GIMPS fundraising/prizes as incentives

[ixfd64]

For non-profit organizations, web hosting costs can be pretty high. That is one of the reasons that the GIMPS forum has limited functionality, as having many features definitely consumes more resources and thus raises bandwidth costs.

One thing users could do is donate money to GIMPS. However, as there is no formal way of donation (probably except for working out a deal with George via e-mail), there aren't going to be many users who will donate.

There are a few ways for GIMPS to make money (for covering its costs) and remain a non-profit organization. One possible way is to offer more forum functionality for donors. Personally, I think that a reasonable donation for enhanced services would be a one-time payment of $30.

Free members:

• search limited to an interval of 60 seconds
• no automatic signature
• 1,000 private messages
• limited poll options (5)
• limited functionality

Supporting members (proposed):

• no search limit
• automatic signature
• 5,000 private messages
• more poll options (up to 10)
• special "Supporting Member" rank
• increased functionality (blogs, etc)

There could be other GIMPS-only features that other forums don't offer, such as dynamic images that show users' exponents and their progress, that can be placed into the signature.

Another idea for fundraising would be to sell GIMPS merchandise. I mean, who could resist a t-shirt with a picture of Marin Mersenne with something like "Because bigger is better" as the caption?

Now, on to the next part:

Last year, there was talk about GIMPS offering payment for sending in results. The good thing is that this would encourage more people to join, and many users would be happy at getting some money back for their computing costs. On the other hand, I personally oppose this system. GIMPS is supposed to be a volunteer project, and I feel that money should not be involved. This would also attract all kinds of unwanted people who are just looking for money. Finally, the costs of paying users for their work are very high, and I don't think that we want to make GIMPS look too commercialized anyway.

However, I do have an "in between" idea. Perhaps we could hold "raffles" (prizes donated by members) for those who turn in work units. Each work unit would count as a "ticket," and once used, cannot be used in another "raffle."

To encourage other activities besides LL-testing, there should be prizes for each category (LL-testing, double-checking, and trial factoring). Each TF-ed exponent (without factor) would count as one "ticket" and each exponent cleared by TF-ing would be five "tickets."

Example prizes:

• TF-ing prizes: GIMPS t-shirts or other similar merchandise (20 winners)
• DC-ing prizes: $100 gift certificates (10 winners oer raffle)
• a new processor (2 winners)

What do you think? Just a crazy idea from me. :D

[Xyzzy]

Please note that the the forum is a separate entity from the GIMPS project, although if George says to jump, we do. (We can't remember him asking us to, though.)

The forum has functionality built in to accept various forms of payment and various levels of membership.

http://www.vbulletin.com/docs/html/subscriptions

Most of the ideas you suggested for higher levels of membership are not available, but there are a few that are.

However, the forum only costs a few dollars a month. To get all the goodies turned on we'd have to upgrade to a dedicated server that would cost more than a hundred dollars a month. Plus, it would be a lot more work to manage, and more work is always bad in our book.

Personally, we like the forum the way it is now. It is inexpensive, fast, simple and the content of the forum is not obscured by excessive features. Is it really worth spending a lot more money to get a little bit more speed or functionality?

We've tried to not limit our users in any way unless it directly affects performance. Members already have 1,000 private messages, which seems like a lot to us. All users are pretty much treated equally, except moderators get a few additional editing options and super moderators get 10,000 private messages and, of course, keys to the magnificent Hasselhoff forum.

We have forum maintenance and upkeep down to about an hour a day. The forum runs stable, is up 99.9% of the time (hidden at the bottom of every page is a link to our uptime stats) and the cost is easily hidden amongst our stack of monthly bills that we pass off to the wife.

That's our opinion. We're lazy and we have a menagerie of assistants to help us and they don't like change either. Especially the trolls. If you think Bob is gruff then obviously you've never lived with a family of trolls.

PS - If you need more poll options or private message storage, let us know. If you can make a good argument for it we can try to make it happen.

[Xyzzy]

Here is a current list of our private message usage. The first column is the number of messages and the second column is the number of users who have that many.

421  1
383  1
281  1
265  1
215  1
192  1
184  1
160  1
152  1
144  1
141  1
137  1
123  1
116  1
111  1
110  1
109  1
107  1
104  1
101  1
 99  2
 96  2
 87  1
 84  1
 83  1
 82  1
 81  1
 78  1
 76  1
 75  1
 74  1
 73  1
 72  1
 71  1
 70  2
 69  1
 67  1
 64  1
 63  4
 62  2
 61  2
 60  1
 59  1
 55  1
 54  3
 53  1
 52  2
 50  1
 45  1
 43  3
 42  1
 41  2
 38  1
 36  1
 32  3
 30  1
 29  5
 28  4
 27  6
 26  2
 25  4
 24  1
 22  5
 21  7
 20  5
 19  9
 18  6
 17  8
 16  9
 15  3
 14  6
 13 10
 12  5
 11 10
 10 16
  9  8
  8 10
  7 15
  6 22
  5 26
  4 19
  3 45
  2 62
  1 76

[jinydu]

I personally don't like the idea of rewards for turning in results because it may encourage cheating. As it is now, the only incentive to turn it fake results is to move up the rankings. But it might be a lot worse if there is a material (especially monetary) incentive to do so.

[Fusion_power]

If there is a real need for money to help run the show, I would prefer to see a donation based system. We had a donation round a few years ago to buy a new computer for George to test code on. I don't see a reason why we could not do that again. I contributed a few dollars for that computer and would contribute again if needed.

Paying for exponents completed is IMHO not worth considering. It would run into quite a few dollars, there would be questions if - as some of us do - someone else's computers are used, it would attract people who were not really serious about doing the work increasing churn of exponents. This is too much trouble for any gain made. Besides, it would reduce my chances of winning the big prize!

There are very few incentives for "increased functionality". One major forum I frequent uses an advertisement blitz on all non-contributing members. You pay the danegeld and you gets rid of the ads. It didn't take long for some of us to come up with ways to block the ads eliminating that problem. Modifying the hosts file on your computer or using Firefox with adblock are easy to implement. I would hate! for this forum to become an advertising nightmare. From my perspective, this forum is pretty nice as is, barring the occasional gremlin or even troll attack.

Just for curiosity, xyzzy, would you mind posting (or maybe pm'ing me) the amount it costs to have the forum per month? I'm curious because my website costs $5 per month or $60 per year. This is entirely reasonable from my perspective.

Fusion

[ixfd64]

I definitely agree that such incentives are not worth the trouble. On another note, there are EFF prizes of $100k, $150k, and $250k given for the discovery of the first known primes with at least 10m, 100m, and 1b digits, respectively. GIMPS may be within a year of winning the $100k prize, half of which will be given to the user who discovers the winning number. The EFF prizes are given by an outside entity and GIMPS isn't guaranteed to win them, although it is commonly believed that GIMPS will win at least the $100k prize.

What I think that GIMPS needs is more publicity. Since GIMPS is a non-profit project, other websites might be more inclined to offer us free/discounted advertising.

[Mr. P-1]

Originally Posted by Fusion_power

...I'm curious because my website costs $5 per month or $60 per year...

If your site uses less than 5GB/Month bandwidth, and doesn't require a huge amount of filespace, then have a look at https://www.nearlyfreespeech.net. They charge no minimum monthly fee, only for what you use.

I'm not connected with the company, other than as a satisfied customer

[R.D. Silverman]

Originally Posted by ixfd64

What I think that GIMPS needs is more publicity. Since GIMPS is a non-profit project, other websites might be more inclined to offer us free/discounted advertising.

Please explain *why* you think GIMPS needs more publicity.
What would it accomplish?? Current resources certainly seem adequate.

I view the cash prizes with hostility. People should participate out of
interest and not because they hope to win a lottery.

If the purpose of the cash prizes is to spur research into better methods
than LL, I see no evidence that this is happening. Indeed, the few
people who are competent to do this kind of research do not do it for
money. Furthermore, there are strong arguments that LL is theoretically
the best possible algorithm already.

[TimSorbet]

Originally Posted by R.D. Silverman

Current resources certainly seem adequate.

There are a few things you can never have enough of, and one of them is people/computers contributing to GIMPS. Are you saying that we don't need or want more people to join? Imagine how quickly we could find new primes of ever-increasing size if there were many more people in GIMPS.

[R.D. Silverman]

Originally Posted by Mini-Geek

There are a few things you can never have enough of, and one of them is people/computers contributing to GIMPS. Are you saying that we don't need or want more people to join? Imagine how quickly we could find new primes of ever-increasing size if there were many more people in GIMPS.

And??? So we find more primes. So we do it a little quicker. BFD.

Neither finding more primes nor doing it faster will add to our UNDERSTANDING
of the mathematics. So what value does it have???

If, on the other hand, participants want to improve George's code, I am
all for that.

Computational number theorists really don't care about new Mersenne primes
unless it leads to better METHODS. Simply throwing more machines at the
search does nothing.

I for one certainly do NOT want more people to join, especially if they are
doing it for money. It takes ZERO intelligence to take someone else's code
and run it on a computer, and makes no intellectual contribution to mankind's
knowledge base.

I also believe that the $100K prize for a 10M digit prime should only be given
if the discoverer wrote his/her own code or contributed in some way to
the GIMPS code. Otherwise, the prize is nothing more than a lottery.

We should encourage intellectual contributions to the art, and not just
the blind running of someone else's code. (hoping thereby to get lucky)

[akruppa]

Originally Posted by Xyzzy

421  1

Alex