Wiktionary talk:Votes/2020-07/Removing letter entries except Translingual

Letter entries to be Translingual
FYI Votes/2020-03/Letter entries to be Translingual exists, though it doesn't appear to have started --DannyS712 (talk) 20:30, 1 August 2020 (UTC)
 * That's the 'unused vote' that Μετάknowledge mentioned at the start of the referenced discussion! --RichardW57 (talk) 21:12, 1 August 2020 (UTC)
 * Yeah, just wanted to make sure it was highlighted here --DannyS712 (talk) 22:15, 1 August 2020 (UTC)

Keeping Derivatives
How do I vote to keep derivatives that are not spelt the same way as the letter and are not excluded by other rules. For example, in English x's is clearly a possessive (excluded), plural (not sure of the Wiktionary rule here) or contraction with is or has, and excluded by convention. On the other hand, Pali can be the bare stem or the vocative singular of the word meaning "short", or it is the genitive/dative singular of the Pali letter name, probably to be excluded as it coincides with translingual. I want us to be able to have an entry so that someone looking up will be advised that it could be the genitive singular of the letter - especially as this fact is missing from most grammars and most dictionaries. --RichardW57 (talk) 11:27, 5 August 2020 (UTC)

Scope of User Guide

 * I think it is asking too much for the user guide to necessarily give either the spelling or the reading rules for a language. I started by thinking about giving the spelling rules for Pali, and then realised that I may not know all the Tai Tham and Lao script rules, and remembered the unresolved arguments over the abugidic Thai script spelling of Pali.  And even for reading rules, who's going to volunteer for English?  English reading rules need a book.  Presumably Japanese escapes by not being alphabetic. --RichardW57 (talk) 14:42, 10 August 2020 (UTC)

I've therefore changed the Option 3 to not require the giving of spelling or reading rules. --RichardW57 (talk) 14:40, 10 August 2020 (UTC)
 * To be honest, I don't understand the problem. The English user guide would have 26 letter entries, all of which would include how the letter is pronounced when spelled. ( being /eɪ̯/ and  being /biː/ etc.), that doesn't seem like a lot of work to me. Thadh (talk) 14:58, 10 August 2020 (UTC)
 * The text read, "Create a user guide per language, containing both the spelling (pronunciation) rules and why some entries are preferred in or excluded from the dictionary...". It doesn't say give the pronunciation of the name of the letter.  For example,  has the following entry for the pronunciation of the letter: --RichardW57 (talk) 18:03, 10 August 2020 (UTC)


 * That's a very first step at explaining how the letter is pronounced in words in the Portuguese language. Unsurprisingly, there's no attempt to give the pronunciations of the letter 's' in Portuguese words. (If my memory serves me right, it has 4 different pronunciations.)  By 'spelling rules', I would expect at least a guide to working out the pronunciation of a word from the spelling, though I wouldn't complain if I got a useful explanation of how to spell words instead.--RichardW57 (talk) 18:03, 10 August 2020 (UTC)
 * As to 26 letters, you're overlooking the English letter . At present, its name (and therefore the pronunciation of the name) is a tad hidden under .  Explicitly giving the names of all the letters we describe as 'English' is more than we do now!  You'll find a few more at Category:English_letters - and it still doesn't cover words like .--RichardW57 (talk) 18:03, 10 August 2020 (UTC)
 * By "containing the spelling (pronunciation) rules" I meant "containing the spelling and spelling pronunciation rules". I'm sorry if that caused a misunderstanding. Furthermore, I am not sure whether the lemma is even entry-worthy under the current CFI, and if it is, that is exactly why we should have this vote. I could borrow the surname  into English and thus say that "The letters 'ą' and 'ł' are English letters chiefly in borrowed terms", which is obviously crazy. sidenote: the entries for these letters are part of the original problem Thadh (talk) 18:32, 10 August 2020 (UTC)
 * : I now think you meant "containing the spelling out rules". Of course, they aren't always known - I don't think we even know the names of the Ugaritic letters.  Keyboards have had a dramatic effects on some systems - the Thai system I learnt seems to have been replaced by key by key instructions!  We can dictate the home of the information - we can't compel its recording. --RichardW57 (talk) 00:22, 11 August 2020 (UTC)
 * : I now think you meant "containing the spelling out rules". Of course, they aren't always known - I don't think we even know the names of the Ugaritic letters.  Keyboards have had a dramatic effects on some systems - the Thai system I learnt seems to have been replaced by key by key instructions!  We can dictate the home of the information - we can't compel its recording. --RichardW57 (talk) 00:22, 11 August 2020 (UTC)

Rare Letters and the original problem

 * Borrowing the surname Bącławek into English unchanged is not easy. Your adopting the Polish spelling into English would not make it an English word - you need others to do it too.  Borrowinɡ Ostojić typically results in the accent being lost, and Wałęsa (remember him?) was rapidly naturalised to Walesa.  On the other hand, there is a significant social difference in the establishments termed café and cafe in British Enɡlish, which Wiktionary totally fails to capture.  I for one find it very uncomfortable to write fiance for fiancé.  Indeed, the Firefox spellchecker offers to correct both unaccented words to having e acute. --RichardW57 (talk) 09:42, 11 August 2020 (UTC)


 * That is partly true, but at the other hand, Wiktionary does have the English entry ... Furthermore, although the United States do handle the politics of not adding any diacritic marks to names, most countries don't. Thadh (talk) 12:27, 11 August 2020 (UTC)

How is any more a part of the problem than ? As far as I understand it, the problem is not the number of entries in Wikipedia, but the number of entries in a page. --RichardW57 (talk) 09:42, 11 August 2020 (UTC)


 * It isn't, but it is a part of the problem. The point is, Wiktionary having (almost) duplicate entries on one page brings many problems in the long term, as much as in the short term if you look at a|ve|oi|wo|eu|l entries, and as a regular user of Wiktionary I personally can vouch that the moment I want to find the definition of the Turkish lemma and I get Lua error because of the many identical letter entries that precede it, I start doubting my choice of this dictionary. Thadh (talk) 12:22, 11 August 2020 (UTC)
 * As an experiment, I cleaned the letters out from o (except for translingual), with untidy result in User:RichardW57/o. That got the memory usage down to 46.7M compared to the 50M limit. Even with the letters gone, the vowel pages will still be heading for trouble.  (I'm not sure how reliable the figures are - I got weird changes in Lua memory usage as I restored or removed bits of the page.) --RichardW57 (talk) 15:21, 11 August 2020 (UTC)
 * That does look good and tidy ^_^. But I understand that that doesn't necessarily fix the problem of Lua (see also si, which is overflowing without being a letter), but it does firstly help and secondly it doesn't seem to bring many disadvantages with it. Although with the case of o, another problem is the vocative particle that is just borrowed in almost every language. Thadh (talk) 15:32, 11 August 2020 (UTC)
 * I did wonder about making the vocative particle translingual! However, there do seem to be a fair few language-specific nuances to it.  At least we don't seem to have an entry for white space - widely borrowed from Irish Latin. --RichardW57 (talk) 16:01, 11 August 2020 (UTC)

Nouns
“Entries of nouns that stand for letters”. – so what would we delete/move away from the mainspace? Fay Freak (talk) 19:20, 11 August 2020 (UTC) It seems like the noun deletions are impracticable, unless great offence isn’t shunned. Fay Freak (talk) 19:20, 11 August 2020 (UTC)
 * (probably not)
 * (because this is “not a real word”, or less of a real word? Although both are equally used as nouns)
 * ? Probably not, but would have to be deleted/not created if  were to be deleted?
 * ? Probably not, but would have to be deleted/not created if  were to be deleted?
 * ? Probably not, but would have to be deleted/not created if  were to be deleted?
 * As I have already said somewhere, I proposed it in the spirit of deleting such nouns as, , etc. These are nouns that do not differ enough from the letter itself to form full names (as opposed to ,  or , .)  On the other hand, I do object to , for it conveys only the meaning of the symbol, the name itself would have been written als.
 * So to sum up,, , i griega and Eszett or (probably, if it isn't SOP) scharfes S are safe in my opinion; But I have not asked for help out of boredom, I wasn't sure if everyone had the same idea, and the vote should include all proposed options. Thadh (talk) 19:37, 11 August 2020 (UTC)
 * I've rewritten the vote to capture earlier discussions. I came to the conclusion that irregular inflections should be kept.  If you want the option of deleting aes, you will need a fourth removal option. --RichardW57 (talk) 20:19, 11 August 2020 (UTC)
 * , I haven't contributed much to this iteration, but I strongly oppose the removal of and similar noun entries. Those are lexical, can host etymology and pronunciation sections, and would probably be something most of the community would vote to keep. —Μετάknowledge discuss/deeds 20:37, 11 August 2020 (UTC)
 * I would also not choose the combination of 1A, but for example 3A wouldn't sound too strange in my opinion. But once again, I am trying to find some combination of options that would be able to suit at least the majority of the community. Thadh (talk) 20:46, 11 August 2020 (UTC)
 * I think we have a fundamentally different view of what a vote should accomplish. This vote tries to find out what the community position is from a range of options, but it is likely to fail to produce a consensus because of how many camps it divides voters into. In my opinion, that is better suited to discussions. A vote should be aimed at getting the community to settle on something in stone, maximising the chance of passing by presenting something that garnered wide support in discussions. —Μετάknowledge discuss/deeds 20:59, 11 August 2020 (UTC)
 * The problem is that we don't have enough discussions to determine what the community wants. What we know is that the community wants to get rid of Lua errors and it wants to make the dictionary useable. And that is all I have understood from all this, as we have only three to four people actively discussing this. If you think that the option of deleting "bee" is against the community's wishes, you have my wholehearted blessing to delete that option and reduce the amount of possibilities (especially given the fact that you know this community much longer and better than I do). However, when writing the vote, I did not have that certainty, and thus I decided upon adding too many options rather than too little. Thadh (talk) 21:31, 11 August 2020 (UTC)
 * , the problem is that too many options tends to slow down progress. But it sounds like your goal is to avoid memory errors, and in that case, you're barking up the wrong tree; this would help with that, but its main thrust should be streamlining the dictionary. After all, 一 and 水 aren't written with letters at all. As I see it, there are only two ways to solve the memory problem for good: 1. Beg the devs to lift the arbitrary limit (this has been tried and they don't seem to care about Wiktionary). 2. Split all pages by language, so cano/la, cano/it, cano/es, etc (this has been proposed many times, but I don't think it has ever been put to a vote; it would require a massive, fundamental change to all of Wiktionary). —Μετάknowledge discuss/deeds 22:09, 11 August 2020 (UTC)
 * Actually my personal motivation is that I don't see the point of duplicates and they make using this dictionary a pain. However, as with the eye dialect entries, I completely understand if others don't feel that way, and so to make the picture complete and not only about myself, I came up with also the more universally positive side of getting rid of these duplicates, namely memory. I think this is a good step towards solving that problem as well Thadh (talk) 08:19, 12 August 2020 (UTC)
 * The problem is that it is hard to define something that the community or you want that is still effective in decluttering the pages. If we remove the letter headers but put noun headers on every vowel character (as that is how the name of a letter will be written in the case of a vowel) little is gained. Whereas you fail to define why a noun entry would need to be deleted but not . I think they try with examples (“highly akin to (e.g. bee and be but not )”) is a bit better; though we would still like to have better vocabulary to describe the particularities of the elaboratenesses of letter names: What’s the qualitative difference between  and ? Basically we can imagine what many want but to write rules is a different story. Fay Freak (talk) 21:42, 11 August 2020 (UTC)
 * The qualitative difference between and  is that the first is generated by rule.  However, eliminating the letter name 'bee' gains us nothing.  Nor does eliminating the almost translingual letter name  - it seems to be nothing like as common as one might expect.  That's why I would vote to remove only 'b' on the basis of the letter.  The problem with 'b' is that many languages have it in some fashion.  Once the spelling deviates, the number drops drastically.  Likewise, there's no need to eliminate Esperanto . --RichardW57 (talk) 22:34, 11 August 2020 (UTC)
 * The other issue is that many instances of 'b' almost all meaning the letter is rather tedious. Otherwise, I would suggest replacing some of the module calls with templates.  It seems to me that most of the letter-related module invocations can be replaced by simple templates.  I'm assuming that  is cheap for ASCII lemmas.  Maybe the module invocation needs to be replaced by templates.  That's the bad news we keep getting - Lua modules may be more powerful and more fun, but our resource constraints scream out that we should use template encoding.
 * I would like to point out that the fact that such entries as be or ce aren't "as common as one might espect" might be due to the reason that we don't yet have every entry of every language and that most editors don't bother adding these entries due to their understanding that it is practically useless. But, unless we settle this vote, in the future they surely will be Thadh (talk) 08:19, 12 August 2020 (UTC)
 * The reason is that the spelling has fractured. Quite a few languages have bé instead.  English has similarly reshaped the spelling of stop consonants' names; Esperanto has reshaped them to vaguely fit its morphology.  These changes push them down to the level of widespread words like auto and zebra.   --RichardW57 (talk) 10:09, 12 August 2020 (UTC)

Why translingual
Also, why doesn’t the vote have the option to keep nothing of all that in the mainspace? Why “keep only Translingual”? We save more by not even keeping translingual. Especially with option 4 it makes sense to keep no translingual, because the editors can just link the appendix per letter. Fay Freak (talk) 19:20, 11 August 2020 (UTC)
 * Because if we delete translingual, that would be deleting the whole concept of, which I thought was a step too far for such a vote. However, if you think it is wise to settle it here, too, I don't oppose to adding the options of deleting translingual entries as well. Thadh (talk) 19:41, 11 August 2020 (UTC)
 * We will still have translingual for non-letters; it seems odd to make letters an exception. Additionally, not all characters are always letters - is pi a symbol or a letter?  Add an Option 4c if you want to be rid of translingual letters. --RichardW57 (talk) 20:19, 11 August 2020 (UTC)
 * Actually, for Latin letters, we'll just separate the letter and the IPA symbol. We'll probably then see the IPA symbol sprout an etymology link to the translingual letter. --RichardW57 (talk) 21:49, 11 August 2020 (UTC)

What’s 4b
And I don’t understand 4b “Link this at the beginning of the page” – which page or link what? A link whence whither? Fay Freak (talk) 19:20, 11 August 2020 (UTC)
 * Let's take the page a. At the beginning of the page, there would be:
 * Now, with 4a on the other hand, it would look something more like this:
 * Thadh (talk) 19:51, 11 August 2020 (UTC)
 * Why is that 4a? And if this is a typo and it is 4b then it is not “from the beginning of the page” as written in the vote option (which page? beginning of each language section?). Fay Freak (talk) 20:11, 11 August 2020 (UTC)
 * Oh right, let me fix this, I was confused again by the many options. Thadh (talk) 20:14, 11 August 2020 (UTC)
 * I've been thoroughly bold and rewritten the options. I've said to link to the appendix.  I've left it implicit that the link would be in the page of the letter.  I presume the link would immediately follow the  if any at the start of the page. --RichardW57 (talk) 20:19, 11 August 2020 (UTC)
 * What do we derive X-ray from? At present we derive it from the English X.  I would rather have Appendix:Language specifics of "A".  However, I think Option 4 is a bad idea - we move the problems from the letters to these appendices. --RichardW57 (talk) 20:28, 11 August 2020 (UTC)
 * I am not sure that we do; By moving them we create two logical pages (one for duplicates and one for "normal" entries), and as such we can restructure the Appendix to not taking up as much memory as the Main pages do. Now, I don't know Lua, neither do I know how precisely to make this happen (that would be up to our technical geniuses), but I do believe creating an appendix would make things easier. Not to mention the fact that it buys us time.
 * What are you getting rid of? You'll still get the position in the alphabet and list of letters in the alphabet for each language, and the occasional inflection template.  You may remove a heading level or two per language, but they and the table of contents are not processed by Lua. --RichardW57 (talk) 21:00, 11 August 2020 (UTC)
 * The  does, though. And by the way, if we have the Appendix another possible step would be trying to get rid of the Category: LANG letters. But, once again, as the memory restriction we have is per page (or so I have understood), moving over half the page would buy us time to figure out what to do with the Lua in Appendices or words like si. Thadh (talk) 21:38, 11 August 2020 (UTC)
 * As for X-ray, I think it would be quite easy to create an Etymology-only template for words derived from letters in a specific language. Thadh (talk) 20:37, 11 August 2020 (UTC)
 * I think the current templates would work. The naming problem is where the 'Derived Terms' section goes.  It's neither the pronunciation (in any sense) nor an inflection of 'X'. --RichardW57 (talk) 21:00, 11 August 2020 (UTC)
 * I think the “X” in X-ray does not belong to any language but was intended as a translingual variable, which does not mean there ever was dictionary-fit content imagined under this letter (as it could rather be an encyclopedic concept).
 * The  does, though. And by the way, if we have the Appendix another possible step would be trying to get rid of the Category: LANG letters. But, once again, as the memory restriction we have is per page (or so I have understood), moving over half the page would buy us time to figure out what to do with the Lua in Appendices or words like si. Thadh (talk) 21:38, 11 August 2020 (UTC)
 * As for X-ray, I think it would be quite easy to create an Etymology-only template for words derived from letters in a specific language. Thadh (talk) 20:37, 11 August 2020 (UTC)
 * I think the current templates would work. The naming problem is where the 'Derived Terms' section goes.  It's neither the pronunciation (in any sense) nor an inflection of 'X'. --RichardW57 (talk) 21:00, 11 August 2020 (UTC)
 * I think the “X” in X-ray does not belong to any language but was intended as a translingual variable, which does not mean there ever was dictionary-fit content imagined under this letter (as it could rather be an encyclopedic concept).

Fay Freak (talk) 21:25, 11 August 2020 (UTC)
 * A-frame might be a better example - not every 'A' is the right shape! Anyway, most of the lemma-referencing templates seem to have an argument for the link and for the display, so we should be able to link to section titles in an appendix and give a suitable display, as in something like  yielding  --RichardW57 (talk) 21:41, 11 August 2020 (UTC)

Voting procedure
Can some one please review the voting procedure. I'm particularly bothered by the mechanism for counting the vote, but I'm not sure whether we have anything resembling STV. I can also conceive needing a run off if one could argue that different schemes had sufficient support. --RichardW57 (talk) 20:49, 11 August 2020 (UTC)

Split
Seeing the far-too-great amount of options and disagreements, I propose we split this vote in two: Of course, the question of where the Appendix (or User Guide) should be linked (and what it should include) doesn't disappear; it should either be sorted out before the vote or by the vote (as a suboption). Thadh (talk) 17:17, 21 August 2020 (UTC)
 * 1) (This one) Removing all letter entries: so all entries that are preceded by the H2-header   and defined by the tag  (or a derivative).
 * 2) Remove
 * 3) Move to Appendix
 * 4) Not Remove
 * 5) (Another) Removing all noun entries: so all entries that are preceded by the H2-head , defined by the tag  (or a derivative) and being eligible as a doublet of the letter.
 * 6) Regarding all letter-meaning
 * a) remove
 * b) move to Appendix
 * c) not remove
 * 1) Regarding all letter-influenced
 * a) remove
 * b) move to Appendix
 * c) not remove
 * 1) Regarding all identical to letters
 * a) remove
 * b) move to Appendix
 * c) not remove
 * Done. I propose that whether we go through with the second vote depends on the result of this one. Thadh (talk) 22:00, 10 December 2020 (UTC)

Very complicated
Just a note on understandability. Perhaps this is inherently a complex decision, or set of decisions, that cannot be made simple, but I am struggling to understand these options. Mihia (talk) 23:42, 5 December 2020 (UTC)

Appendix question
This vote looks good to me - omitting the portion about removing nouns for letter names has made it much more understandable. I think I'm still a little confused on what exactly happens with Option 2, though. Would the headers for each applicable language still exist, but now refer to an appendix instead? Or would there just be some link at the top of the page like "Appendix: Alphabets that use the letter X". I would support the latter, and suggest make the planned actions a little more explicit in the wording of Option 2. Looks good though! Hope this doesn't fizzle out! – Guitarmankev1 (talk) 13:05, 24 June 2021 (UTC)
 * What I had in mind was more like "Appendix:Foobarian alphabet" (not linked in mainspace) with the appropriate letters in alphabetic order and possibly names and pronunciations. In that case the actions on constructing said appendices would be left to the languages' respective communities. I don't think the appendices concerning the specific letters would be handy, since that would mean about three thousand entries on said appendix for the letter "a", which would possibly break the appendix' memory, but I'd love to hear what you think about it! Thadh (talk) 13:34, 24 June 2021 (UTC)
 * I like the idea of having separate appendices for each alphabet, maybe something like Appendix:Greek_script, but pertaining to individual alphabets/languages rather than greater scripts - which would include letter forms, letter names, pronunciation, etc. But if I land of the page for Š, then ideally I'd still want to see some list of alphabets/languages which include that letter. I agree that listing these languages directly on the namespace could get messy (the page for A currently lists 225 non-translingual languages) so maybe there could just be a link to "Appendix: List of alphabets which include Š" or something would be a good solution. But I should be able to navigate my way to that information from the letter page somehow. – Guitarmankev1 (talk) 14:02, 24 June 2021 (UTC)
 * Maybe we could combine these: Make an "Appendix:List of languages using the letter X" which then provides a list of appendices of the specific languages, in alphabetic (+English) order, like "Appendix:List of English letters#X". I'm wondering also what others think, ? Thadh (talk) 14:40, 24 June 2021 (UTC)


 * That sounds like a great idea! – Guitarmankev1 (talk) 17:12, 24 June 2021 (UTC)
 * It looks good. I think we can handle large lists of redirects, but if we can't, we can always use categories. --RichardW57 (talk) 19:33, 25 June 2021 (UTC)

Sign languages
For the sake of being as formal as possible, is it okay with you, since no-one else has voted yet, if I add a small paragraph couple of words about excluding sign languages for this vote? I don't think it's a good idea to include them because they are quite different from "traditional" letters, and since nobody from the sign language communities has weighed in, I think we ought to just let them be for the time being, at least until someone can say anything reasonable on the subject. Thadh (talk) 15:10, 13 October 2021 (UTC)


 * Yes, that's fine with me, and I will strike that portion from my oppose comment. AG202 (talk) 15:14, 13 October 2021 (UTC)
 * I won't oppose this change. &mdash; surjection &lang;??&rang; 15:25, 13 October 2021 (UTC)
 * Done, thanks for the quick responses! Thadh (talk) 15:31, 13 October 2021 (UTC)

Terms in various languages derived from their languages' letter entries
If we go ahead with removing all the language-specific letter entries, what are we going to do with the non-translingual derived terms of said language-specific letter entries (which currently go in the "Derived terms" subsections of their respective language sections)? The derived terms can't follow the letter entries to the translingual sections, as the derived terms themselves aren't translingual; are we going to simply delete the derived-terms listings for these letters? Whoop whoop pull up Bitching Betty ⚧ Averted crashes 03:02, 17 October 2021 (UTC)
 * X-ray is a derivation of Translingual X and the English ray. Wouldn't be the first time a language borrowed only a part of the word from another language. The derived terms will in that case be listed as descendants at the Translingual entry. We could maybe think about classifying borrowings from Translingual as just derivations from the target language, but I don't know if that executable. Thadh (talk) 09:24, 17 October 2021 (UTC)

Implementation
Since the vote is likely to pass, is it possible that entries could not be deleted until that specific community has made their own Appendix page? (Maybe before a specific deadline?) I'm very wary of mass bot-deletions and the loss of information if it's done too hastily, as some implementations have been done in the past. AG202 (talk) 22:42, 30 October 2021 (UTC)


 * I think that's reasonable, but I wouldn't know what a good deadline is for that. I wasn't expecting any hasty implementation anyway, we've had this discussion for years now and we're only now having an actual vote on the issue. Thadh (talk) 12:22, 31 October 2021 (UTC)


 * We no what we don't want, so we just ban that and then sort it out. It's simple, like the UK leaving the EU. --RichardW57 (talk) 13:15, 31 October 2021 (UTC)


 * What's a community? Seriously, what to stop any user just creating a quick index page and copying the letters to subpages of that index page and putting up a notice on the old entries asking that they not be edited any further.  For good measure, he could just copy the first four letters and then get bored. --RichardW57 (talk) 13:15, 31 October 2021 (UTC)


 * What's to stop you from adding fake Pali words? Nobody is checking your contributions any more, after all. The answer, of course, is that you want to help the dictionary, not harm it. We all want that, which makes this idle fearmongering pointless. —Μετάknowledge discuss/deeds 19:08, 31 October 2021 (UTC)


 * Adding fake Pali words is clearly not helpful; a simple move as I suggested for letters can be seen as helpful. (Also, I suspect my Pali entries are being monitored with a view to inserting a Sanskrit etymology.)  It also helps issues such as marginal letters.  Thinking of Welsh, 'j' as in Jones was not reckoned a letter of the Welsh alphabet when I was a lad.  It seems that 'v' still isn't, even though "Evans is Welsh for Jones".  I always find accented vowels being linked to as Welsh letters, even though there are currently no such entries. --RichardW57 (talk) 22:35, 31 October 2021 (UTC)
 * @RichardW57 Is there a reason you haven't voted yet (in either direction)? I've noticed that you've been active in discussion about the vote but haven't voted directly yet. AG202 (talk) 20:25, 1 November 2021 (UTC)
 * I planned to complete a trial implementation for Welsh (results at User:RichardW57/Appendix:Welsh Letters) before voting, but the mopping up was getting tedious. One thing that became clear was the inconsistencies that are obscured by scattering the information over dozens of pages.  If  is honest, he has underestimated the work to be done in gathering the data, but its sanely doable.  (I seriously scared myself by the Brexit analogy.) However, I am bothered about the concept of 'community', which sounds a good idea, but seems too informal for formal operation.  Is it just an editor and his mates in discussion on his user talk page?  One might once have hoped that  and its tooling would have enabled self-enrolling communities, but it seems not to have happened.  The big question is, how many languages use just the letter as the name of the letter.  I fear it's lots, so we may still be left with a lot of tedium in the one letter pages.  To yield dividends, we probably need to try to keep the entries terse, while doing justice to things such as A-frames.  We could probably cut things down a lot by defining k not as the eleventh letter of some alphabet, but giving, as we are supposed to, its translation, i.e. 'k'. --RichardW57 (talk) 21:02, 1 November 2021 (UTC)
 * I definitely agree that it'll be quite the undertaking. I'm not sure if you're in the Discord, but it was discussed there further, about how there's a much larger issue that needs to be fixed. You're right that a lot of languages have a letter name for that specific letter, especially for the vowels a, e, and so on. And so, the problem will keep continuing on, which is why I wanted to shift towards focusing on an actual solution towards the problem at hand. However, this vote is also important to how things will be done in the future, so I'd encourage participation. AG202 (talk) 21:13, 1 November 2021 (UTC)

Proposal for keeping letter entries in mainspace while still solving the oversize-pages problem (copied/paraphrased from comments on vote)
During the voting,, as part of backing up their oppose vote, brought up an important point regarding the technical problem that we're trying to solve:

"And before anyone mentions Lua Memory Errors, this is not the solution for that, if you look at many of the entries like o or e, most of the entries are in fact not letters, but other POSs, and then, mi, for example, that has no letters has the same issue. Thus, deleting letters & moving them to translingual is not a solution, but a wide-sweeping move that wouldn’t fix that issue at hand to begin with, a short-term solution to a long-term problem." (AG202, emphasis added)

In discussion with regarding the point brought up by AG202, I came up with an alternate proposal (sorry if it looks like I'm tooting my own horn a bit here), and brought up another point that would further reduce the benefit of nuking the letter entries from mainspace:

"Removing the non-translingual letter entries would be of minimal help with the OoM errors; as detailed in AG202's oppose vote below, the non-translingual letter entries make up only a small fraction of the total sizes of the letter pages (and this becomes an even smaller fraction if, as suggested on the talk page in response to my concerns regarding the derived-terms sections for the non-translingual letter entries, these sections are kept on the pages and moved to the translingual letter entries rather than being excised as part of the non-translingual letter entries), and many of the pages running into these errors aren't letter pages at all. My preferred option for dealing with the Lua OoM errors would be to split huge pages into several smaller pages (like we already do - for example - with the Appendix:Unicode pages for blocks with thousands of character assignments)." (me, emphasis added; the proposal to move language-specific derived-terms sections to the translingual sections upon removal of the non-translingual entries from mainspace is on this page, two sections up)

This could be implemented by leaving the translingual and English entries on the main letter pages, and moving the rest to subpages such as (for instance) A/Languages: A-F, A/Languages: G-M, A/Languages: N-S, A/Languages: T-Z.

I realize that this is probably a bit late to be bringing this up, and my sincerest apologies if I'm out-of-bounds with doing this, but I figured that, given that the main impetus for nuking the non-translingual letter entries from mainspace is to keep the letter pages small enough to keep Lua and lower-powered web browsers from choking on them, and given that (as per the evidence provided by @AG202) the situation currently proposed wouldn't actually solve the problem at hand, this was probably a big enough issue to warrant rethinking the proposal. (Again, feel free to trout me if I'm breaking the rules by doing this!) Whoop whoop pull up Bitching Betty ⚧ Averted crashes 04:08, 31 October 2021 (UTC)


 * 100 real words is still more manageable than 100 letter entries + 100 real words. —Suzukaze-c (talk) 04:21, 31 October 2021 (UTC)
 * Also, we know it doesn't solve the memory problem. Go read WT:Lua memory errors if you want more on that topic. —Μετάknowledge discuss/deeds 06:02, 31 October 2021 (UTC)