Re: [Re: Mersenne: Serious problems with v.18]
Definitively. But complaining doesn't really do much without a proof. It's easier to prove a hack (in the form of illegal entry) or spam mail than proving a troll, because a troll is defined by those being trolled (us) whereas a hack or a spam is found in different log files. Still, if nobody complains, no one will give the guy a healthy reminder. So, why not? The worst thing that could happen is that they don't care. Gustav On Sun, 6 Jun 1999 11:29:12 -0400, Pierre Abbat wrote: >Is it okay to complain about a troller to his ISP? > >phma > In the immortal words of the captain of the Titantic, 'Where did all this fucking ice come from?' Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [[Re: Mersenne: Serious problems with v.18]]
Stephen Whitis <[EMAIL PROTECTED]> wrote: > He claimed to be a programmer, but then over and over he claimed that > bugs should *always* be caught in testing, prior to release. That > proves he's lying, because no programmer can believe that happens in > the real world. A complete set of test-cases should catch all major (e.g. showstopper) bugs. If a bug of such magnitude makes it into a release product testing cannot have been thorough. And someone admitted that testing for v.17 was in fact not thorough, as the routines that kick in above 2^22 never were tested since the test suite didn't include an exponent above 2^22. > Maybe we'll get lucky and his ISP will screw him again... :^) You should grow like a turnip, with your head in the ground and your legs in the air. Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
I second that. At 16:33 06/06/1999 +0100, you wrote: >NO MORE. The end.. > > >Chris Jefferson, Girton College, Cambridge, [EMAIL PROTECTED] > >I have a proof that x^n+y^n=z^n never has integer solutions for n>2. >However, it won't fit into my signature file > Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
NO MORE. The end.. Chris Jefferson, Girton College, Cambridge, [EMAIL PROTECTED] I have a proof that x^n+y^n=z^n never has integer solutions for n>2. However, it won't fit into my signature file Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [RE: [Re: [Re: Mersenne: Serious problems with v.18]]]
On Sat, Jun 05, 1999 at 11:07:04PM -0600, Paul Derbyshire <[EMAIL PROTECTED]> wrote: > > Erm? Regardless of when I found out, it remains a fact that to release > non-beta software to the general public without a thorough testing *is* a rash > act. The software *was* tested - the testing missed this particular problem. The testing has been improved since then - if you'd like to help it improve even more you're been invited to contribute. George has apologized, at length. Another prime has (most likely, double-check pending) been found. The world continues to turn. > >From my point of view, it isn't "long after the fact". It was yesterday. Yay you. Can we let this topic drop now? Bryan -- Bryan Fullertonhttp://www.samurai.com/ Core Competency Samurai Consulting "No, we don't do seppuku." Can you feel the Ohmu call? Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [RE: [Re: [Re: Mersenne: Serious problems with v.18]]]
My motivation is to help. This spectacular bug caused some of my help to have been wasted. Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [RE: [Re: [Re: Mersenne: Serious problems with v.18]]]
> Here are my ideas on bugs: > Bugs happen! They're a fact of life, omnipresent in all software. Showstopper bugs should not slip through testing and into release software. > Bugs should allways be caught in the testing, but so often they aren't. Minor bugs yes, massive showstoppers no. > In the meantime, a subgroup of testers have been created which should > (hopefully) ensure that things like this cannot take place again. Yes... a good idea. With a shiny new lock on that barn door, perhaps these horses won't escape a second time and cost GIMPS hundreds of P90-years of time... I still think it would have been a good idea to have had this lock from the outset. But it's water under the bridge... Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
RE: [Re: [Re: Mersenne: Serious problems with v.18]]
> Paul Derbyshire wrote: > > Sounds like a lot to me. This CPU I have can finish LL tests in > that kind of > > time, except that I have a habit of using it for other things a > lot of the > > time. > > It may seem like a long time to you, but I've been interested in Mersenne > Primes as an observer since around 1978-9, and as a participant > in GIMPS before primenet, > back when exponents were more like 700,000 rather than 7,000,000, > which has been > what, 3, 4 years? > > Moreover, George and Scott have always been very clear and > forthright about > what's going on in GIMPS/Primenet. I've never met George and Scott in > person, but I've dealt with them in regard to GIMPS/Primenet, and I want > to take this opportunity to say publically that they are a primary reason > that almost all my idle cycles remain dedicated to the GIMPS. I also, as someone whose been in GIMPS since nearly the beginning, don't consider a few weeks of lost time to be very much. I run NTPrime/Prime95 on anywhere from 30-50 processors at any given time. The version 17.x bug was annoying, and I had a few machines that were just about finished. But it's not a big deal. I just started over. Cumulatively, I probably wasted nearly a couple P-90 CPU years, but so what? I do this for fun...before there was any prize, whether I'm able to run on just 1 machine or several thousand :-) or whether I'm doing just factoring work, double-checks, or LL tests (or a healthy mix of all 3). If losing a few weeks work is so upsetting, perhaps you should examine your motivation. Just my $0.02 worth... Aaron Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [RE: [Re: [Re: Mersenne: Serious problems with v.18]]]
"Rick Pali" <[EMAIL PROTECTED]> wrote: > I don't think that anyone's trying to flame anyone at all, but I can't [the rest is missing] ??? > ... might it not be a good idea to ask to be apprised of the situation > before accusing the very people who make all this possible of > being "completely unconscionably rash." Erm? Regardless of when I found out, it remains a fact that to release non-beta software to the general public without a thorough testing *is* a rash act. > And even when a single problem has occurred, there was a fix within a > day or two. Yeah. But with spectacular showstopper bugs, that is still locking the barn door after the horses have escaped. Spectacular showstopper bugs shouldn't ever make it into a release product... testing (alpha or beta) should be able to catch all of those. > I don't think that he deserves to be called to the carpet for it so > long after the fact. >From my point of view, it isn't "long after the fact". It was yesterday. Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
RE: [Re: [Re: Mersenne: Serious problems with v.18]]
All, alright, everybody break it up! >> Well, you've dropped the annoying Mandelbrot quote, but you're still >> trying to stir up trouble on my favorite mailing list. > > I'm not trying to stir up trouble. However, between your insulting of my sig > file (absent temporarily, due to my ISP's mail trouble and my consequently > being stuck with usa.net's crummy web interface) and your accusing me of > flame-baiting, I begin to suspect that you, who I formerly had not even > noticed, are in fact trying to flame-bait the list. I doubt that anyone was trying to flame-bait, [EMAIL PROTECTED] was mearly over-reacting to Paul Derbyshire's over-reacting to a software bug. Hey, I understand on both counts. Most of us deeply respect George as a brilliant organizer, and (at the very least) a very competent programmer, as well as very good leader of a *fun* and *volunteer* project. Here are my ideas on bugs: Bugs happen! They're a fact of life, omnipresent in all software. Bugs should allways be caught in the testing, but so often they aren't. In the meantime, a subgroup of testers have been created which should (hopefully) ensure that things like this cannot take place again. >> a) This incident was well discussed when the bug was detected >> about 9 weeks ago. > > I wasn't there then. Because of my F---ING ISP and their worthless mail > system... Ok, hopefully, you can see how your remarks seemed especially inflamatory, as well as rude and stupid to someone without this knowledge. However, now that we all have this knowledge about your email, your remarks seem more reasonable, just an over-reaction. and remember, A debugged program is one for which you have not yet found the conditions that make it fail. -- Jerry Ogdin -Lucas Wiman Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: [Re: Mersenne: Serious problems with v.18]]
Paul Derbyshire wrote: > Sounds like a lot to me. This CPU I have can finish LL tests in that kind of > time, except that I have a habit of using it for other things a lot of the > time. It may seem like a long time to you, but I've been interested in Mersenne Primes as an observer since around 1978-9, and as a participant in GIMPS before primenet, back when exponents were more like 700,000 rather than 7,000,000, which has been what, 3, 4 years? Moreover, George and Scott have always been very clear and forthright about what's going on in GIMPS/Primenet. I've never met George and Scott in person, but I've dealt with them in regard to GIMPS/Primenet, and I want to take this opportunity to say publically that they are a primary reason that almost all my idle cycles remain dedicated to the GIMPS. --Joe Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
RE: [Re: [Re: Mersenne: Serious problems with v.18]]
From: Paul Derbyshire > and your accusing me of flame-baiting, I begin > to suspect that you, who I formerly had not even > noticed, are in fact trying to flame-bait the list. I don't think that anyone's trying to flame anyone at all, but I can't > > Evidently you haven't been monitoring the list. > > No thanks to Globalserve... Well that's not the list's fault, not is it George's or Scott's so might it not be a good idea to ask to be apprised of the situation before accusing the very people who make all this possible of being "completely unconscionably rash." Not only would've it prevented that egg on your face, but it's just plain courteous. > A problem of such magnitude should have been caught a lot > sooner... but I repeat myself. Yea, you sure do. I'd consider the software that the project has produced as *outstanding* when compared to the commercial norm. And even when a single problem has occurred, there was a fix within a day or two. It's a pipe-dream with software that we pay hundreds of dollars for...and we can download this stuff for free. After two years, I've only lost two numbers (that weren't my fault). With a record like that, the reasonable among us have nothing at all to complain about. And if 100% accuracy is what you demand, Ken has extended an invitation to you to join the testing team. > > George made COPIOUS apologies REPEATEDLY when the bug was found. > > Well, no thanks to my ISP, I was not aware of this. We're got nothing at all to do with that, and you can't claim that your ISP has anything to do with what you've written since your return either. Like the previous poster said, George obviously regretted the error, felt bad about it, and apologised repeatedly. I don't think that he deserves to be called to the carpet for it so long after the fact. Rick. - [EMAIL PROTECTED] http://www.alienshore.com/ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: [Re: Mersenne: Serious problems with v.18]]
[EMAIL PROTECTED] wrote: > Well, you've dropped the annoying Mandelbrot quote, but you're still > trying to stir up trouble on my favorite mailing list. I'm not trying to stir up trouble. However, between your insulting of my sig file (absent temporarily, due to my ISP's mail trouble and my consequently being stuck with usa.net's crummy web interface) and your accusing me of flame-baiting, I begin to suspect that you, who I formerly had not even noticed, are in fact trying to flame-bait the list. I think that is a bad idea. This list is a good thing and causing a flamewar on it will be a bad thing. In the interests of not allowing a flamewar to start, I shall henceforth not respond to anything you post that appears to have inflammatory intent. I recommend that everyone else do likewise. Then there won't be any flamewars. Of course, if I have misunderstood your intentions, you presumably won't post anything else inflammatory, and the foregoing will be rendered moot. > Evidently you haven't been monitoring the list. No thanks to Globalserve... > Errors mounted with each and every iteration, making "correction" > impossible and worthless anyway. A problem of such magnitude should have been caught a lot sooner... but I repeat myself. > George made COPIOUS apologies REPEATEDLY when the bug was found. Well, no thanks to my ISP, I was not aware of this. [Various inflammatory stuff deleted.] Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: [Re: Mersenne: Serious problems with v.18]]
[EMAIL PROTECTED] wrote: On 5 Jun 99, at 8:23, Paul Derbyshire wrote: > If you check the source code... I don't have it. > And, because of a "feature" dating back to days when exponents were > small, none of the test data sets in the program self-test would > trigger the bug. So "normal" testing procedures failed to detect it. The self tests go to arbitrary levels of precision! I've seen it. Before beginning an exponent it does a self-test for that precision, if it hasn't already done one there. > a) This incident was well discussed when the bug was detected > about 9 weeks ago. I wasn't there then. Because of my FUCKING ISP and their worthless mail system... > b) I can assure you that George did not feel at all good about it at > the time. Nevertheless, accidents happen. Always have done, > always will do. That's life! Accidents will happen. This is why they invented crash test dummies... I think I'll be staying away from any ".0" versions in future. > (Though, as I say, if you saw the offending code section, I > think you'd need to think hard about it for a few minutes before you > could see that there was a potential problem). I'd bet you're right. Because, although I am a programmer myself, I don't know the code like the back of my hand. I would hope each person that does work on the code does know it like the back of his/her hand... > d) There is a maxim amongst conservative computer users, "don't > fix what ain't broken". There was no _need_ to upgrade to v17, > unless you wished to run double-checking assignments. I was running double checking assignments until I got the new K6-2 400MHz CPU. > ...the setback was actually only four or five weeks. Sounds like a lot to me. This CPU I have can finish LL tests in that kind of time, except that I have a habit of using it for other things a lot of the time. Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
Well, you've dropped the annoying Mandelbrot quote, but you're still trying to stir up trouble on my favorite mailing list. In a message dated 6/5/99 10:27:51 AM Eastern Daylight Time, [EMAIL PROTECTED] writes: > Are you sure of that? What if the bug didn't happen to strike my run, or the > errors could be corrected? Evidently you haven't been monitoring the list. The version 18 software was DESIGNED, immediately after the bug was found, to determine whether any given run contained corrupt data. Errors mounted with each and every iteration, making "correction" impossible and worthless anyway. > > If what you say is true, then whoever designed version 17 acted in a > completely unconscionably rash manner by releasing it without thoroughly > testing it for problems as serious as that. And has therefore shot the whole > GIMPS effort in the foot by setting it back many weeks. More inflammatory statements showing your non-monitoring of the mailing list. George made COPIOUS apologies REPEATEDLY when the bug was found. If you think his behavior is unconsionable, you are welcome to withdraw from the project. Remember that 100% of this effort is volunteer. > > Hopefully, realization of the impact on GIMPS is punishment enough to make > sure this won't happen again... > Hopefully, you'll get a clue and stop being so insulting. I don't hold out a whole lot of hope though. Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
At 08:23 AM 1999/06/05 MDT, Paul Derbyshire <[EMAIL PROTECTED]> wrote: >Peter Doherty <[EMAIL PROTECTED]> wrote: > >> This is normal. Because of the bug in v17, all the math it was doing >> was wrong, so using that 77% would have been a waste since it was >> incorrect data. There is no need to try and retrieve that data. It's >> useless. > >Are you sure of that? What if the bug didn't happen to strike my run, or the >errors could be corrected? > >If what you say is true, then whoever designed version 17 acted in a >completely unconscionably rash manner by releasing it without thoroughly >testing it for problems as serious as that. And has therefore shot the whole >GIMPS effort in the foot by setting it back many weeks. > As I recall, there were times when new versions of primenet or prime95 were needed by a certain date to prevent exhaustion of all exponents below the usable limit. Remember that all code development is by volunteers, primarily George Woltman for prime95, and Scott Kurowski for primenet. I do not have you on my list of QA testers or code reviewers. Would you like to participate in helping V19 meet the code quality you desire in this free software? Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
At 08:23 6/5/99 MDT, Paul Derbyshire wrote: >Peter Doherty <[EMAIL PROTECTED]> wrote: > >> This is normal. Because of the bug in v17, all the math it was doing >> was wrong, so using that 77% would have been a waste since it was >> incorrect data. There is no need to try and retrieve that data. It's >> useless. > >Are you sure of that? What if the bug didn't happen to strike my run, or the >errors could be corrected? This is a known bug the consequences of which were hashed and rehashed at least one month ago. V18.1 was designed to take appropriate action with "work in progress." Your work to date on that exponent was disgarded (though you might still get credit for time spent--Scott?) because it was invalid. >If what you say is true, then whoever designed version 17 acted in a >completely unconscionably rash manner by releasing it without thoroughly >testing it for problems as serious as that. And has therefore shot the whole >GIMPS effort in the foot by setting it back many weeks. Umm, this is a volunteer effort on the part of George, Scott, and others. Programming of this type is difficult, and bugs are a part of life. Let's save phrases such as "completely unconscionably" for more important things in life. GIMPS, while real science, is supposed to be fun. >Hopefully, realization of the impact on GIMPS is punishment enough to make >sure this won't happen again... Is the phrase "Lighten up" too strong? Kel <--hoping that one of the 6,000,000s I'm checking is a Mersenne prime as per the island theory :-) Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Re: [Re: Mersenne: Serious problems with v.18]
Peter Doherty <[EMAIL PROTECTED]> wrote: > This is normal. Because of the bug in v17, all the math it was doing > was wrong, so using that 77% would have been a waste since it was > incorrect data. There is no need to try and retrieve that data. It's > useless. Are you sure of that? What if the bug didn't happen to strike my run, or the errors could be corrected? If what you say is true, then whoever designed version 17 acted in a completely unconscionably rash manner by releasing it without thoroughly testing it for problems as serious as that. And has therefore shot the whole GIMPS effort in the foot by setting it back many weeks. Hopefully, realization of the impact on GIMPS is punishment enough to make sure this won't happen again... Get free e-mail and a permanent address at http://www.netaddress.com/?N=1 Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
