Wiktionary:Votes/2023-08/Changing how the section "References" works

Changing how the section "References" works
Voting on: the way the sections "References" and "Further reading" work, as well as what content is in them.

One vote may be cast for each proposal. A vote of "no consensus" or against all proposals would result in none of the proposals being implemented, and the status quo as described at and  would remain in place. As expanded upon at, this status quo does not imply any obligation to format the "References" section in a specific way. If more than one proposal achieves consensus, the proposal with the highest support ratio will be implemented.

Schedule:
 * Vote starts: 00:00, 8 September 2023 (UTC)
 * Vote ends: 23:59, 7 October 2023 (UTC)
 * Vote created: Vininn126 (talk) 19:13, 21 August 2023 (UTC)

Discussion:
 * [[Image:Wikt rei-artur3.svg|20px]] "Further reading" vs. "References" for Belarusian

Previous votes:


 * [[Image:Wikt rei-artur3.svg|20px]] "References" and "External sources"
 * [[Image:Wikt rei-artur3.svg|20px]] "External sources", "External links", "Further information" or "Further reading"

Proposal 1: Add a new subheader
Proposal 1: Both the header "References" and "Further reading" stay, but when references are added that are not cited inline, "References" is split in two, containing a section with only inline references and a section for non-inline references.
 * 1) Without subsections,   may contain either inline references, handled by   or, or a list of non-inline references.
 * 2) If both inline and non-inline references are listed, the non-inline references must be listed under the subheading , nested under.
 * 3) Allow only supplementary material which is not used in creating the entry under.

Rationale: This gives editors more control and discretion while providing enforceable criteria for what kinds of links can go where. By having a subsection, all works whose materials are used in the entry can be listed as a reference while not cluttering up the style and functionality of the text printed by  or. It also clarifies that non-inline references used as sources for an entry are still references and not supplementary reading. This comes at the cost of being less bot-enforceable than proposal 3.

===References===   21:22, 8 September 2023 (UTC)
 * 1)  Re-read the proposal and it really isn't an improvement to the current situation. Thadh (talk) 21:52, 8 September 2023 (UTC)
 * 2)  I don't really think adding new or different (sub-)headings is going to reduce confusion. We might as well stick to the current situation. — Sgconlaw (talk) 05:37, 10 September 2023 (UTC)
 * 3)  Vininn126 (talk) 22:28, 11 September 2023 (UTC)
 * 4)  Virtually all Bulgarian entries pretty much exclusively use reference templates (e.g. ) in the References section vs inline references. We'd have to rename every References section for Bulgarian entries to Additional sources, which isn't an improvement. Chernorizets (talk) 09:44, 17 September 2023 (UTC)
 * No! If there are no inline references, no change is required.  Moreover, such references can be converted to inline references, or for one and  for the others. --RichardW57m (talk) 12:23, 20 September 2023 (UTC)
 * @RichardW57m Thanks. Apparently I hadn't fully parsed the if-then-else branches of this proposal. It seems I'd only have to move all non-inline references under "Additional sources" if I add an inline reference, unless I had already artificially made everything an inline reference. I don't think this would come up a lot in Bulgarian entries, but I also don't think people (including myself) would remember to split the section in the cases when it should be split. Chernorizets (talk) 22:31, 20 September 2023 (UTC)
 * But the splitting is something that can (and I think would) be done by bots. It's also easy for passers-by to fix.  And non-inline references wouldn't have to be moved.  It's just that two lines (plus a blank line? - again a bot job) would be inserted at the top of the references section.  --RichardW57m (talk) 08:28, 21 September 2023 (UTC)
 * 1)  Adding an L4 header that is used among the other two existing L3 headers will only make matters more confusing. The wording "Additional sources" is also a bit mouthful and out of place. – wpi (talk) 13:41, 21 September 2023 (UTC)
 * 2)  Per my talk message.  ·~   dictátor · mundꟾ  15:08, 21 September 2023 (UTC)
 * 3)  -- Huhu9001 (talk) 04:13, 28 September 2023 (UTC)
 * 4)  A new subheader creates more complexity. IMO the optimal approach is to find a way/style of bullet-pointing non-inline references in the   section such that they are consistently indented with the numbered lists of inline references generated by /, coupled with a strong policy that any material that is intended to provide general corroboration of all or part of the entry content must be placed under  . The unsightly inconsistency of indentation between bullet-pointed and numbered lists (the "ugly mismatch" mentioned by User:RichardW57m above) is currently what inhibits me from placing non-inline (unspecifically corroborative) references and inline (specifically corroborative) references next to each other under  . Voltaigne (talk) 19:05, 28 September 2023 (UTC)
 * One way to align them is by adding a colon before the non-inline references like it's done on starve the beast. lattermint (talk) 15:05, 1 October 2023 (UTC)
 * Thanks for the tip—I have implemented it on . The text in the  section is consistently indented but the digit/bullet-point alignment still looks a tiny bit off. Certainly an improvement, though. Voltaigne (talk) 22:41, 1 October 2023 (UTC)
 * 1)  I don't see a need for the   subheader. It will probably only make things more confusing. Netizen3102 (talk) 17:23, 29 September 2023 (UTC)
 * 2)  कालमैत्री (talk) 06:04, 2 October 2023 (UTC)
 * 3) . DonnanZ (talk) 19:52, 3 October 2023 (UTC)

Abstain

 * 1)  I may have missed the problem that this proposal is intended to solve, so I wont support or oppose this solution, but it is the most agreeable to me in that it doesnt have the two problems that led me to oppose the other two proposals. — Soap — 17:00, 11 September 2023 (UTC)
 * Perhaps deliberately omitted it.  The problem this approach solves is the formatting mismatch between in-line references and non-in-line references.  The change of format is hidden by the new L4 header within the L3 section.. --RichardW57m (talk) 12:42, 20 September 2023 (UTC)
 * 1)  This is actually an issue I'm interested in, because I've grappled with the question of where to put references and how myself, but I don't think the solution is adding or renaming headers. (Of the available options I prefer this one, proposal 1, because I imagine most entries won't need to change.) I don't see the case made well enough for separating reference works used/not used in the entry. I think it'll increase confusion.
 * I'd rather see the status quo formalised by updating Entry_layout to be clear about what goes where. If you used the reference for the whole entry and you don't know where to put the footnote, you could have a guideline for that. —Caoimhin ceallach (talk) 15:48, 27 September 2023 (UTC)

Proposal 2: Add a new header

 * 1) Add the L3 header   above.
 * 2) Allow only references cited inline in the   section, handled by   or.
 * 3) Allow only non-inline references under.
 * 4) Allow only supplementary material which is not used in creating the entry under.

===Citations=== , but oh well. Catonif (talk) 14:08, 8 September 2023 (UTC)
 * 1)  Thadh (talk) 19:58, 8 September 2023 (UTC)
 * 2)  Vininn126 (talk) 22:28, 11 September 2023 (UTC)
 * 3)  I'm not fully convinced we have a problem that really needs solving, but out of the listed options, this one is least disruptive to Bulgarian entries. Chernorizets (talk) 09:48, 17 September 2023 (UTC)
 * 4)  per AG202, though I would prefer using a different wording to avoid confusion with the citations tab, perhaps "Notes" or "Footnotes". – wpi (talk) 13:41, 21 September 2023 (UTC)
 * 5)  Ioaxxere (talk) 03:49, 22 September 2023 (UTC)
 * 6)  Seems like a good distinction. Netizen3102 (talk) 17:43, 29 September 2023 (UTC)
 * 7)  but third better कालमैत्री (talk) 06:05, 2 October 2023 (UTC)
 * 8)  --  18:39, 4 October 2023 (UTC)
 * 9)  Mcph2 (talk) 06:12, 7 October 2023 (UTC)

Oppose

 * 1)  This is only partially bot enforceable as opposed to fully bot enforceable: A bot cannot tell the difference between something that should be classified under "References" or under "Further reading". Also, it's unrealistic that a distinction will be maintained in practice between sources used in creating the entry and sources which are not (we already have this problem). Furthermore, the initial BP discussion that sparked this vote was created because for Belarusian, sometimes editors would put a given dictionary under "References", and sometimes the same dictionary would get put under "Further reading", which appeared inconsistent. Yet, according to this proposal, such placement would be 100% consistent if the editor who placed the dictionary under "References" used it in creating the entry, and the editor who put it under "Further reading" did not. However, for the average reader, this "meaning" will not be conveyed; instead it will look disorganized. Megathonic (talk) 04:03, 8 September 2023 (UTC)
 * Isn't it a good thing that it is not bot enforceable? References to construct entries are added by users so makes sense that the judgement is also make by users, not bots. I think it would be a very big mistake if we delegated to bots the enforcement of something that, so far, still requires human judgment. - Sarilho1 (talk) 15:36, 8 September 2023 (UTC)
 * We don't have enough humans, let alone willing humans, and least of all willing humans with good judgment willing to enforce rules that are not obviously a net benefit. DCDuring (talk) 21:07, 8 September 2023 (UTC)
 * 1)  for reasons I outlined on talk, and reasons given by Megathonic. It's not realistically maintainable: people already don't maintain any coherent distinction between the two headers we have now; in practice, adding a third header means references will be placed haphazardly under any of three headers instead of under any of two headers. I suppose from the perspective of supporters this is not a problem, because there will usually be plausible deniability by which they can think that the user who added their references to whatever header was present really did mean whatever ideal distinction between "I consulted and used this" vs "I didn't use this but am citing it anyway" is being envisioned, since the user's mind will usually be unknowable to anyone else. So then we can present haphazardness as meaningful (to people who have read there is supposed to be a distinction, while just looking as randomly disorganized as ever to other people, as Megathonic says). I would rather not. - -sche (discuss) 20:26, 8 September 2023 (UTC)
 * I think the wording in the vote was a bit rushed, and from the talkpage discussions it seems clear to me that "used in the creation of the entry" really just means "includes the headword or term and can be used for attestation". This is not something that should be difficult to maintain - either the source work you include actually mentions the term (and is not a wiki) or it doesn't, and based on that, you put it in either of the two headers. Thadh (talk) 21:47, 8 September 2023 (UTC)
 * No, that's not what it means; it means exactly what is written. If a source can be used for attestation, but it was not used in creating the entry itself, then it would go under further reading. I added this clarification based on what Al-Muqanna said.. This wording was unchallenged for 8 days prior to the vote starting. Megathonic (talk) 03:52, 9 September 2023 (UTC)
 * I think from five days later, which is also unchallenged, takes priority, wouldn't you say? Just because nobody replied to Al-Muqanna's message with "You should clarify that" doesn't mean that it is to be taken as a definite description of the vote as understood by the ones composing it. From my understanding, "supplementary material which is not used in creating the entry" means that it has to be 1) supplementary and 2) not used in creating the entry (so, supplementary material that is used, is added to references). It's not a description of the word "supplementary". Thadh (talk) 08:30, 9 September 2023 (UTC)
 * I meant that my wording in the vote proposal went unchallenged. I am the one who wrote "source not used for creating the entry" and "supplementary material which is not used in creating the entry" and stated my reason for adding this clarification in the edit summary. I wrote that precisely so that people wouldn't walk away from the vote with different interpretations as to what just happened. In retrospect I would have left the word "supplementary" out since it seems that word itself is causing confusion. Megathonic (talk) 13:13, 9 September 2023 (UTC)
 * 1)  Introduces confusion with Citations tab, where the meaning is "attestation". DCDuring (talk) 21:01, 8 September 2023 (UTC)
 * 2)  I don't really think adding new or different (sub-)headings is going to reduce confusion. We might as well stick to the current situation. — Sgconlaw (talk) 05:37, 10 September 2023 (UTC)
 * 3) per my vote below. — Soap — 16:59, 11 September 2023 (UTC)
 * 4)  Given these three choices and the fourth (the current regime), I am in support the current regime despite its flawed implementation and common misuse. I oppose this proposal's (Proposal #2) artifical distinction between a section of materials that was "used to create" an entry and one of sources that were not used to create the entry- this is so far from an objective standard that the current regime seems awesome by comparison. What if I re-read something in Further reading and it confirms something I added on the entry- is it now a References seciton entry? What if I re-read a References and change something that I added from it, such that now there's no close connection to the entry- is that material now moved to Further reading? Absurdity of absurdities. --Geographyinitiative (talk) 10:04, 17 September 2023 (UTC)
 * 5)  Ditto.  ·~   dictátor · mundꟾ  15:08, 21 September 2023 (UTC)
 * 6)  per my vote above, and per DCDuring. –Caoimhin ceallach (talk) 16:27, 27 September 2023 (UTC)
 * 7)  -- Huhu9001 (talk) 04:13, 28 September 2023 (UTC)
 * 8)  per DCDuring and my comments above in opposition to proposal 1. Voltaigne (talk) 19:10, 28 September 2023 (UTC)

Proposal 3: Rename the headers
Proposal 3: Two sections are maintained, one specifically for inline references and one specifically for sources and supplementary material that have not been cited inline. The distinction between non-inline "References" and "Further reading" is abolished.


 * 1) Merge   and   into one category called.
 * 2) Add the L3 header   above.
 * 3) Allow only references cited inline in the   section, handled by   or.
 * 4) Allow only a plain list of reference works in the   section.

Rationale: Such a system is easily enforceable by a bot, allows non-inline references to be marked as references, and removes the need for editors to make a judgement call between listing something under either "References" or "Further reading". Sources from which lexical information can be derived (e.g., a dictionary) would be categorized uniformly across all entries. This comes at the cost of no distinction being made between reference and non-reference works.

=== Citations ===      === References ===      *  [source used for creating the entry]   === Further reading ===    *  [source not used for creating the entry]   *

Support

 * 1)  100% bot enforceable. Keeps everything simple and consistent. I would have preferred a better name than "References" for the second section, but this is an improvement over what we have currently. Two big items often categorized as further reading are dictionaries and encyclopedias. Dictionaries convey lexical information and are still a general reference source. To a lesser extent, an encyclopedia is too, as the entry itself is an illustration of how the word is used. I'm not sure what other types of "further reading" would belong on this project; a general list of "further reading" is something for wikipedia. If this vote passes, "reflist" could be renamed to something like "cite-list" or whanot to keep the name consistent with the Citations section. Megathonic (talk) 04:03, 8 September 2023 (UTC)
 * 2)  Simple. Easy for humans to understand more or less correctly. The distinction between inline footnoted references and general references can me made visually, as it is now. A better look should be achievable by our CSS masters. DCDuring (talk) 21:03, 8 September 2023 (UTC)
 * 3)  without unnecessary stuff कालमैत्री (talk) 06:01, 2 October 2023 (UTC)

Oppose

 * 1) . On balance I don't think abolishing the further reading/citations distinction is helpful to readers. —Al-Muqanna المقنع (talk) 08:23, 8 September 2023 (UTC)
 * 2) . Removing the longstanding distinction because a few users are confused goes too far; if anything, we should be educating users more and making sure that it is abundantly clear. AG202 (talk) 08:35, 8 September 2023 (UTC)
 * 3) . We should not bring our human capabilities down to bots'. I don't see any reason to let them wipe out the distinction we have and value, as has already happened in Italian entries without consensus. Moreover, the headers' names would be very wrong: a Wikipedia link is not a "reference" (and if it is, it cannot, and should be substituted with the original source, as WMF projects aren't eligible resources). Catonif (talk) 14:08, 8 September 2023 (UTC)
 * 4) . I don't hold much love the "Further reading" section and I would be fine to see it go, but only if those links under it were removed. I cannot support categorizing those links (Wikipedia in particular) as references. - Sarilho1 (talk) 15:30, 8 September 2023 (UTC)
 * 5) . Thadh (talk) 19:58, 8 September 2023 (UTC)
 * 6)  I don't really think adding new or different (sub-)headings is going to reduce confusion. We might as well stick to the current situation. — Sgconlaw (talk) 05:37, 10 September 2023 (UTC)
 * 7)  There is a bot which places "Further reading" below "References" (God knows why). I don't agree with that. No to ===Citations===, which can be entered at the top of each entry. There are still some ====Quotations==== headers with content which needs rehoming. DonnanZ (talk) 10:04, 11 September 2023 (UTC)
 * 8) based on naming. The   tag is hard-coded into MediaWiki. We could create a new template that renames it, but people are used to using   and that'll be a difficult habit to break; even with a new template, people might not recognize it for what it is and will try to add it.  Also I think having a header for citations is confusing when we already have a Citations namespace which is more in line with what the word citation usually means on this project.  I think the status quo is fine. All the best, — Soap — 16:58, 11 September 2023 (UTC)
 * 9)  Vininn126 (talk) 22:28, 11 September 2023 (UTC)
 * 10)  per Catonif and Soap. lattermint (talk) 17:34, 13 September 2023 (UTC)
 * 11)  per AG202 and Catonif. It is important to maintain the distinction. – wpi (talk) 13:41, 21 September 2023 (UTC)
 * 12)  Ditto.  ·~   dictátor · mundꟾ  15:08, 21 September 2023 (UTC)
 * 13)  Ioaxxere (talk) 03:49, 22 September 2023 (UTC)
 * 14)  -- Huhu9001 (talk) 04:13, 28 September 2023 (UTC)
 * 15)  per my comment above in opposition to proposal 1, a better approach IMO would be to somehow align the indentation of inline and non-inline ref lists, rather than the more drastic course of renaming sections and thus potentially creating more confusion. Voltaigne (talk) 19:17, 28 September 2023 (UTC)
 * 16)  Support proposal 2. Netizen3102 (talk) 17:47, 29 September 2023 (UTC)
 * DonnanZ (talk) 20:04, 3 October 2023 (UTC)
 * @Donnanz You've opposed this twice now. —Caoimhin ceallach (talk) 12:15, 4 October 2023 (UTC)
 * Striking double vote. AG202 (talk) 18:56, 4 October 2023 (UTC)
 * OK. DonnanZ (talk) 20:08, 4 October 2023 (UTC)

Abstain

 * 1) I wouldn't object to this, but also wouldn't object to enforcing this type of distinction while keeping the current headers. Perhaps what would resolve the competing concerns of (a) not having sister-wiki links presented as 'references', and (b) being maintainable without having to give some of our users The Boys' Compound V so they can read every random IP's mind to know how deeply that user consulted a reference and hence which header it belongs under, is: put (1) inline references under one header, (2) non-inline, non-wiki references under another header, and (3) links to other wiki projects under a third header. That would be bot-maintainable. - -sche (discuss) 20:44, 8 September 2023 (UTC)
 * The issue here is that there are also some links that can be added as further readings but not be used for attestation - actual encyclopedias that are not wikis and (particularly for reconstructed terms) descriptions of the principles of reconstruction and/or soundlaws relevant to the entry but not mentioning the reconstructed term itself. Thát is the main problem I have with this proposal. Thadh (talk) 21:50, 8 September 2023 (UTC)
 * I'm not sure I follow what you mean. Aren't reconstructed terms by definition not attested? It's not necessary for the encyclopedia (or book, paper, etc.) to specifically mention the term itself in order for it to serve as a reference: The sound laws/reconstruction principles would themselves be a reference that was used for reconstructing the term. In this case, wouldn't it actually be incorrect to list it under further reading? Megathonic (talk) 03:22, 9 September 2023 (UTC)
 * A reconstructed term is not attested, but a reconstruction is (it is reconstructed elsewhere).
 * And I don't think that reconstruction principles that are explained in a paper serve as proof of the reconstruction, and thus should be included under the "reference" header - since essentially that makes it extremely difficult for anyone to figure out if we're just pulling the reconstruction out of our [name your favourite body part], or if it is actually a solid reconstruction used elsewhere. There is nothing wrong with original research, but readers should be aware that we are doing it.
 * Compare : To anyone familiar with the Permic languages, it is clear that this term should go back to PP, since it is present in every branch of Permic, and is completely regularly formed. However, nobody has reconstructed it yet. If I were to add the etymological dictionary to "references" though, every reader would have to check with the reference whether or not this is something we have reconstructed ourselves or if it has already been reconstructed by someone else. Thadh (talk) 08:38, 9 September 2023 (UTC)
 * @-sche Is this intended to be an abstain vote? AG202 (talk) 21:15, 13 September 2023 (UTC)
 * Functionally, yes; it's more of a comment than a vote, but since it's neither a vote of support nor a vote of oppose, I put it in the abstain section. My actual vote/preference would be to implement the bot-enforceable distinction described by this proposal 3, but while using the current header names (or at least, continuing to call the header that &lt;references/&gt; and reflist go under "References", and maybe renaming "Further reading" if people think it sounds too much like "Not references" to put references under). - -sche (discuss) 21:44, 13 September 2023 (UTC)
 * , due to not really knowing how frequently/effectively the "Further reading" heading is used today. Chernorizets (talk) 09:55, 17 September 2023 (UTC)
 * 1) . I am sympathetic to any of the three in principle, but I'm not wild about any of the headings. As Soap notes above and DCDuring points out in Proposal 2, a "Citations" heading and a separate "Citations" page with different functions is ambiguous and potentially confusing. Proposal 1 avoids that ambiguity, but I agree with Megathonic and others that using three headings seems more complicated than is necessary. Cnilep (talk) 06:48, 19 September 2023 (UTC)
 * 2)  Mcph2 (talk) 06:12, 7 October 2023 (UTC)

Decision

 * Proposal 1 - Add a new subheader: Failed 5-15-2.
 * Proposal 2 - Add a new header: No Consensus 12-10-0.
 * Proposal 3 - Rename the headers: Failed 3-16-4

Kutchkutch (talk) 00:00, 8 October 2023 (UTC)