JSH: Posting realities, residues result
in [Math]
|
From: MichaelW on 12 Jun 2010 22:18 On Jun 13, 8:54 am, Joshua Cranmer <Pidgeo...(a)verizon.invalid> wrote: > > What I have seen more in the most recent cycles is people who just take > the opportunity to insult JSH and do nothing but that. However much I > like or dislike JSH, and however much he may or may not warrant it, such > actions are just unacceptable in my view. Everyone would be better off > if they just shut up: those posts just waste bandwidth and adds fuel to > the file. I blame MMM. This group was increasingly being taken over by the bandwidth wasters but with the arrival of MMM almost everyone else has left. It would be nice if there were a moderated forum somewhere to go to. The problem I suspect is that mathematical knowledge and skill is on such a broad spectrum that our particular level (interested amateur, capable but realistically not trained enough to operate at the academic level) does not have a home. The best I have found is Project Euler. I have merrily wasted many hours of my time and computer time at that site. Regards, Michael W.
From: Rotwang on 12 Jun 2010 22:46 MichaelW wrote: > > [...] > > I would note that step 7 in the cycle is especially creeping me out. > He recently claimed to have never agreed that his TSP algorithm is > wrong (despite posts clearly stating just that) I think you've missed something here - the version of the TSP algorithm which he agreed was wrong after I posted code implementing it and an example graph for which it fails is not the same as his current version. If my implementations are to be believed then the current version is actually worse than the old version, but AFAIK he never actually acknowledged that the new version doesn't work (despite my pointing out that it didn't at the same time as I pointed out that the original didn't, since I had anticipated he would come up with the variation in question).
From: MichaelW on 12 Jun 2010 22:53 On Jun 13, 12:46 pm, Rotwang <sg...(a)hotmail.co.uk> wrote: > MichaelW wrote: > > > [...] > > > I would note that step 7 in the cycle is especially creeping me out. > > He recently claimed to have never agreed that his TSP algorithm is > > wrong (despite posts clearly stating just that) > > I think you've missed something here - the version of the TSP algorithm > which he agreed was wrong after I posted code implementing it and an > example graph for which it fails is not the same as his current version. > If my implementations are to be believed then the current version is > actually worse than the old version, but AFAIK he never actually > acknowledged that the new version doesn't work (despite my pointing out > that it didn't at the same time as I pointed out that the original > didn't, since I had anticipated he would come up with the variation in > question). Ah, I stand corrected. We are in the "keep iterating through variations of the same failed idea until enough people loose interest and then claim victory" space. Still creepy. Regards, Michael W.
From: Joshua Cranmer on 12 Jun 2010 23:26 On 06/12/2010 10:46 PM, Rotwang wrote: > I think you've missed something here - the version of the TSP algorithm > which he agreed was wrong after I posted code implementing it and an > example graph for which it fails is not the same as his current version. I base my knowledge of what I see off of Usenet. He makes little distinction between different iterations of his same work, so you can lose track of which algorithm he's referring to and which one you've referred to. I constantly qualified my example of disproving TSP with "all of the algorithms I've seen" or similar wording, and JSH only rarely indicated the existence of work outside of Usenet. The last time, it took a few weeks before he admitted in Usenet that he had another algorithm elsewhere which is what he had been referring to. I still don't know the difference between that algorithm and another one I had also disproved. > If my implementations are to be believed then the current version is > actually worse than the old version, but AFAIK he never actually > acknowledged that the new version doesn't work (despite my pointing out > that it didn't at the same time as I pointed out that the original > didn't, since I had anticipated he would come up with the variation in > question). The problem with his original version of the TSP algorithm (at least, the original that I recall seeing) was noted to be avoided only if the algorithm hit two points at the correct time. His solution was to modify the algorithm to try starting at every point in order, which fixed this particular instance by causing one time to hit it at the right time. It took a surprising amount of work to convince him that this wouldn't actually fix it, as the feature in question could be pretty trivially duplicated and a new graph created that defeated it. This was when I realized that specifying an example input with words like "all other weights have edges sufficiently high to be not considered by the program" was insufficient to convince him of incorrectness. The TSP algorithm work is the case study I'm most familiar with; it's the primary basis for my conclusion that he does not attempt to stress results too hard, as many of the errors found would have been easily spotted had he worked hard to figure out where the algorithm was most likely to fail: I think Patricia found the failing case within 6 hours or so of him publishing his algorithm. As I think about more, I'm starting to develop yet another alternative explanation for this behavior. I've noted before that JSH seems to require rather explicit examples to sway his mind about his results, and I've seen recently that he is adamantly refusing to reason about his results without using examples. I theorized in the past that this was an unwillingness to be proven wrong or the lack of an imaginative mind; perhaps it is an inability to reason (or maybe metareason fits better here) abstractly. This may explain the elaborate hypothetical situations: in lieu of presenting proper mathematical analyses, asserting that the results are important and then explaining why that means we all suck deflects attention from the lack of rigorous mathematical analysis. A fair amount of his incessant recitation of Google results could also be an attempt to back his assertions, lacking the analysis truly desired, with yet another example and covering up the aforementioned inability. Or this could all be a product of my presumably incredibly deranged imagination. I think I really ought to sleep now... -- Beware of bugs in the above code; I have only proved it correct, not tried it. -- Donald E. Knuth
From: Jesse F. Hughes on 13 Jun 2010 16:42
JSH <jstevh(a)gmail.com> writes: > My coverage by myself on a yearly basis is 120+ countries to JUST my > math blog alone. Do you have any idea how many of those visitors are human beings and how many are spiders, either indexing the site for legitimate reasons or scanning the site for email addresses for some spammer somewhere? -- "Yup, you guessed it. If worse comes to worse, I *will* turn to the Army to help me with mathematicians. And then mathematicians don't think the NSA or CIA can save your asses, as generals LIKE me." -- James Harris's latest foray into mathematical logic. |