This site is supported by donations to The OEIS Foundation.

User talk:James G. Merickel

From OeisWiki
Jump to: navigation, search

If I can edit now ...

... I won't be here much till at least 2018. If I can't there's only so much this site has for me, but I'll use whatever is useful. Right now it appears I can't add terms or make changes/repairs to sequences already done, and I have not tried adding new sequences.James G. Merickel 19:02, 22 September 2016 (UTC)

Notices by Editors to me

I'm just putting this here in case anything not really to be responded to directly is deemed to be something I should or might want to know. Recently the offset of one of my older sequences was changed from 2 to 1 (with the addition of a trivial first term) by Neil Sloane, and I have at least one related sequence I am changing likewise myself as a result of my happening to discover this. If I had been told, this chance occurrence would not have been necessary. Simply edit such under the divider from this remark.James G. Merickel 04:30, 26 July 2015 (UTC)



Sequences Requiring Small Changes

These all contain small errors (or editorial weaknesses, such as weak commenting on search limits in the original -- see A211405) that an editor could easily fix. They are last on the queue of things I plan to do while my freedom to edit is restricted by number of sequences and pace of approval. A246718, A230054, A211405, A217409, A216873 (dated), A209296, and A260075 (very beginning of program). There are probably others, but most planned future editing involves extensive fixes or improvements, new terms for my own and other people's sequences, and more new sequences.James G. Merickel 06:50, 3 August 2015 (UTC) A211404(7)=46115 (Already should be credited elsewhere, so this is usable as basis for somebody else submitting just to save time, no need to credit).James G. Merickel 00:12, 10 August 2015 (UTC) A103787(32)=290220889. Feel free to add this one for me for joint credit. (I assume one cannot submit for me without being registered as one of those editing, and unless computed oneself putting my name to it protects you in the event I'm wrong.)James G. Merickel 04:42, 11 August 2015 (UTC)

Probable Primes versus Primes

Unless explicitly spelled out how probable a probable prime is according to heuristics, in my sequences it should be assumed that 'prime' and 'probable prime' both mean that inaccuracy would need to be due to an astronomically improbable false positive to a set of tests comparable--and usually identical, to date--to PARI/GP's ispseudoprime function, unmodified.

As well, in some cases I may have terms that are only extremely probable of being correct, and in this case--not questions of primality, but of some other kind--this should be explicit in the COMMENTS section and possibly also the TITLE of a sequence.James G. Merickel 13:03, 9 October 2013 (UTC)


One example arose actually regarding primality but of another kind. Searching 8338 and my name will yield specifics, where I signed off on this value as a term without doing certain primality tests at all, where cases were so reduced that multiple primes had to coincide, ones of very large value. This tentative term has at least not been fully checked in a way that totally satisfies me, but I am still confident of its correctness. One sequence of two that are related is more theoretically affected than the other. Computer crashes and premature approval left me wanting to complete the project but unable to for a while, and then no corrected version despite the open warning of possibly erroneous approval led me to conclude it is more likely sound than before without doing anything or being specifically aware of computations concerning it by others (just because of some chance these checks were done more fully). This is not an ideal situation, and it generally is not a good way for mathematics to function or appear to.James G. Merickel 00:24, 2 August 2015 (UTC) To be a little clearer, where 8338 = a(5) I am relatively satisfied. As I recall, I was down to some hundreds of values where only 1 or 0 primes exist under a certain size, with very very small chance any of these have 4 or 5 more primes being the term basis. But the chance there was a bit larger of course that a total of 4 primes exist (so that the other sequence might have missing terms, with some probability still small but not where science would generally accept the terms in order).James G. Merickel 04:06, 2 August 2015 (UTC)

Excess data for A230360

Don't know if I can create the needed b-file without my own webspace somewhere, which I don't have; so here are terms beyond those being submitted in data, beginning with a(66): 132727, 301356, 0, 120794, 545340, 215112, 0, 0, 1506887, 0, 0, 936531, 744207, 0, 0, 1571930, 4160428, 0, 0, 2987556, 1263881, 0, 0, 9965736, 6744787, 0, 0, 16901659, 39697315, 0, 17793339, 99276869, 0, 0, 177599974, 0, 0, 0, 30123558, 128499559, 0.

ANYONE is welcome to expand the b-file with this. I will eventually but it's low on a queue right now.James G. Merickel 18:50, 30 July 2015 (UTC)

I've decided this will be relatively soon because of somewhat poor commenting at the sequence that also needs a little work. And running the program that generated these terms has shown me there is no point to it -- I had already let it run a pretty long time to get the terms I did. James G. Merickel 09:14, 2 August 2015 (UTC) And the program needs to be moved.James G. Merickel 09:18, 2 August 2015 (UTC)

(About trios of same-omega squarefrees)

A093550(9)=22192526378762466. The means by which this was discovered and ascertained involved a little over a week computing on my i5 laptop using a search over increasing largest prime factor of a trio member not divisible by 2 or 3. I expect to link a program at that sequence that shouldn't differ much from what was used. The value was only the 2nd of 2 improvements over the value given several years ago at the primepuzzles site, and it came out very early. So the programs scarcely indicated they were doing anything. It was mostly split over each of the 5 congruence classes modulo 5 for the index of the largest prime of the hypothetical trio's member of largest smallest prime. It completely covered the possibilities in that time! I expect that A093550(10) will also be proved. I have no improvement over the problem's given value yet. No value at all has so far come out under 10^22 for the 11 case through largest prime 199. I guess I'll complete that case too, but not soon.James G. Merickel 06:10, 23 July 2015 (UTC) This is as slow as molasses. The search on 10 has yet to pass 500 for largest prime factor of one not divisible by 2 or 3, with no improvement still, using an entire i5's power; and one window on the 11 case is only past 211 for largest prime. For the latter, I still don't even know if 10^22 is conservative enough.James G. Merickel 13:29, 25 July 2015 (UTC) 223 passed in 11 case.James G. Merickel 19:53, 25 July 2015 (UTC) 227, and no first case for 11. Past 500 in all 5 (by class mod 5 of prime indices), but have to type them in and continue runs again because of power outage. Also no improvement over the one published.James G. Merickel 18:50, 28 July 2015 (UTC) Actually, that was just stupidity. Turned my air conditioner off when I was sitting next to it naked. Next thing I know I'm waking up in a sweat and the computer I know is touchy has crashed. Took forever to re-type 5 times right using onscreen keyboard. So, lost around 3 days work kind of, since the windows were working at different paces and I didn't know where to start them to not repeat stuff and not skip stuff either.James G. Merickel 14:53, 29 July 2015 (UTC) 229 passed in 11 case, and what I was saying for 10 is I set them all to just past 500 even though one or two were past 600 when I made my dumb a.c. decision.James G. Merickel 02:22, 30 July 2015 (UTC) 233 passed.James G. Merickel 04:03, 31 July 2015 (UTC) Past 241, but I won't be updating all that frequently. No news in either case, and the other is moving along slowly.James G. Merickel 00:36, 2 August 2015 (UTC) 263 gone and still nothing under 10^22. I expect it's still a conservative enough bound to have one come out. As for the 10 case, most of the 5 windows have passed 700. No smaller value than that computed by Fred Schneider years ago yet.James G. Merickel 09:52, 6 August 2015 (UTC)

On size of smallest n with prime reversals of n^k, k=1 to 8

Candidate values at a particular order of magnitude lie in two connected locations: around the power of 10 with lower bound the power of 10 times (9*10^7)^(1/8) and upper bound the power of 10 times 2^(1/8), and around the square root of 10 times the power of 10 with lower bound 3000^(1/7) times the power of 10 and upper bound 4000^(1/7) times the power of 10. I have it known that the smallest number is at least in the middle 12-digit range, but the following calculations assume only it's known to be at least 12 digits. Very nearly 5/11 of all numbers are eliminated on grounds of being divisible by 3 or 11 or ending in 0. All possible final 2 digits are broken down to 90 possibilities and the effects on the power's reversals are arrayed with an additional 1/2 added and an array of offsets to the integer parts of the logarithm base 10 is created. The 2 connected segments are both split in 2 to have the lengths of the powers all the same over a range. The number with specific final pair of digits that meet the requirements vis a vis 2, 3, 5 and 11 is for each about 1/165 times the whole number of numbers in a given range. The problem itself is for NON-existence of simultaneous primes to have repeatedly occurred for all numbers up to some point. With 2, 3, 5, and 11 not factors of any of the powers reversed, the PNT-based heuristic factor has an extra part 33/8. As a result of this described method (again, without the info that smaller 12-digit numbers have been ruled out), the heuristic calculations are for 85% chance the number is at least 13 digits, 38% that it's at least 14, and less than 0.5% for at least 15 (with virtually no chance it is not to be found then with leading digit 1). The most probable is a 14-digit number with leading 1, at 28.6% under conditions stated, which would go up to 33.5% if 12-digit numbers are removed (and 22.9% for 13-digit with leading 1, 19.9% for leading 3, 12.5% for leading 9, and 10.5% for larger 14-digit number).James G. Merickel 15:25, 15 July 2015 (UTC) Being sought is A165696(8). The search has now progressed to the transitional range from 12 to 13 digits. If a fail occurs here the maximum likelihood would still be for 14 digits with initial 1, but the chances for larger will have increased a bit.James G. Merickel 04:08, 20 July 2015 (UTC) No movement here recently.James G. Merickel 10:34, 25 July 2015 (UTC) Still not on to middle 13-digit range yet. It will be months if it goes beyond 14 with leading 1 (unless a dedicated search using multiple windows is employed).James G. Merickel 21:52, 29 July 2015 (UTC) I expect the signaling asterisk to appear soon, but nothing so far.James G. Merickel 04:11, 2 August 2015 (UTC) I don't quite get this. I suppose a combination of increasing the load on the involved computer and not knowing the relative size of the ranges (in my head). Some day or perhaps hour this will make it to the next range.James G. Merickel 09:48, 6 August 2015 (UTC) ispseudoprime with an additional order of magnitude may also be a significant cause.James G. Merickel 11:16, 6 August 2015 (UTC) Quite annoyed at how long this is taking just to go another step!James G. Merickel 07:36, 8 August 2015 (UTC)

Off until Friday, August 14

OEIS vacation.James G. Merickel 01:05, 9 August 2015 (UTC)

If I look and find things approved, I am going first to my list of minor edits required. Lengthier fixes and new stuff are set to continue on that date, but if I have space for fixing a little typo or grammatical thing needed, it will be edited and re-submitted in a jiffy.James G. Merickel 01:13, 9 August 2015 (UTC)

GP script at A260075

I just ran your script at A260075 and noticed that it gave only a string of 5s. Is there a bug in the program?

Charles R Greathouse IV 20:38, 25 September 2015 (UTC)

Test. Just seeing I can edit here. Was trying to edit a sequence now, but couldn't. Adding terms to A171740 and changing its comments and removing a claim I could prove something that I'm sure was a mental mirage (another sequence I don't have the A-number of at present). were the first things I was going to do back. As far as the named sequence, that data is posted at h2g2.com in 'useless facts' 3 pages from the last edited. [Note: I was only able to reply to this originally by email, having been banned for a year for 'wasting editors' time'.]