust. Yet you advocate that GIMPS should post false data, i.e., data which count as untested, exponents which, in fact, have been tested???

The wording could be changed. Instead of "Countdown to proving M(24036583) is the 41st Mersenne Prime: 585", the milestone page could say "Less than 586 exponents left to proving M(24036583) is the 41st Mersenne Prime". So if someone poaches M21065063 and no one else turns in a result, that statement could be left unchanged while still being true. Yes, it will be slightly misleading at times, but that is much better than removing it completely.

With that said, I will still walk away from the project if data from the milestones page is censored. I and many others have spent dozens of CPU years on GIMPS, and we're not even allowed to know how much the project is progressing? If that becomes the case, I'd rather contribute to Primegrid or another project which makes its data freely available to all users.

[Mr. P-1]

Originally Posted by ellipse

The wording could be changed. Instead of "[SIZE=2]Countdown to proving M(24036583) is the 41st Mersenne Prime: 585", the milestone page could say "Less than 586 exponents left to proving M(24036583) is the 41st Mersenne Prime". So if someone poaches M21065063 and no one else turns in a result, that statement could be left unchanged while still being true. Yes, it will be slightly misleading at times, but that is much better than removing it completely.

Well yes, your "Less than 586 exponents left" formulation wouldn't be too bad.

On the other hand, "All exponents below 31,494,937 have been tested at least once." would be pretty misleading if the real trailing edge was above 37 Million. It would be useless as an estimate for the real value of this metric and useless for keeping you in the project.

There does not seem to be much to commend the idea.

With that said, I will still walk away from the project if data from the milestones page is censored. I and many others have spent dozens of CPU years on GIMPS, and we're not even allowed to know how much the project is progressing? If that becomes the case, I'd rather contribute to Primegrid or another project which makes its data freely available to all users.

I have a lot of sympathy for your POV, but at the same time, didn't that boat already sail with the switchover to the V5 server? Before then the full database was uploaded to the mersenne FTP server at regular intervals, and you could download it then filter and search on whatever criteria you liked. Now we're limited to the status of hundred or so exponents at a time, and just the search options they've chosen to implement.

[davieddy]

Originally Posted by Prime95

I've sent an email to the user letting him know that his exponent reservation is likely to be poached. I'll let you know if I hear back from him.

OK George.
I suspect youv've read this thread as carefully as I have!

I repeat: I'll miss that "1" when it goes.
I hope your response will inspire the core (as we refer to him/her these days) to continue its enthusiasm/participation in GIMPS as much as it did me.

Can't you think of anything better to do? Better take another vacation

David

[Oddball]

one suggestion for an anti-poaching measure: Remove all the "progress to the next GIMPS milestones" links from the milestone report. In particular the two lines that say "All exponents below xx,xxx,xxx have been tested..." are almost an invitation to poach.

A distributed computing project should never hide information regarding the contributions of its members.

I've never poached or attempted to poach an exponent, but if the countdown info for the milestones or the "All exponents below xx,xxx,xxx have been tested..." info is removed, I will leave GIMPS immediately

Originally Posted by Brian-E

By the way, I support Mr. P-1's suggestion of not publishing the type of progress which encourages poaching. But I do understand that it's a radical step which might well be unpopular with contributors.

I think there's room for compromise on this issue. Keep everything as is, with two exceptions:

1.) When the number of exponents needed to complete a milestone drops below 10, say "Countdown to testing all exponents below MXXXXX once: less than 10" instead of giving an exact number. Remove it only when all exponents below that milestone are complete.

2.) Round down the lowest completed exponent. Instead of saying "all exponents below M31494937 have been tested at least once" say "all exponents below M31400000 have been tested at least once".

That should provide fairly accurate progress reports while also discouraging poachers.

[davieddy]

At least till we find another prime.

The "milestones" are dominated by " how many more before suchandsuch
is proved". If (like me, inter al) you find it more (maybe not a lot) interesting to monitor "The Gimp's Progress" by observing the change
in these numbers by the day/week/year, you would see that the "countdown to proving" any "Milestone" above 26M progresses at rough
the the same rate as all the bigger ones.

I would like to see a countdown to "Knowing the status" of all exponents
below 60M.

You may think of other ideas.

David

[davar55]

Originally Posted by davieddy

You may think of other ideas.

I still think Che..seH..d's argument is right - don't send the letter.
Poaching is just wrong, morals/ethics trump law.

[axn]

Originally Posted by davar55

I still think Che..seH..d's argument is right - don't send the letter.
Poaching is just wrong, morals/ethics trump law.

Holding up progress is wronger. Discuss.

[chalsall]

Originally Posted by davar55

...morals/ethics trump law.

Says who?

Jesus?

Or the "great" US of A?

[davar55]

Originally Posted by chalsall

Says who?

Jesus?

Or the "great" US of A?

Morals/ethics are determined by the nature of humanity
and the nature of reality.

Not by a single person of the past.

And the USA is indeed great.