Talk:Clear-cut examples of keywords

nonn vs sign

Keyword sign is used for

A052435 Round(li(n) - pi(n)), where li is the logarithmic integral and pi(x) is the number of primes up to x.

The first crossover (P. Demichel) is expected to be around 1.397162914*10^316. The first negative term will thus forever be out of term visibility since there are only ~ 10^80 protons in the observable universe... — Daniel Forgues 04:20, 30 November 2011 (UTC)

A literal interpretation of the official description of the keywords would mean that that sequence should have keyword:sign. But I am inclined to agree with Charles's position on the matter:

"I think that this keyword [keyword:nonn] should not be used when there are known to be negative members of the sequence, even if they do not appear in the sequence data."

Actually, I agree wholeheartedly with that position. But if it is believed that a sequence contains negative terms, it might be better to go with nonn. Such a sequence would not make for a clear-cut example. Alonso del Arte 19:55, 30 November 2011 (UTC)

cofr

Can we use keyword:cofr for generalized continued fractions? — Daniel Forgues 01:29, 5 December 2011 (UTC)

No. It might be nice to have a keyword for that, but until that's done the keyword cofr should be used only for simple continued fractions. Charles R Greathouse IV 06:55, 5 December 2011 (UTC)

walk

Should sequences related to meanders, e.g. meanders filling out an n-by-k grid (reduced for symmetry) have this keyword or not? — Daniel Forgues 05:18, 12 December 2011 (UTC)

Practically speaking they do not have this keyword. Should they? Charles R Greathouse IV 15:46, 12 December 2011 (UTC)

more example not clear-cut

The current example at Clear-cut examples of keywords#more is A005820 and says "presumably extending this sequence is just a matter of letting a computer search run". But the sequence has comments indicating it may be full. Several sources say it is believed to be full, although this is unproven. Jens Kruse Andersen 10:23, 31 July 2014 (UTC)

Yeah, probably not the best example considering the title. Do you have a better? - Charles R Greathouse IV 04:33, 1 August 2014 (UTC)

There are lots to pick from at https://oeis.org/search?q=keyword%3Amore. We probably want a hard sequence with low risk that 3 lines are filled later without updating the example. Apart from that it's not terribly important but I suggest A123692. It has a simple definiton, it's not bref but has less than a line, it has large terms found by computer and states a search limit, it's probably infinite, and it doesn't rely on probable primes like many other hard sequences. Jens Kruse Andersen 12:05, 1 August 2014 (UTC)

Let's go with A123692 until someone finds an even more clear-cut example. In any case it's a much clearer example of keyword:more than A005820. Alonso del Arte 17:17, 1 August 2014 (UTC)

Repetitive sequences

So is there any metadata for distinguishing repetitive sequences from non-repetitive ones? "fini" is sort of appropriate, but implies that the sequence stops, not that it repeats from the start, and isn't used in https://oeis.org/A011655. Just search for the word "repeat" or "period"?

A011655 is most definitely NOT a finite sequence, for it is well-defined for all positive integers. What if n is a googolplex? Skewes number? Graham's number?

On the other hand, A003173 is not well-defined for all positive integers. What is the tenth value of d such that ${\mathcal {O}}_{\mathbb {Q} ({\sqrt {-d}})}$ is a unique factorization domain? There isn't one, because only nine imaginary quadratic integer rings are UFDs.

Maybe there are periodic sequences that are finite, though I can't think of one at the moment. I'm pretty sure almost all periodic sequences are infinite.

So yeah, look for "period" or "periodic." I don't know about "repeat," I think that might give you too many red herrings. Alonso del Arte 04:07, 17 August 2014 (UTC)

Actually, with the stemming we use, period = periodic, so you can't search one without searching the other.

I like the idea of adding more keywords (and have a page about this). I don't know offhand if one has been suggested for sequences with a finite image. Charles R Greathouse IV 20:35, 18 August 2014 (UTC)

Ah yes, there could be finite and repetitive sequences, so they really are separate things. Jonathan Bright 22:53, 10 September 2014 (UTC)

frac

https://oeis.org/A002487 does not have the frac keyword even though it makes the Calkin-Wilf fraction sequence when combined with a shifted version of itself?

"For sequences of unit fractions"

Do these get the frac keyword? Can you list an example in the text here? https://en.wikipedia.org/wiki/Unit_fraction#Series_of_unit_fractions has some pretty trivial examples.

"only the sequence of denominators need be added; the numerators are A000012."

But that sequence doesn't have a frac keyword because it's too simple? Jonathan Bright 22:53, 10 September 2014 (UTC)

You can take any sequence of nonzero integers and interpret it as a sequence of denominators for unit fractions, so, that means, for example, that A000027 shouldn't have keyword:frac even though we can say it consists of the denominators for the harmonic series.

But your question raises other points I'll have to consult the other Editors on. Alonso del Arte 00:34, 11 September 2014 (UTC)

Maintenance

The table of contents (and the complete page) could look better, if the six #Maintenance keywords allocated, changed, dupe, new, probation, and recycled are simply sorted alphabetically with all other keywords. One footnote <ref name="auto">Never added manually</ref> for allocated could help, also used as <ref name="auto" /> for changed, new, etc. in no particular order. The footnote could be inserted immediately above or below #See_also as <references />. Nothing is special with recycled, it's the documented way to retire "own" allocated sequences, "own" hopeless submissions, or (as editor) to delete NOGI, rifo, etc. –Frank Ellermann 04:28, 13 October 2017 (UTC)

Implemented on "probation"; this keyword is apparently really unused at the moment, ditto "dupe". Very vaguely I recall that "dupe" was not some special case of "dead", it was really some kind of dupe to be recycled soon (like 'rifo' deletions today). OTOH "dead" was and maybe still is supposed to stay as tombstone/redirection, because external references (offline as in printed) of the dead sequence exist. Fun fact, keyword:recycling or keyword:recycl find keyword:recycled, but keyword:recyc doesn't. –Frank Ellermann 08:17, 14 October 2017 (UTC)

JFTR, wikipedia:OEIS mentions probation, if or when keyword:probation is seriously obsolete here it should be also fixed there. It's a nice workaround for "could somebody else please delete this, or maybe even rescue it", some kind of uned for the 3rd millennium. –Frank Ellermann 22:29, 17 October 2017 (UTC)

Thanks for your edits! I mostly support them, except for this part: I believe that since the only proper clear-cut example for keywords not currently in use such as probation is don't use them at all, they shouldn't even have their own sections. --Andrey Zabolotskiy 12:51, 25 October 2017 (UTC)

Okay, let probation go to the dupe section here, it's the garbage bin for historic keywords. Assuming that you are certain that it's historic, i.e., it should be also fixed on enwiki. Just in case, the garbage bin is (still) only visible in the draft. –Frank Ellermann 17:41, 13 November 2017 (UTC)

Section links

Section links, i.e., section titles consisting of a wikilink, are generally frowned upon, and here they are mostly useless wikilinks to empty subcategories, cf. category talk:keywords. Suggestion: Let's kill them all, and add nice OEIS lookup links as found on User:Charles_R_Greathouse_IV/Keywords to the prose; maybe skipping some keywords too broad for the purposes of a simple lookup (nonn, sign, etc.) –Frank Ellermann 07:17, 13 October 2017 (UTC)

I agree with you about the section titles, I have edited accordingly. Charles has many great ideas and I might get around to adding lookup links if someone else doesn't get to it first.

But I disagree with you about omitting nonn and sign. New contributors don't know what those mean and are unaware that it might sometimes be necessary to override the system's choice. Alonso del Arte 17:16, 13 October 2017 (UTC)

What I really wanted to say: nonn + sign need no lookup-link, because together these keywords are supposed to cover the complete database. –Frank Ellermann 18:13, 13 October 2017 (UTC)

I'm sorry, that's what I should have understood the first time. Thank you very much for clarifying. Alonso del Arte 02:04, 14 October 2017 (UTC)

Stopped editing after reaching 107 pending changes for obvious reasons, and running out of steam. ;-) –Frank Ellermann 05:18, 15 October 2017 (UTC)

Some observations

See #Maintenance above for the probation business, I've erroneously wikilinked this section in two edit histories.

1. tabf/tabl/walk: Blanked and replaced by lookup links with a fresh EXAMPLE GOES HERE  blurb. Apparently folks mis-understand this and add content discussing not yet existing sequences with not yet approved tabl keywords, or similar. I'll try to find preliminary examples a.s.a.p., simply replace them by a (or two) clearer examples for the relevant keyword.
2. Dated fini vs. full info removed, the "new" 2013 rules make sense, no further details needed.
3. Arguments about hard as proper subset of more: Possible, but the page should cover the state of the art plus relevant historic/technical info wrt keywords, it's not a wish list or talk page.
4. Cross-namespace links to user sub-pages: That's a bad idea, user pages are typically not supposed to be edited by other users (excl. talk pages or maintenance issues, of course). If a user sub-page should be moved to the article namespace just suggest it on its talk page. Copy+paste might be also possible if it has only one contributor, but typically it would be not okay (lost edit history, needs manually added credits, etc.)
5. Apparently all users with review rights gave up on (not only) this page in 2014, is there a category for revived review requests?
6. The wiki always wants me to look at the reviewed version, I always want the draft. Is there a preference or another trick to get always the draft?
7. I do not grok eigen, same as it ever was, even after submitting an eigen sequence only to figure out what it is back in 200x. No issue if it's only me.
8. One bogey in category:keywords not yet fixed (see #4 above), and #8 is no octal digit.

–Frank Ellermann 22:51, 14 October 2017 (UTC)

As an associate editor, I could mark reviewed, but I don't know if at this point it needs an EiC like Charles to bring this updated version to the front.

However, I want to see the tabf example restored, I think it's a very good example. Neil and I are fond of Losanitsch's triangle, but Pascal's triangle is a better example of tabl. But we need to show Pascal's triangle formatted as a triangle, and to mention the "table" link that shows these sequences as triangular, square and upper triangle. The walk example needs to show a lattice diagram.

Lastly, I reassure you that I frequently refer new contributors to this page, because we can't just expect them to somehow know this stuff already. - Alonso del Arte 16:23, 15 October 2017 (UTC)

Okay, it really will take Charles validating to get all these changes reflected in the stable version. For what it's worth, I would sign off on it without a good walk example, but to remove the tabl and tabf examples are steps backwards.

I'm still not sure organizing the keywords alphabetically is such a good idea, but I concede that a lot of people will not be reading this in a sequential manner beginning to end. - Alonso del Arte 17:39, 16 October 2017 (UTC)

To see keywords in action folks can click on the A-number(s) given as example(s), the walk in an older version did not even mention a sequence by its A-number, and the layout of tabl in that version was broken. The current tabf+tabl examples work for me. –Frank Ellermann 17:31, 17 October 2017 (UTC)

Then maybe we're ready to ask Charles to look it over? Alonso del Arte 18:40, 17 October 2017 (UTC)

Admittedly I don't get how this "sight"+"validate" process on the Wiki (here) works:

On dewiki "sighted" is the publicly visible state, it indicates "not obviously nonsense or vandalism". On commons "patrolled" means "somebody with patrol rights checked that it's not obviously..." (as above), on enwiki it's similar. Folks get auto-patrol rights before they are ever sighters/patrollers, and mostly that means that the sighters/patrollers do not more expect that an auto-patroller turns into a spammer or vandal.

Slightly odd for me, because for me it's easy to spot issues created by others, and hard to judge my own contributions. Apparently you use some four eyes strategy, any editor can "review", and after a review any other editor can "approve". Fine for sequences and critical wiki pages not limited to wiki pages with incoming links from sequences.

But everything else on the wiki, who cares, "undo" is easy, "redo" is also easy, and after that folks should discuss what's better; with a minimum of admin or editorial oversight, because the admins/editors are supposed to work on more interesting stuff than "clear-cut examples of keywords", broken hints trying to find ~njas, or disclaimers shown at the bottom of all wiki pages and ending up as a redirect to /dev/nul. –Frank Ellermann 19:33, 17 October 2017 (UTC)

I didn't choose this system, that's my excuse for not fully understanding it. Sometimes I can set the stable version, sometimes, like here, I can't. I do consider this one a much higher priority than the Some integers template, which, by the way, I have to reply to your suggestion there. - Alonso del Arte 21:15, 17 October 2017 (UTC)

tabf

The tabf example is arguably ugly, because I didn't try to prettify it. How about this:

Example: A027746 Triangle in which first row is 1, $n$-th row ($n>1$) gives prime factors of $n$ with repetition. Flipping rows and columns this yields…

 2  3  2  5  2  7  2  3  2 11  2 13  2  3  2 17  2 19  2  3  2 23  2  5
       2     3     2  3  5     2     7  5  2     3     2  7 11     2  5
                   2           3           2     3     5           2
                                           2                       3  

Now it's obviously nice even for me, i.e., I finally understand what it actually is. ;-) Sorry, old eyes, but the vintage August 2014 quality page, the current draft with 112 pending edits, and my ASCII art here all missed the first row with a 1. –Frank Ellermann 02:23, 22 October 2017 (UTC)

Thank you very much for that, I've used it replace the irregular table in the current draft.

For what it's worth, I've reviewed your changes to four pages you've edited recently and now those are the stable versions (maybe tomorrow morning I go through the other three).

I currently have the authority with those pages, I currently don't for this page. Of course this page is more urgent, but not so urgent that I'm deterred from referring a new contributor to it. Alonso del Arte 03:46, 22 October 2017 (UTC)

Thanks, next step, shall I create User talk:Hilarie Orman to get the numerous waiting drafts and missing reviews on their way, or is that basically a bad idea like pestering NJAS etc. with Wiki stuff? –Frank Ellermann 18:11, 13 November 2017 (UTC)