Talk:On-Line Encyclopedia of Integer Sequences

OEIS linking syntax help

I would like to make a page along the lines of "Wikipedia:How to cite the OEIS" and include a link to that from this article. That page would explain the five different ways of citing and linking sequences in the OEIS and illustrate how to accomplish each one. PrimeFan 22:46, 14 July 2005 (UTC)

I've just spent 20 minutes searching for a template, convinced I'd seen one before. I'd welcome a synopsis. Hv 01:07, 17 August 2005 (UTC)

Using Sloane's A012345 as an example, the table below shows the five different ways. PrimeFan 19:32, 18 August 2005 (UTC)

[http://www.research.att.com/projects/OEIS?Anum=A012345]	[1]
[http://www.research.att.com/projects/OEIS?Anum=A012345 A012345]	A012345
[[OEIS:A012345]]	OEIS:A012345
{{OEIS|id=A012345}}	(sequence A012345 in the OEIS)
{{OEIS2C|id=A012345}}	OEIS: A012345	This format is used for second and subsequent citations in a given article.
{{SloanesRef|sequencenumber=A012345}}	Sloane, N. J. A. (ed.). "Sequence A012345". The On-Line Encyclopedia of Integer Sequences. OEIS Foundation.	Seen in Arnold's_cat_map

I don't think the first two are really worth mentioning, and the third is a maybe - I'd class the templates as generally preferable, both for brevity and for the meta-information provided by having a reference to the template.

My own search turned up nothing via Wikipedia:Template, and I eventually stumbled across the templates in another page.

I certainly wouldn't have guessed what "OEIS2C" stood for, so some explanation somewhere would surely help. I did look in Wikipedia:Template messages/Links, so a reference there would be a start. Hv 02:03, 19 August 2005 (UTC)

I'm not even sure how things are sorted there at Template messages/Links. I'd have to carefully examine that page before I dared add anything to it. PrimeFan 20:39, 22 August 2005 (UTC)

How about mentioning these templates in the OEIS article? HenningThielemann (talk) 16:00, 9 January 2011 (UTC)

They are Wikipedia-specific, the article should be about OEIS itself and not list templates related to OEIS that can be used on Wikipedia. If someone created a page describing their use in the Wikipedia namespace, that would be okay. See also Wikipedia:SELFREF#Community_and_website_feature_references. Svick (talk) 19:14, 9 January 2011 (UTC)

Make encyclopedic

The current text has a mix of encyclopedic information and some very low-level "how-to" material. It needs to be cleaned up.... --Macrakis 23:31, 26 September 2005 (UTC)

Oh, how easy it is to complain about a problem and not do a thing to fix it! "It needs to be cleaned up," easy to say. "Make encyclopedic," easy it is to give that command.

At any rate, there isn't a problem. What you call "low-level how to" is a much better way to explain a potentially cryptic mathematical topic. A professional mathematician is given to saying things like "It is obvious that $\int _{-N}^{N}{\sqrt {k\pi \lambda }}\,dx$ hence $k=x^{2\mu }$." The lexicographic order of the OEIS is here explained far more clearly than in the OEIS itself! Robert Happelberg 22:06, 27 September 2005 (UTC)

OK, I've started fixing it. --Macrakis 01:51, 28 September 2005 (UTC)

Good job so far, Macrakis. Bob makes a good point, too. With Macrakis's and Bob's help, this article will eventually strike a good balance between spontaneity and professionalism. PrimeFan 15:48, 28 September 2005 (UTC)

Errors in the database

The article says "For a database of its magnitude, the OEIS is relatively free of errors. ... The more common mistakes in the OEIS occur in fields other than the sequence or signed field."

I have reason to doubt this. In early January 2000 or 2001, I took the database and ran a program on it to look for sequences whose terms were all prime except one. There were something like 100+ of these, perhaps as many as 300 (I can't remember for sure). I started checking these to see if the non-prime term was in error. I think I checked a little over 100 sequences. I found 10 or 11 errors. The errors were about what you would expect:

1. transpose two digits
2. spurious digit or digit omitted
3. repeating the wrong digit, i.e. 355 instead of 335.
4. leaving a comma in a term with four or more digits, making it appear to be two separate terms
5. omitting a comma between two terms

So around 10% of the sequences I checked had obvious errors, and there could have been others. Bubba73 (talk), 23:26, 15 January 2006 (UTC)

Then I'm in trouble. For the vast majority of Wikipedia articles I've written about various kinds of numbers (e.g., Carol numbers, prime quadruplets), I've taken the listing from the OEIS. In a few cases I take the listing from Mathematica.

For my part, I will double-check the sequences listed in articles I've written. For your part, could you re-run your "sequences whose terms were all prime except one" search on a more recent copy of the table? PrimeFan 21:05, 17 January 2006 (UTC)

Bubba73 did an interesting approach for finding errors in the database! Another way might be to check the sequence samples against the given PARI, Mathematica, and Maple code pieces. HenningThielemann (talk) 16:26, 9 January 2011 (UTC)

Self-referential paradox

In the 'Self-referentiality' section, "sequences A053873, n is in An, and A053169, n is not in An" are referred to as "delightfully paradoxical" because "[t]he paradox is, which sequences do 53169 and 53873 belong to?" Surely only 53169 in A053169 gives rise to a paradox?

Sendhil 01:24, 20 July 2006 (UTC)

Alright, prove that 53873 does belong in OEIS: A053873. Anton Mravcek 16:41, 20 July 2006 (UTC)

Or prove that it does not.

And 53169 not in OEIS: A053169 also gives rise to a paradox --Rumping 19:12, 22 September 2007 (UTC)

Looking at this a year later, I stand by calling these issues "paradoxical." My problem now would be the word "delightfully." I find these paradoxes delightful but that might just be my POV. Some people might find these annoying. Anton Mravcek 20:32, 22 September 2007 (UTC)

I don't see anything paradoxical regarding 53873. It's true that we can't prove whether 53873 is in the sequence or not, but that just means that the sequence is not completely defined. On the other hand, I believe that A053169 is truly a paradox: both the assumptions that 53169 belongs to the sequence and that it doesn't lead to contradictions. -- Jitse Niesen (talk) 14:15, 29 December 2008 (UTC)

I have re-written the section to try to improve its clarity. I removed the "delightfully paradoxical" phrase because it seemed to me to be an unencyclopedic description. I agree with Jitse Niesen - A053873 cannot be completed unless we create an axiom that says "53873 is (or is not) a member of A053873", but only A053169 is truly paradoxical. Gandalf61 (talk) 16:22, 26 January 2009 (UTC)

In the unreferenced quote attributed to Sloane in the self-referentiality section of the article, the article source contains an HTML nonbreaking space before the exclamation mark in "known to 11 terms !" Is this a punctuation given in the quotation source? if not, I can't see that it should be there: it's not going to be misconstrued as a factorial, for instance. --Kay Dekker (talk) 22:12, 27 February 2010 (UTC)

Italics

The article currently uses ''n'' (n) for indices and A_''n'' (A_n) for members. I think this looks bad, and doesn't match the OEIS (which uses no italics). The 'right' way would seem to be n and A_n, or perhaps <code>n</code> (n) and <code>A_n</code> (A_n).

There is value to italicizing n to draw it out in the text, but when paired with the upright A it has a very unpolished appearance.

CRGreathouse (t | c) 15:58, 27 January 2009 (UTC)

what does at&t have to do with this?

if someone knows the history of how at&t is related to this topic, it would be useful to add. i clicked a link on a link on a wiki page and was taken to an att website, which prompted me to look for more info on this page. the company obviously have something to do with the sequence ecyclopedia, but there is no real reference to this fact in this article. HantaVirus (talk) 17:27, 27 July 2009 (UTC)

The database is hosted on an AT&T web site and run by Neil Sloane, a long-time AT&T employee. I've added some text to the article explaining that. —David Eppstein (talk) 17:32, 27 July 2009 (UTC)

Does anyone want to use footnotes instead of inline links to link OEIS sequences?

A suggestion to change the template in that way has been made at Template talk:OEIS#This is an external link. Lipedia (talk) 15:57, 19 October 2011 (UTC)

Normalizing the sequences for lexicographic order

How is the normalization done? I did a cursory search but could not find out how. Thank you in advance. — Preceding unsigned comment added by Anorderofmagnitude (talk • contribs) 00:17, 16 May 2013 (UTC)

Did you read the article's section on Lexicographical ordering?!?

The examples given start like this:

• Sequence #1: 2, 3, 5, 7, 11, 13, ...
• Sequence #2: 2, 3, 5, 7, 11, 101, ...
• Sequence #3: [0, 1, 1], 2, 3, 5, 8, ...
• Sequence #4: [1], 2, 4, 7, ...
• Sequence #5: [1], [−]3, [−]8, ...

The initial 0's, 1's, and the numbers' signs (which i've explicitly marked above) are ignored, thus the order is #1, #2, #3, #4, #5 instead of #3, #5, #4, #1, #2.

-- Jokes Free4Me (talk) 10:25, 16 May 2013 (UTC)

Thank you much. Yes I did read the section. First the square brackets above made the sequencing order very clear. Second maybe the we can change "ignoring initial zeros or ones" to "ignoring initial zeros and ones". I was only ignoring the initial "0" in sequence #3. — Preceding unsigned comment added by Anorderofmagnitude (talk • contribs) 00:35, 17 May 2013 (UTC)

Okay, i've changed the article to make it all clear. Is it okay now? :) -- Jokes Free4Me (talk) 17:15, 17 May 2013 (UTC)

Yes thank you. — Preceding unsigned comment added by Anorderofmagnitude (talk • contribs) 19:27, 17 May 2013 (UTC)

how useful is it in actual research?

Are there any examples of research in mathematics published in fairly prestigious journals which was aided by OEIS? — Preceding unsigned comment added by 2.97.117.210 (talk) 12:27, 3 September 2014 (UTC)

Look at https://oeis.org/wiki/Works_Citing_OEIS. PrimeHunter (talk) 13:20, 3 September 2014 (UTC)

Speculation on OEIS size limit?

This discussion has been closed. Please do not modify it.
The following discussion has been closed. Please do not modify it.
Minor change from 180284 sequences and The On-Line Encyclopedia of Integer Sequences (OEIS), in the future at least 340142 sequences and no higher than contains 3403xx sequences in around 2020 and will be extincted* 180284+(319716/2) = 340142 sequences. With extinct OEIS in 2020 at 340,142 sequences, the A34xxxx serial with only hundreds of sequences, Last possible A-Numbers will be A3406xx or A3407xx. — Preceding unsigned comment added by 59.101.168.179 (talk) 09:45, 23 November 2017 (UTC) (1) Start a new section in the Talk page if you are introducing a new topic; I have invented one for you. (2) Where do these numbers and "last possible A-Numbers" come from? At the very least, you need a reference. Chris Thompson (talk) 16:02, 24 November 2017 (UTC) Please keep to the guidelines for Talk page discussions,, rather than modifying the text of your original posting. Meanwhile, you still haven't explained why you think OEIS is going to be "extinct" in 2020, or when it acquires 340142 sequences. Chris Thompson (talk) 11:59, 27 March 2018 (UTC) Further, talk pages are for discussions on improvements to the article only, not for general-purpose discussion about the subject of the article. Unless someone has published something about limits to OEIS in a reliable source, we can't include it in the article and so it doesn't belong here. —David Eppstein (talk) 12:04, 27 March 2018 (UTC) Milestones (sequence A300000 in the OEIS) March 2018. (sequence A310000 in the OEIS) 2018. (sequence A320000 in the OEIS) 2019. (sequence A330000 in the OEIS) 2019/2020. (sequence A340000 in the OEIS) 2020. OEIS going to extinct in 2020 at 340,142 sequences? Exceeds the limit of 340,142 sequences though, with milestones from A300000, A310000, A320000, A330000 and A340000. — Preceding unsigned comment added by 59.101.158.170 (talk) 14:27, 19 March 2018 (UTC) — Preceding unsigned comment added by 59.101.168.46 (talk) 09:34, 10 March 2018 (UTC) Future milestones? Wqwt (talk) 05:59, 11 March 2018 (UTC)

Removed "Searching the OEIS" section

Nothing in the section was cited and Wikipedia:NOTMANUAL. Wqwt (talk) 05:57, 11 March 2018 (UTC)

Proposed "OEIS in popular culture" section

This might be worth mentioning: https://xkcd.com/2016/ Eastmain (talk • contribs) 07:19, 7 July 2018 (UTC)

Not without third-party reliable sources. Also see WP:TRIVIA. —David Eppstein (talk) 16:07, 7 July 2018 (UTC)

Unneeded inline <math> usage

The article has inline <math> formatting. I changed that into straight html (wikicode), but twice was reverted by David Eppstein [2] [3].

Clearly Eppstein's argument is changing in steps; WP:RETAIN is not applicable (it is about WP:ENGVAR, nor is WP:DATEVAR applicable); saying "to me your new formatting looks significantly uglier" is not an argument, especially since general accepted wisdom is that one should not mix formattings & font families inline when not needed. -DePiep (talk) 20:46, 5 September 2018 (UTC)

I don't think we should be gratuitously changing one style of mathematics formatting to another, and especially not changing <math> to templates. In general, template-based mathematics formatting is both weaker than <math> (unable to express many important types of notation) and semantically less clean, so in the long term we should move towards always <math>. In the short term, the template-based system has some small advantages in terms of more consistent font usage than <math> but in this specific case I think even those advantages are negated by the ugliness of its formatting (too little vertical spacing between digits and the fraction bar). —David Eppstein (talk) 21:21, 5 September 2018 (UTC)

Updating this article

I think we should update the section "Abridged example of a typical entry" because it is too old. Yifan Xie (talk) 10:40, 1 July 2022 (UTC)

Thanks for your suggestion. When you believe an article needs improvement, please feel free to change it. We encourage you to be bold in updating pages, because wikis like ours develop faster when everybody edits. Don't worry too much about making honest mistakes—they're likely to be found and corrected quickly. You can always preview your edits before you publish them or test them out in the sandbox. If you need additional help, check out our getting started page or ask the friendly folks at the Teahouse. - CRGreathouse (t | c) 20:33, 1 July 2022 (UTC)

Size of database

At the end of the second paragraph, we read that: "As of January 2022, it contains over 350,000 sequences, making it the largest database of its kind."

At the end of History section, we read that: "...by end of July 2020 the database contained more than 336,000 sequences."

There is no source provided for either of these statements.

1. Are there any reliable, secondary sources for the size of the database? If so, they should be included and the two statements should be synced. If not, both statements should be deleted.

2. Assuming that there is a source for the first statement, what evidence exists to support the claim that it is "the largest database of its kind?" Indeed, what evidence is there that there are any other databases of its kind? 76.14.122.5 (talk) 06:00, 1 April 2023 (UTC)

Founding date

When was the enclycopedia founded? 1964, as stated in the infobox, or 1965, as stated in the first sentence of the History section? Neither date is sourced, and neither appears in the NYT article (Note 5) 76.14.122.5 (talk) 06:06, 1 April 2023 (UTC)

1964 is the traditional founding date. In Jan-Feb 1964, this was the state of the OEIS: 20 sequences in a notebook. I *think* (but I'd need to check) that the sequences were transferred to index cards in 1965, and then punched cards in 1967, then an early hard drive (can't remember when). The next stages were HIS (1973) and EIS (1995), followed by the Internet: an email service (1995?) and website (1996). The OEIS Foundation was launched in 2009. So pick whichever date appeals to you.

For sourcing purposes, you can probably find the early dates in the HIS or EIS, and the later ones in one of the published papers on the OEIS. - CRGreathouse (t | c) 23:46, 1 April 2023 (UTC)

Actually the hard drive probably came after the HIS, maybe c. 1980. I can dig up sources if any of this makes its way to the article. - CRGreathouse (t | c) 23:48, 1 April 2023 (UTC)

Interesting history! Are there any reliable secondary sources? 76.14.122.5 (talk) 02:56, 2 April 2023 (UTC)

1964: e. g. doi:10.1007/978-3-319-68376-8_9 colt_browning (talk) 11:23, 2 April 2023 (UTC)

Thanks. Updated article with reference 76.14.122.5 (talk) 16:44, 2 April 2023 (UTC)

new sequence

TWPK's FOUR

first 128 numbers: 2, 4, 4, 3, 1, 2, 4, 3, 3, 2, 4, 5, 5, 4, 4, 4, 5, 4, 4, 4, 4, 5, 5, 4, 4, 4, 5, 4, 4, 4, 3, 4, 4, 5, 3, 3, 4, 5, 5, 3, 4, 4, 4, 3, 4, 4, 4, 3, 3, 4, 4, 4, 4, 3, 4, 4, 4, 3, 3, 4, 4, 5, 5, 4, 4, 4, 5, 4, 4, 4, 3, 4, 4, 5, 3, 3, 4, 5, 5, 3, 3, 4, 4, 5, 3, 3, 4, 5, 5, 3, 4, 4, 4, 3, 4, 4, 4, 3, 3, 4, 6, 5, 5, 6, 6, 6, 5, 6, 6, 6, 5, 3, 3, 6, 5, 5, 3, 6, 6, 5, 5, 3, 3, 6, 5, 5, 3, 6 (first appearance I found: https://www.youtube.com/watch?v=ruI5Olajj3w&t=3434s) 84.151.252.41 (talk) 13:36, 4 August 2023 (UTC)

Hi! This isn't the right place to submit a sequence to the OEIS. If you want to do that, register at the OEIS, read the new user information, and submit a sequence there. If you'd like to contribute to Wikipedia, keep in mind that it does not accept original research in general. - CRGreathouse (t | c) 14:15, 4 August 2023 (UTC)

A095646 and A095647 need to be corrected.

Because 1^x is always 1, the lowest possible base is base 2, so A095646 and A095647 need to be corrected.

source for the correct numbers from bases 3 to 10: https://www.rapidtables.com/convert/number/base-converter.html

A095646 after the correction: 11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111, 202201102121002021012000211012011021221022212021111001022110211020010021100121010, 3333333333333333333333333333333333333333333333333333333333333333, 11031110441201303134210404233413032443021130230130231310, 23053353530155550541354043543542243325553444410303, 3115512162124626343001006330151620356026315303, 3777777777777777777777777777777777777777777, 22642532235024164257285244038424203240533, 340282366920938463463374607431768211455

A095647 after the correction: 340282366920938463463374607431768211455, 22642532235024164257285244038424203240533, 3777777777777777777777777777777777777777777, 3115512162124626343001006330151620356026315303, 23053353530155550541354043543542243325553444410303, 11031110441201303134210404233413032443021130230130231310, 3333333333333333333333333333333333333333333333333333333333333333, 202201102121002021012000211012011021221022212021111001022110211020010021100121010, 11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 94.31.82.138 (talk) 16:05, 15 September 2023 (UTC)

This is not the place to correct OEIS. OEIS has its own editorial procedure. You have to create an account there and propose a correction. —David Eppstein (talk) 16:22, 15 September 2023 (UTC)

largest number fully written out

What is the largest number that is fully written out in OEIS?

The largest I found is A094133(100) = 300827525...933912001, which has 2651 digits. 94.31.86.138 (talk) 17:36, 23 October 2023 (UTC)

The page at https://oeis.org/SubmitB.html specifies: "Remember that the numbers cannot have more than about 1000 digits, or some of the programs will fail.". So I don't think that there is a real limit besides the 10Mb limit, but some programs that automatically process some b-files will fail if your numbers exceed 1000 digits, and 1000 might actually be 1023 or 1024 rounded down. It is also possible that the 1000 limit is old and has been changed to some higher value. To be sure you should ask someone at OEIS that understand the internals of the database. Dhrm77 (talk) 20:58, 23 October 2023 (UTC)

In the sequence entries themselves, 100-300 digits is the practical maximum. The database could handle more but the editors will reject entries that are too long to fit nicely on the screen. (Further terms can go in a b-file.)

There is no technical limitation on the size of terms in b-files. The OEIS has recommended keeping their terms to 1000 digits pretty much from the beginning of the b-files, but there are a number of entries which exceed this (mostly for no good reason, but in some cases it's valuable).

CRGreathouse (t | c) 05:28, 24 October 2023 (UTC)

94.31.86.138's question is not about the theoretical limit of a b-file. It is about the largest number that actually has been written out in a b-file so far. The largest I found is A007138(6437), which has 6240 digits. 94.31.88.138 (talk) 17:11, 18 November 2023 (UTC)

The longest as of this moment seems to be A326609(13) which is about 35,667 digits long. - CRGreathouse (t | c) 05:54, 29 November 2023 (UTC)

missing warning in the links section of A000040

The Table of Primes from 1 to 1000000000000 link in the links section of A000040 needs an output size warning, like the Index entries for sequences related to divisors of numbers link in the links section of A174670 (The warning is "Output is ~13 GB." and is in the prog section of A174670.).

In fact, the list of primes from 1 to 1000000000000 is even larger (about 34.3588 times more data and 88.5957 times more entries) than the list of divisors of 808017424794512875886459904961710757005754368000000000 (There are 37607912018 primes from 1 to 1000000000000!).

warning for the Table of Primes from 1 to 1000000000000 link in the links section of A000040: "Output is ~446.66 GB." 94.31.83.138 (talk) 17:05, 16 November 2023 (UTC)