Mersenne: Weighing in on Blosser
Hi all, As you've noticed, I've stayed out of the Blosser controversy on this mailing list. A few reporters have called and I've tried to let them know Mr. Blosser's side of the story from his public statements on this mailing list. After listening to the various arguments, here's what I'll be doing: 1) I've rewritten the freesoft.htm page to include a bunch of disclaimers - one of which is GET PERMISSION. This warning used to be in the readme.txt file. But lets face it, that isn't the most widely read document. I've also taken this opportunity to make installation easier. I've also begun a long overdue FAQ. Comments are welcome. 2) As I cannot be the permission cop, Mr. Blosser's results will stand for now. If he is charged and convicted of computer fraud, then I'll reassign his results to "anonymous". This will occur even though his results predate the May incident. This will be my policy in the future. Anyone convicted will have ALL his results reassigned, no matter how they were obtained. I don't expect this "harsh" threat to deter anyone, that's the purpose of loss of job and possible criminal prosecution. 3) I've removed the link to the Eurocrypt project. It smells of hackery. Personally, I hope that cooler heads prevail and USWest and the FBI drop their case. From Aaron's statements it seems he acted in good faith and did little to no harm. It would seem to be a weak case to charge someone for theft of computer cycles that USWest was discarding anyway. Firing Aaron lets USWest use him as an example to others - and get the word out as to their computer policies. They should probably thank him for uncovering a security hole that some truly disgruntled employee could have used to create some SERIOUS havoc. Still having fun, George
Re: Mersenne: A short gdunken on Aaron B's situation
>I don't really care about the Blosser bit. I just >wish the PrimeNet server would be more reliable. agreed. I discovered most of my HTTP clients at work were busy retrying ever 2 minutes and not getting any work done today. Grrr. And several have results they are trying to post, gr. -jrp
Re: Mersenne: factor found but factorization continues? why?
> For the factor to be about 630, the exponent could be at most about > 315, since the factors of mersenne numbers are of the form q=2kp+1. > (if q=630, for k=1, p=~315, and for larger k, p is smaller) > > But nearly all numbers below 331 have already been not only tested > for factors to at least 2^50 (with most exponents factored up to 2^55 > or more), but Lucas-Lehmer tested also already (except for 72). > > I think it is much more likely that the exponent tested is p=~630, > and the factor found for it is somewhere between 2^55 and 2^62 or so. I calculated a k value of approximately 366,006,826,859 for factoring 2^6,300,000 to a range for 2^62. 3,310,000 values of k would take approximately 11 seconds to factor on a P200. Chip Kerchner
Re: Mersenne: A short gdunken on Aaron B's situation
> others have also called this uncalled for. > As several have mentioned, I was > indeed recalling the rejection of the > "scientific work" by so-called > "scientists" of Germany and Japan during WWII. > These results were tossed out by the Western > powers. The case here is not as serious, not as We are way off topic here, but this has been repeated several times and it isn't true. One of the experiments Nazi scientists performed was to immerse Russian prisoners of war in freezing water to measure how long they would last before dying. Data such as this is apparently still used today, because it is the only source of such data, and it may be helpful in planning rescues at sea, designing protective gear, and saving lives. But it's irresponsible to even hint at any analogy between this and the current situation. If such was not the intent, then what exactly was the point? Introducing the word "Nazi" into any discussion, whatever the context, is rarely helpful. The idea of tossing out mathematical results is not only premature (the facts of the case are still in dispute and under investigation) but silly. If Blosser had actually discovered a prime, would we choose to somehow forget its existence? PS, Someone else wrote: > > the GIMPS server > > machines, which [the FBI] would be entitled > > to by many grounds, weren't confiscated. The PrimeNet Server? I hope this was a joke in poor taste. There are no such grounds for the FBI to take any such action. They may be heavy-handed at times, but they must operate within the law.
Mersenne: Outta here
Not that anyone will notice (I'm somewhere int he 3000's), but I have had enough of the whining regarding whether or not to use the results. I thought I was contributing to a process involving mature individuals. I'm sorry to see that this is not the case. I will remove all my systems from the search after the current LL tests complete. George, that means you can have back the Factor range from my UNIX systems. Clark B. WierdaURL: http://www.io.com/~cbw/[EMAIL PROTECTED]
Re: Mersenne: A short gdunken on Aaron B's situation
From: Mikus Grinbergs <[EMAIL PROTECTED]> >In article <[EMAIL PROTECTED]>, >"Vincent J. Mooney Jr." <[EMAIL PROTECTED]> wrote: >> >> Well, we could argue this for a long time. I vote for discarding the >> results and asking GIMPS to warn its participants to never do this again. >> GIMPS should not credit Aaron for the work. > >I second this proposal. > >Let me explain my reasoning: GIMPS should have a _policy_ of what >are appropriate CPU-cycles, and what are *inappropriate* CPU-cycles. >Then GIMPS should enforce that policy by NOT ACCEPTING the results >of *inappropriate* CPU-cycles. I disagree. I think Aaron should keep all credit for the work done. Aaron asked for permission and got permission. Unfortunately he didn't have everyone's permission that mattered. >My concern is with "how does GIMPS look to the witch hunters?" If >GIMPS does not have a _policy_ of not accepting *inappropriate* >CPU-cycles, then we risk being tarred with "THOSE people don't care >whose CPU-cycles are taken, or what the consequences are!" Maybe GIMPS does need a policy. We shouldn't accept results from a hacker who broke in to a system and installed prime95 without permission. But the policy shouldn't apply to Aaron's situation. He worked for the company. He had permission from someone. He used an admin password that was given to him. I think that if the Prime Server hadn't run out of exponents back in May(?), US West would have never known about it. Aaron would still be getting results today. Wayne Sheppard
Mersenne: Worldwide GIMPS
<> No, it shouldn't. GIMPS, or some distributed computing project, should be on every computer in the world.
Mersenne: Primenet Server down?
Hi All, Is the Primenet server down? My prime95 program has been trying to connet to primenet for some time now. I am getting the message: Contacting PrimeNet Server. ERROR 2250: Server unavailable Will try contacting server again in 1 minutes. Chuck Baker --- Visit our homepage at: http://www.primenet.com/~cbaker ---
Re: Mersenne: A short gdunken on Aaron B's situation
> Let me explain my reasoning: GIMPS should have a _policy_ of what > are appropriate CPU-cycles, and what are *inappropriate* CPU-cycles. > Then GIMPS should enforce that policy by NOT ACCEPTING the results > of *inappropriate* CPU-cycles. The problem with having a policy on correct and incorrect CPU cycles is that GIMPS can no longer claim impartiality. As soon as we abandon the position that results are results no matter where they come from GIMPS is forced into the position of acting as a computer cop. Certainly GIMPS doesn't have the resources to investigate where all the results are coming from. I am willing to bet that a significant number of results come from machines with little or no permisions. For instance all the useres who run NTprime on their office NT machine...must we discard those results because technically these people don't own the computer so don't have decision powers. Does someone in GIMPS have the time to go investigate to decide on the correct policy in all these cases. What if GIMPS has a _policy_ but in some instance GIMPS decides it can use the cycles while the government claims its illegal (i.e. we disagree with the results of a trial or just investigation). GIMPS will take much more heat if it has a policy which is portrayed as a "pro-hacking" policy than if GIMPS represents itself (like journalists themselves when they use stolen data which they themselves didn't steal) as an impartial collector of work. That was the more practical reason. In addition one of the key principles behind science and mathematics is its universability. It does not matter who does the reaserch. If we throw out the results it seems we are somehow violating the spirit of that principle (we shouldn't have thrown out the NAZI results either. Yes they were incredibly evil but how would you like explaining to someone that they can't recieve a mdeicine because the NAZI's killed that persons grandparents). I vote keep them Peter
Mersenne: Primenet server not available?
Hi, I've been trying all morning to connect to the Primenet server to update/obtain new work... is it down? My system keeps trying to open connections to hrothgar.grendelnet.com:2587 but I am only getting an ERROR 2250: Server Unavailable. This is with Prime95 v.16.4.1, and I've been using this method of network connection (ie. DSL line at friend's house) for a while now with no other problems. I know its not the network connection, as I can ping/tracert the site. Thanks, Sean ([EMAIL PROTECTED])
Re: Mersenne: Primenet Server down?
At 11:53 AM 9/18/98 -0700, Chuck Baker wrote: > Is the Primenet server down? Yes. Something is wrong. Scott is on a short trip, but his backup has been notified. The server was re-booted this morning, but the reports are still dated (Thu Sep 17 04:00 Pacific). He is currently re-booting again. We'll just have to be patient. > Will try contacting server again in 1 minutes. I suggest changing that to 30 minutes or more. --Luke
Re: Mersenne: On a lighter note...
>A bit off-topic, but in a recreational-math kind of way. Plus, >I think we all need to step back a few paces from this USWest >thing, take a deep breath, regain our senses of humor (assuming >we had such to start with :) ... Ernst... thank you so much... yes we all need to take a break. Much appreciated after I just downloaded the biggest daily dose of messages since my time on the list. I giggled a lot :) As for the new math problem: > - /\ > | | / \ (imagine a circle here, > | |/\ but that's tough to draw > | | / \ using ascii characters). > - -- >Question: what equation does the above represent? For one horrible moment, I thought all it could represent was an anagram of the Electronic Arts logo. >What could be more obvious? I am very very scared. It reminds me of my last year of 'junior school' in England in 1982. It had already been decided where I was going to go for the rest of my education, so I spent several months doing nothing in math but playing with compasses and constructing involute and evolute curves (I didn't know that was going on, and didn't really care, the patterns sure looked pretty). Oddly enough I'd say that was where I got into number theory as an escape from the mindless tedium. I tried to think of my answer to the square, triange, shape... in the same vein as Ernst's suggestions. 4-3=1 is the best I could see! And we are surprised to find the numbers of college math students are dwindling? Chris
Re: Mersenne: A short gdunken on Aaron B's situation
>We have been having quite a bit of discussion on what to do with >Aaron's results, given that U.S. West claims that he didn't have >premission to use the machines in question. Maybe I can end this right now. If you all think it's fair to remove the *credit* given to my user account, then so be it. Remember, neither I nor anyone else in the GIMPS project *should* be concerned with "credit". My goal was not one of ultimately becoming top producer. I saw what I thought was a decent way to help in the search for new primes. I knew I would be ineligible for the prize money since it requires *written* authorization which I didn't have. My only goal in this was, honestly, to, if not find a new prime, then at least to help eliminate thousands of other numbers from contention. So if you feel it's worth removing the credit from my account, then so be it. Bear in mind that *none* of the US WEST computers actually completed an LL test. There were some machines running trial-factoring that had, of course, finished up, so if Scott feels like it'd be worth the incredible effort of going through his server logs to find out which machines were responsible for how much credit, then more power to him. Remember, it's about finding primes! Not about something so petty as who is in the "top producers" page. At any rate, it wouldn't do any good to throw out the *results* of the work done (even if they were *only* trial factoring results). But I give my hearty approval that, if you all think my credit should be removed for the work done on US WEST machines, then do it.
Re: Mersenne: A short gdunken on Aaron B's situation
In article <[EMAIL PROTECTED]>, "Vincent J. Mooney Jr." <[EMAIL PROTECTED]> wrote: > > Well, we could argue this for a long time. I vote for discarding the > results and asking GIMPS to warn its participants to never do this again. > GIMPS should not credit Aaron for the work. I second this proposal. Let me explain my reasoning: GIMPS should have a _policy_ of what are appropriate CPU-cycles, and what are *inappropriate* CPU-cycles. Then GIMPS should enforce that policy by NOT ACCEPTING the results of *inappropriate* CPU-cycles. One parallel might be with the question of sexual harassment in the workplace. If a company has a _policy_, which it communicates to all its employees, and which it enforces, then a claim of sexual harassment against that company would require very strong evidence. My concern is with "how does GIMPS look to the witch hunters?" If GIMPS does not have a _policy_ of not accepting *inappropriate* CPU-cycles, then we risk being tarred with "THOSE people don't care whose CPU-cycles are taken, or what the consequences are!" mikus p.s. In this specific case, I don't have the answers. Apparently there were people at US West who knew and did not complain. And there were other people at US West who did not know, and who *did* complain when they found out. Now, which of these groups had the ultimate "right of approval" ? If it comes to trial, that could take months or years. How should GIMPS act in the meantime? If somebody authorized by US West tells the press "CPU cycles were stolen from US West", then I believe GIMPS **should** act on that statement.
Re: Mersenne: US West - not a 'prime' example
>> David Beigie, a spokesman for US West, said it was "an encouraging >>sign" that both sides were willing to continue talking through the night. >> The company has arranged for 15,000 managers to fill in Sunday morning >>in case of a strike, Beigie said. David Beigie is the same spokesman who stated quite clearly in the Denver Post article that (I don't have the exact quote) "Blosser was given access to the system". Hmm...so much for hacking or the charge of computer fraud (which must show *unauthorized* access). But, nevertheless, the strike occurred subsequent to the alleged slowdown in Phoenix back in May. Note, I still dispute (as may others of you) that NT Prime could cause a 3-5 second delay to turn into a 5 minute delay, hard drive thrashing or not. Bear in mind that these machines were Pentium 133's and 166's with 64MB of RAM apiece. The only software that these things ran was an XWindow server (RumbaX if I recall). Sigh...
Re: Mersenne: A short gdunken on Aaron B's situation
>GIMPS should not credit Aaron for the work. I hate to drag this thread out further, but isn't this question academic because Aaron's computers as US West were only running for two weeks? He said that wasn't enough to actually return any LL results based on the size of exponents he was running and the speed of the computers. Just my two cents, back to lurk mode. Greg Hewgill
Re: Mersenne: A short gdunken on Aaron B's situation
I don't really care about the Blosser bit. I just wish the PrimeNet server would be more reliable. Randy Given [EMAIL PROTECTED] http://members.aol.com/GivenRandy public key at http://members.aol.com/GivenRandy/pgpkey.asc
Re: Mersenne: A short gdunken on Aaron B's situation
As the one who started all this, I commend Seth L. Chazanoff's comments below. I realize this is for some an intriguing issue, hard to resolve. One person resolved it by a personal, and uncalled for, criticism of me; others have also called this uncalled for. As several have mentioned, I was indeed recalling the rejection of the "scientific work" by so-called "scientists" of Germany and Japan during WWII. These results were tossed out by the Western powers. The case here is not as serious, not as offensive, and is a good one for a different and lower level of ethics discussion. Aaron is not to be compared to the WWII "scientists". He is also very likely to be a really nice person as well. (Footnote: I seem to recall that the use of penicillan as very beneficial was give to Germany after 1939). My vote is to disallow the results when 1) The air is clear on what happened and if 2) Aaron cannot convince GIMPS he had ALL needed permissions. I emphasize ALL needed permissions. Like many of you, I too work in an environment where there are scads of idle cycles at night and on the weekends. But the company only lent me one machine for my use, not the whole network. Getting the network would be impossible, as it should be. A) Someone wants or needs to work at night or on the weekend. B) The GIMPS program is is on their machine and who is to maintain it? C) What about the firewall that the company has? A major problem is that of the appearance of a "hacker" in the unsophisticated minds of too many people running the network, running the company, or people otherwise involved (purchasing did not buy computers for GIMPS). I don't think that Aaron is even up to the level of a "hacker" yet I worry that the perception will be advanced by the unsophisticated (which includes uninformed newspaper writers who know far to little to write on this subject, probably using a laptop.) Maybe some of the unsophisticated are likening the event to the danger of spam. Of couse it isn't equal to that or even like that, but perceptions are important. It may well be true that Aaron Blosser caused no damage to anything or anyone. Yet that cannot be the sole deciding issue. Surely Seth L. Chazanoff's method of expression, improper use of comany resources, should also be used. This is where getting ALL NEEDED permissions comes in. Also we need to respect the secretary who says "don't put that thing on my machine" because the secretary does not know _anything_ about computers beyond the keyboard (but viruses are a worry). Again, ALL NEEDED permissions raises its head. Well, we could argue this for a long time. I vote for discarding the results and asking GIMPS to warn its participants to never do this again. GIMPS should not credit Aaron for the work. At 09:38 PM 9/17/98 -0700, you wrote: >We have been having quite a bit of discussion on what to do with Aaron's > results, given that U.S. West claims that he didn't have permission to use > the machines in question. > > Let's try this: > > Assume that I am your supervisor, and you come to me and and > make your pitch for permission to put a program on all of the department's >computers. I like it and I tell you go ahead. > >Someone else (an auditor? a P.Oed co-worker) comes along, discovers >this program on the computers, escalates a complaint to a "pointy >haired" someone who will listen to the complaint of the improper use of >company resources, and shows you the door for "stealing" the company's >valuable excess CPU cycles. > > Did you have permission? > > If not: > Did you act in good faith? > Should we throw out your results? > Should we not credit you for your work? > Vincent J. Mooney Jr. [EMAIL PROTECTED]
Re: Mersenne: US West - not a 'prime' example
jt <[EMAIL PROTECTED]> wrote... >Hasn't anyone else considered that the slowdown at US West may have >something to do with the fact that the 35,000 union people trained and paid >to run the computers were on strike, and thus... NOT THERE! ? > >Read the following email. Talks failed - there was a strike. I won't >comment on what I went through trying to get a phone line. Thousands of >others complained to the government. An investigation was opened and >announced last week. > > >>Email: [EMAIL PROTECTED] >>Date: 1998/08/15 ^ only, the events with A. Blosser occurred back in MAY. That was quite clear.
Mersenne: A short gdunken on Aaron B's situation
We have been having quite a bit of discussion on what to do with Aaron's results, given that U.S. West claims that he didn't have premission to use the machines in question. Let's try this: Assume that I am your supervisor, and you come to me and and make your pitch for permission to put a program on all of the department's computers. I like it and I tell you go ahead. Someone else (an auditor? a P.Oed co-worker) comes along, discovers this program on the computers, escalates a complaint to a "pointy haired" someone who will listen to the complaint of the improper use of company resources, and shows you the door for "stealing" the company's valuable excess CPU cycles. Did you have permission? If not: Did you act in good faith? Should we throw out your results? Should we not credit you for your work?
Mersenne: my big mouth Re: Aaron Blosser
I have been a lurker on the digest version of this list. But the manner in which the media has gone overboard on this issue is sickening I know for a fact that the vast majority of people who work in Fortune 500 IT depts are inept. I have seen them give up on issues and proclaim "it's a virus, etc.". Prime95 may have caused some problems, and the gentleman involved may have overstpped his bounds, but the manner in which the media controls the story must be stopped... We must manipulate the media rather than the opposite. Tomoorow they will reporting how many millions of $$$s Mr. Blosser cost US West, a company I must add made over 1billion in profits last year along with the other Baby Bells. I chose to call David Beigie, US West's director of communications. His name was in the original, error-ridden Denver Post story and the AP wire version carried on www.cnn.com, etc. His number is (303) 896-5528. I politely called during the night and left a message on his voice mail. In my message, I politely explained that the GIMPs project is not a one-man obsession, but a project of sentimental value to mathematicians, prfessional and amateur alike. While problems like this could be solved on supercomputers, no official bodies would fund such purely intellectual work. I also added that the GIMPs software was designed to run at a level not to affect user productivity. While not excusing unauthorized use, perhaps US West should research the problem before announcing "a hacker intrusion affected our QoS" with such authority. I then politely left my name and phone number if he had any questions. Surprisingly, he called back today personally and left a message on my answering machine. He said that he agrees the GIMPS Project should not be belittled (he used the word "noble"). However, he feels that they must pursue full prosecution because a precedent must not be set. US West's computers are asset's dedicated solely to their business mission. While I understand this fully, I agree with others that a better PR move would be to punish Mr. Blosser internally and then arrange for GIMPS to continue to use US West computers in a manner that does not affect their performance. Unfortunately, corporate knee-jerk is now in full swing. Mr. Beigie added that he agrees the media is hyping this story by using the word hacker. Later, a reporter for the Rocky Mountain News called me. My name and phone number were passed on fro US West. I explained that I was anobody in the GIMPS project (ranked 5000+). I described the errors in the Denver Post article. I directed him to www.mersenne.org. I tried to explain the basic premise of the various distribited projects on the web and how spare CPU cycles were being used. I tried to stress that this story was not the story of a hacker obsessed with numbers. I mentioned the recent movie "Pi" and how this story is being written with a similar theme. He basically let me ramble on and on while I heard him clicking notes in the background. I ended by telling him to be fair and not hype this story. Unfortunately, he said that an FBI search warrant is a serious thing and is news. Perhaps some of you in the top 10 should call John Akala(sp?). His phone number is 303-892-2666. I'd first ask him to prove he understands what a Mersenne prime is. But most of all, please be polite... hopefully the article won't say: Matthew Lewis, another gimp hacker, also like to search for the next 350-year old prime, when he's not downloading serialz and crackz from top-secret, kewl warez sites. mlewis Iteration:5225200/5478241. Clocks: 9655915=0.483 sec.
Re: Mersenne: factor found but factorization continues? why?
For the factor to be about 630, the exponent could be at most about 315, since the factors of mersenne numbers are of the form q=2kp+1. (if q=630, for k=1, p=~315, and for larger k, p is smaller) But nearly all numbers below 331 have already been not only tested for factors to at least 2^50 (with most exponents factored up to 2^55 or more), but Lucas-Lehmer tested also already (except for 72). I think it is much more likely that the exponent tested is p=~630, and the factor found for it is somewhere between 2^55 and 2^62 or so. Ken At 09:03 AM 1998/09/17 -0400, you wrote: >> >Computer running prime95 ver 16.4 intel 486 today found a >> >factor of the number it was processing, contacted primenet, >> >sent the result, obtained a new number to factor, and then >> >continued to factor the same number it had just found a factor >> >for. I do not have the files immediately to hand but the >> >number was in the 630 range and the factorization had >> >reached the 6th or 7th iteration of 16. Can someone tell me, >> >> The factoring test will continue on the chance that a smaller factor >> may be found. I believe this is in the interest of doing some kind >> of analysis on the smallest factors of composite Mersenne numbers. > >Well, is it aware enough not to try any factors larger than those >already found? In the case above, iterations 7-16 should go _very_ >fast if the iteration 6 factor found was only 630. > >-- Tim >
Mersenne: US West - not a 'prime' example
Hasn't anyone else considered that the slowdown at US West may have something to do with the fact that the 35,000 union people trained and paid to run the computers were on strike, and thus... NOT THERE! ? Read the following email. Talks failed - there was a strike. I won't comment on what I went through trying to get a phone line. Thousands of others complained to the government. An investigation was opened and announced last week. >Email: [EMAIL PROTECTED] >Date: 1998/08/15 >Forums: alt.society.labor-unions >Bargainers for US West Inc. and the union for 35,000 of its workers reported >progress in labor talks with a federal mediator Friday, but big differences >remained a day before a strike deadline. > The current contract between the regional telephone company and the >Communications Workers of America expires at 12:01 a.m. Sunday, and US >West's union workers have already authorized their leaders to call a strike >if no agreement is reached. > Both sides reported some movement in talks Friday. But disagreements >over mandatory overtime, scheduling and US West's proposal to link pay to >performance remain unresolved. > "I wouldn't call it [the movement] significant," CWA spokesman Bill >Thornburg said. > Thornburg said negotiations were expected to continue around-the-clock >ahead of the strike deadline. The union represents more than half the phone >company's 51,000 employees in 14 states. > David Beigie, a spokesman for US West, said it was "an encouraging >sign" that both sides were willing to continue talking through the night. > The company has arranged for 15,000 managers to fill in Sunday morning >in case of a strike, Beigie said. > Denver-based US West, the dominant local phone service provider in 14 >Western states, asked late Thursday for a federal mediator to take part in >the talks. The union agreed, and negotiators and the mediator started >working Friday morning.
