This site is supported by donations to The OEIS Foundation.

Features Wishlist

From OeisWiki

Jump to: navigation, search

This page is here to centralize all suggestions for features for the OEIS (wiki and main site).

Please, add your signature stamp (next to rightmost button of the editing toolbar) each time you comment on a suggestion. --Olivier Gérard 17:33, 15 November 2010 (UTC)

Contents

Main Site (sequences)

Sequence Editing (General User)

  • Send an email with a copy of the edit.
  • Increase login duration. Charles R Greathouse IV 02:40, 15 December 2010 (UTC)
  • Automatically delete "has not been edited for a week" messages from the server when a sequence is proposed for review. -Franklin T. Adams-Watters, Sep 15 2011

Sequence Editing (Associate Editor)

  • Search sequences with regular expressions
To expand: there are many times when I would like to search for something in the OEIS but which go beyond the capabilities of its (excellent) search engine. One possibility would be to download the entire OEIS and use standard tools like grep. But there are two problems with this approach. First, it takes a lot of bandwidth and time. Second, the OEIS cannot be downloaded in bulk at this time...
Use cases:
  • Instances where the Porter word-stemming prevents a search (for example, finding "decimated")
  • /\b[^0-9Aab]\d{6}\(n\)/ to find A-numbers missing their A
  • /([-a-zA-Z0-9]+\.)*[-a-zA-Z0-9]+\(AT\)[-a-zA-Z0-9]+(com|edu|org|[a-z][a-z])/ (or a slightly more complicated 'correct' version) to search for email addresses
  • /[^\x20-\x7E]/ to find non-ASCII characters like … or ²
  • Finding multiple links per line, sometimes intended but other times not
  • (many others)
I put this in the "Associate Editor" section in case this was deemed too performance-draining to be allowed more generally. But allowing it for everyone would be even better, assuming a good algorithm and implementation (like re2).
Charles R Greathouse IV 21:08, 17 June 2012 (UTC)
  • Global replace with regular expressions.
This is obviously more dangerous than the above, but I think it is still desirable.
Use cases:
  • Search for unlinked user names
    • /(?!_)Charles R Greathouse IV(, (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) [0-3][0-9] 20[01][0-9])/_Charles R Greathouse IV_\1
  • Fixing typos like /mentionned/ as requested by D. S. McNeil (ordinary search fails due to stemming)
  • Standardizing names which aren't searched for properly:
    • Erd\"osErdős (Erdos, Erdös, and Erdős are all searchable, but Erd\"os stems to erd os which does not appear in a search for [erdos])
    • Sz\'ekelySzékelys
    • Szeg\"{o}Szegő
  • Fixing GP scripts using the deprecated & or | syntax. Basic idea (not yet production-ready):
    • /^(\(PARI\).*[^,&\s])\s*&\s*([^&\s])/\1 && \2
    • /^(\(PARI\).*[^|\s])\s*\|\s*([^|\s])/\1 || \2
  • Fixing abbreviations:
    • \bpp. ([0-9]+) (ff?)\bpp. \1\2
  • Fixing attributions:
    • \(from Robert G\. Wilson v ([A-Z][a-z][a-z] \d\d \d\d\d\d)\)(* _Robert G. Wilson v_, \1 *)
Charles R Greathouse IV 14:12, 7 January 2013 (UTC)

Sequence Browsing

  • Ability to search by date ranges, or at least month + year
  • Searches for relations between sequences, checks based on formulas and cross-references, checks for dead links, etc. Joseph Myers 22:30, 21 November 2010 (UTC)
  • Offer an option to extend search over b-files. Max Alekseyev 21:13, 18 December 2011 (UTC)

Sequence Formatting

  • More ways to format tables than traditional triangular ones, the most general way being to use another sequence for the number of terms of each rows.
  • Viewing contents of any sequence in b-file format, or as a decimal constant, or in a different base for "base" sequences related to a base other than 10 (as sequences are sequences of integers not digit strings, entering the integers in base 10 is the mathematically correct way of storing sequences related to any base - but not the most convenient for spotting patterns; some sequences are present in both the mathematically proper base 10 form, and in the own-base form). Joseph Myers 22:30, 21 November 2010 (UTC)

It would be nice to have a good way to search 'formatted' sequences (triangles, arrays listed by antidiagonals, and generally any tabl or tabf) Alexander R. Povolotsky 15:09, 1 November 2011 (UTC)

On another page of this wiki I gave a link to the PHP script I wrote for this and which I'm ready to share. — M. F. Hasler 14:19, 7 November 2011 (UTC)
OEIS table search? Charles R Greathouse IV 18:00, 7 November 2011 (UTC)

Sequence Watching

  • rss flux mechanism to receive updates about sequences.
  • RSS for all recent changes is useful; there should also be personal watchlists so you can get (by web, RSS etc.) lists of recent changes to the particular set of sequences of interest to you. Joseph Myers 22:30, 21 November 2010 (UTC)
  • Move from the explicit "more" markers for sequences needing extension to an automatically generated list of sequences in ascending order of the number of terms present (including in any b-file), together with some marker to indicate that a sequence should not be extended because the next term would exceed some limit on the size of terms present. Joseph Myers 21:52, 21 November 2010 (UTC)
    This might be nice for the "help needed" sequences, though -- it could display both sequences that have been marked "more" and ones it decides could also use extra terms. I'm not sure that it should replace k:more; there are sequences which don't need extra terms, even though few appear, as well as sequences where extra terms are desired even though we already have many. Charles R Greathouse IV 22:23, 21 November 2010 (UTC)
    Like Charles, I believe the "more" keyword is still needed. But I suggest at least 2 "help needed" queues: one with "more", "hard", and non "fini, full" prioritized to induce contributors on working on them. Another with "uned", "obsc" or similar idea (such as ref needed, conjecture to prove, etc.) to induce users to enhance and edit them. Olivier Gérard 06:05, 30 March 2011 (UTC)
    It would be nice to distinguish between the various types of 'help needed's. You mention conjectures: I can imagine many types of conjectures, from ones which just need eyes to ones equivalent to major unsolved problems. I don't think I have a good idea yet on how these should be handled. Charles R Greathouse IV 03:40, 11 July 2011 (UTC)

Sequence Metadata

  • Categorization feature, possibly building on the current links to the Index
    The Index should be built out of annotations in the individual sequence entries. These might get a bit complicated for the cases where there are multiple sets of related sequences, each set appearing together in the index in a consistent order (e.g. a group of related sequences for base 2, the corresponding sequences for base 3, etc.), but I think it's the best way to keep the Index up to date whenever corrections are made to sequences mentioned in it. Joseph Myers 12:43, 24 November 2010 (UTC)
    It's hard to search for these compared to searching for keywords. They're also hard to add: you can just type "mult", but can you remember the wording and the link to linear recurrence relations? Also I have concerns about "related to" vs. "is a member of"; the former is what the index is best at, while the latter seems more suited to a keyword/category/other metadata. We have thousands of linear recurrence relations; it's not practical to browse them directly, not even when broken down by signature. Charles R Greathouse IV 15:10, 26 January 2011 (UTC)
  • Expanded keywords, unless included in the above
  • More information for each keyword: base(10), base(factorial), fini(conjectured), cons(10); these should be standardized in some way.
  • Better integration of b-files: changing a b-file should automatically change the associated sequence terms and changing sequence terms or offset should automatically update the b-file; it should be possible to use a b-file to provide the terms of a new sequence instead of pasting in those terms; b-files should be version-controlled like the sequence entries properly; "full" keyword should apply to the set of terms in the b-file. Joseph Myers 21:52, 21 November 2010 (UTC)
    Better integration sounds good, but I would be careful here. If a b-file is submitted which does not match the current terms of the sequence, the most likely reason is that the b-file is wrong, not that the sequence is wrong. (For example, because the user uploaded the wrong file or misunderstood the definition.) I don't oppose changes here as long as they retain the error-correcting features of the current version. Charles R Greathouse IV 17:48, 5 July 2012 (UTC)
  • Similarly, automatically processed relations between sequences; if a set of sequences are known to be the same except for offset and initial terms, extending one should extend all of them (and all their b-files); likewise for linear relations between sequences. Joseph Myers 21:52, 21 November 2010 (UTC)
  • Some way to describe the density of the sequence, when appropriate: cofinite, density 1, positive lower density, fractional power density, log density, etc.
  • Microformats integration: hCalendar? hCard? Appropriate rel attributes?
  • There should be a more keyword not only for "more terms" but also specifically for "more crossrefs", "more programs" in particular (maybe others). Somehow a rough equivalent to the "stub" category of Wikipedia. (Or maybe that could be done via a link on an appropriate (wiki) page, but would certainly receive less attention.) — M. F. Hasler 14:17, 7 November 2011 (UTC)
  • Sort keywords lexicographically in the keywords field of each entry. Currently there is no sorting imposed on the keywords. Max Alekseyev 09:13, 7 December 2011 (UTC)
    I think this would be nice. Conceivably it would make searching faster, too, if keyword lookup is ever a hotspot. Charles R Greathouse IV 17:48, 5 July 2012 (UTC)

Other features

  • Reference and external link databases separate from the uses of those references and links in individual sequences, so only one place needs changing to update a widely used reference or link (e.g. changing a reference to a preprint to refer to the final version, updating a link to a website that has moved). Joseph Myers 22:30, 21 November 2010 (UTC)
  • Support for DOI-linking in the References field. E.g., doi:10.2307/2007151 in A061034 should be hyperlinked to http://dx.doi.org/10.2307/2007151 Max Alekseyev 08:00, 19 May 2011 (UTC)
    • Nice idea. Similarly (but 10 times as common) would be arXiv links. Charles R Greathouse IV 19:06, 2 June 2011 (UTC)
+1 from me (until I have my browser(s) doing this on any web page...) — M. F. Hasler 12:42, 28 March 2012 (UTC)
  • It would be nice to have a third Plot in the button “Graph”. The appearance of the third Plot should be the same as "Plot 2" or “Plot pairs of sequences in the OEIS” with the following configuration. Sequence 1: A?????? untransformed (our sequence). Sequence 2: A000004 (always) untransformed. Contents of Plot: “Plot Seq2(n+shift) vs n and Seq1(n) vs n”. Appearance of Plot: “Draw line segments” (only). For example see the third Plot of A005132 Recaman’s sequence. See also A064413 EKG sequence. Many other sequences may show a nice graph using this Plot. Omar E. Pol 11:43, 14 April 2012(UTC)
    • Similarly, I think it would be nice to have a link to Plot 2 which filled in the name of the sequence in one of the slots. Charles R Greathouse IV 15:11, 24 August 2012 (UTC)
How come I don't have a third plot in A005132/graph and A064413/graph? I only have the standard two plots. — Daniel Forgues 02:34, 25 August 2012 (UTC)

OEIS Wiki

Logo, Appearance

(...)

Main page

See Main Page/Prototype. — Daniel Forgues 05:10, 24 November 2011 (UTC)

Help Pages

  • Translated version of the main help pages
    • The pages should probably be finalized before attempting translation. Charles R Greathouse IV 00:45, 31 January 2011 (UTC)
      • They're probably as finalized as they're going to get now. Anyone up for translation? Charles R Greathouse IV 21:14, 17 June 2012 (UTC)

Information Pages

(...)

User Pages

  • It would be useful if a mechanism could be put in place to enable unreviewed user Wiki pages to be renamed by authors or editors so that they can be made to comply with an approved convention. Daniel Forgues emailed me about a month ago stating that he tries to follow the convention:
    • only capitalize the first word (unless proper names, or some exceptional Conjecture, Hypothesis, Theorem, ...)
    • space between words (instead of underscore or no space at all)

Such a mechanism would preclude the need for redirections.

Chris Gribble 20:25, 16 September 2011 (UTC)

Special Pages

Various known OEISwiki bugs and limitations would be fixed by the current MediaWiki version. –Frank Ellermann 13:56, 23 September 2011 (UTC)

Links

(...)

Special Syntax

(...)

Categories

  • The wiki categories previously contained the wiki sequence pages; now that all the sequence pages have been removed from the wiki, the wiki categories contain no more those sequence pages. It would be great to have a way to have the sequence pages listed in the wiki categories like before. — Daniel Forgues 20:26, 17 November 2010 (UTC)
    • I am looking for Guidelines for creating categories (right now for wiki pages, not for sequences, although I favour this idea): Naming convenions concerning capitalization, singular/plural (Category:Style sheet or Style Sheet or Style sheets ?), etc...) Didn't find such.... — M. F. Hasler 19:49, 20 January 2013 (UTC)
I've mostly named categories and pages like Category:Perfect numbers and Perfect numbers (only capitalize first word and proper names) and plural form. Sometimes exceptions were made for e.g. Category:Riemann Hypothesis and Riemann Hypothesis, Category:Fermat's Last Theorem and Fermat's Last Theorem (actually not, it is Category:Fermat's last theorem and Fermat's last theorem) ... — Daniel Forgues 00:58, 21 January 2013 (UTC)

Typesetting formulas

Cf. Talk:Features Wishlist#Using MathJax. — Daniel Forgues 03:53, 29 July 2012 (UTC)

Other features

Miscellaneous

  • It would be nice to be able to log in to both sides (sequences and wiki) of the site at once, instead of having to enter the password twice. Alonso del Arte 01:43, 17 November 2010 (UTC)
    • (On a related note, I wonder if I'm the only one for who has to be careful not to log on to both IMSLP and OEIS wiki at the same time?) Alonso del Arte 01:43, 17 November 2010 (UTC)
      • What do you mean ? What happens when you do ? Olivier Gérard 06:11, 30 March 2011 (UTC)
        • You get kicked off the other one (you have to log in again). But it doesn't matter if you're logged in to OEIS Main, so Clark Kimberling might not have noticed this. Alonso del Arte 05:28, 17 September 2011 (UTC)

Completed

2012

  • The links to the Maple and Pari programs on the page Transformations are broken. There is no link to the Mathematica version; instead this version can be found on a separate page Trans-MMA. See also [1] and [2]. Peter Luschny 23:32, 16 January 2012 (UTC)
    What I imagine is one wiki page introducing integer transformations, then a page for each transformation so that all aspects of the transformation can be discussed: definition, history, uses, properties, implementation issues, etc. Then a separate wiki page for each programming language, on which the code for all transformations can be entered. Possibly the individual transformation pages should redundantly include code (in all languages) for that particular transformation; it sounds like that could be a lot of clutter, though.
    It absolutely breaks my heart to see pages like Srinivasa Ramanujan‎ and Template:Divisor function on the OEIS wiki, but not the transformations. The MacTutor History of Mathematics archive and Wikipedia will always be a better source than the OEIS for biographies, and computer algebra packages and sites like Wolfram|Alpha will always be better at computing arithmetic functions. But where else can you go to learn about the "logarithmic derivative of exponential generating function" transformation, or the uses of the Boustrophedon transform?
    Charles R Greathouse IV 03:06, 17 January 2012 (UTC)
    See Sequence transforms. Charles R Greathouse IV 17:34, 5 July 2012 (UTC)

2011

  • http://oeis.org/recent.txt is no longer populated. Please re-implement and maintain recent.txt. This is very useful backwards compatibility feature User:Alexander R. Povolotsky
  • Now that /wiki/-entries of A-entries have been removed, would it be possible to to hack quick-references to sequences, like e.g. [[A033538]] to point http://oeis.org/A033538 instead of old http://oeis.org/wiki/A033538 ? Or is there another "magical notation" which shows the title of the sequence with mouse-over? (Like in "Classic View".) — Antti Karttunen 16:51, 17 November 2010 (UTC)
    • I would strongly recommend not to do this. I see wiki pages like http://oeis.org/wiki/A033538 as places in which it would be possible to store additional explanations (construction rules, examples etc.) to sequences for which the definition is too long to be put into the comment field (as an example see A048888). For these sequences the exact definition could be elaborated in the wiki page, and the comment in the OEIS page could get by with a descriptive summary and a link to the wiki page. --Jens Voß 08:37, 18 November 2010 (UTC)
    • Using the template {{oeis|A000040}} (e.g. A000040) to link to sequence pages instead of [[OEIS:A000040]] (e.g. A000040,) or [[OEIS:A000040|A000040]] (e.g. A000040) to omit displaying the prefix, is more convenient if we don't want to display the OEIS: prefix. Should we use the template throughout and avoid displaying the OEIS:prefix on links to sequence pages or do we want to keep the prefix visible to emphasize the fact that the sequence pages are not in the wiki? Also, if the sequence pages are ever moved in the future (e.g. from oeis.org/ to oeis.org/en/ for the english version, oeis.org/fr/ for the french version...) by merely updating the template all the links to sequence pages will be working again (actually the template is not necessary for this since the OEIS: prefix functionality can be updated...) — Daniel Forgues 17:40, 18 November 2010 (UTC)
    • It would be nice to have title-text for sequence links in the wiki. I don't think there's any easy way to do this, unfortunately. Charles R Greathouse IV 00:41, 31 January 2011 (UTC)
  • Backlinks from sequence to all linking to it (requested by F. Chapoton, 11:44, 18 August 2011)
  • Some way to modify many sequences at once, as was possible for Dr. Sloane with the old system. (Did Russ say he wanted to do this? I seem to recall that, but I have no SeqFan archive link.) Charles R Greathouse IV 00:30, 22 November 2010 (UTC)
  • If it's not complicated to do or for other reasons undesirable, it would be nice if http://oeis.org/?q=1234 would be the same as http://oeis.org/search?q=1234, i.e. when a nonvoid ?q=... "argument" (query string) is given, it should do the search also on the welcome page. This would also make some redirections work, e.g. http://www.research.att.com/~njas/sequences/?q=A007185, which currently does not work although the query string is preserved upon redirection to oeis.org. — M. F. Hasler 00:10, 17 November 2010 (UTC)
  • Upload b-files for sequences. (now available - but still manual intervention to replace an old one Olivier Gérard 05:56, 30 March 2011 (UTC))
  • Would it be better to use keyword "modified" for changes (instead of using keyword "new" - while keeping it to be applied only to "brand new" sequences" ) ? User:Alexander R. Povolotsky
    Keyword:changed added by Russ Cox, Jan 24 2011. Charles R Greathouse IV 20:00, 24 January 2011 (UTC)

2010

  • I strongly request implementing emails sent from OEIS to authors/editors, confirming the fact of the sequence's submission and/or reflecting the change of the sequence status ( such as "being edited","rejected" or "approved") and containing changes made. Besides for asking for bringing back that feature for the backwards compatibility reason alone, those confirmation emails also in the previous OEIS format served authors to retain on their end as some way of the history record keeping. This was a very important OEIS feature and needs to be reimplemented in OEIS wiki. User:Alexander R. Povolotsky
    This is apparently implemented now. Charles R Greathouse IV 02:58, 22 November 2010 (UTC)
  • Have a way to explain the reason of a rejection (note Russ Cox said on seqfan he was going to implement this feature ).
    This is apparently implemented now. Charles R Greathouse IV 02:58, 22 November 2010 (UTC)
>29,000 sequences were edited, mostly to update 
>links to supporting files, so there are 29,000 sequences with 
>keyword:new. 
>keyword:new is automatically applied to all sequences that have 
>changed in the last two weeks.  This is the same condition that 
>was used before the transition to the wiki format. 
  • Downloads of sequence terms (stripped.gz link is currently broken), b-files and the other parts of sequence entries for use in automatic processing Joseph Myers 22:30, 21 November 2010 (UTC)
Personal tools