Wiktionary:Votes/bt-2018-10/User:Robin Bot for bot status

User:Robin Bot for bot status
Nomination: I hereby request the Bot flag for User:Robin Bot for the following purpose:
 * Creating pages for alternative forms of Esperanto words. Esperanto has 6 special letters (ĉ, ĝ, ĥ, ĵ, ŝ, and ŭ) which can be represented in 3 different ways: with Unicode, with the H-system and with the X-system. With the help of pagefromfile.py I want to create all missing pages. (test run)

Schedule:
 * Vote starts: 14:58, 22 October 2018 (UTC)
 * Vote ends: 23:59, 5 November 2018 (UTC)
 * Vote created: 14:58, 22 October 2018 (UTC)

Discussion:
 * [[Image:Wikt rei-artur3.svg|20px]] Wiktionary talk:Votes/bt-2018-10/User:Robin Bot for bot status

Support

 * 1) . This is badly needed. Wiktionary currently is less convenient than other vortaroj because they automatically resolve at least the x system, while we do not. It is very difficult to search for Esperanto terms with special characters; the only option, without switching keyboard layouts, is to enter them without special characters and click one of the "see also" links at the top of the page. Also, regarding whether to have entries for x- and h-system entries at all, this is currently inconsistent; some common words have them as alternate forms, but most don't. Any resolution of this would be good for Wiktionary. One more thing for the proposer: looking at the test run, I'd be careful of archaic words with ĥ - adding alternate forms for these would probably require adding similar alternate forms for the equivalents with k. If the bot can't do this, maybe those should be left out of the set. ARR8 (talk) 02:48, 31 October 2018 (UTC)
 * 2)  If the orthographies are equivalent for all intents and purposes, distribution of their attestation is perfectly meaningless and is not worth reproducing. Crom daba (talk) 12:28, 31 October 2018 (UTC)
 * 3)  per above. —Suzukaze-c◇◇ 00:34, 1 November 2018 (UTC)
 * 4)  per above (and per the comment made in the "Abstain" section). Andrew Sheedy (talk) 01:30, 1 November 2018 (UTC)
 * 5)  per above. J3133 (talk) 17:14, 3 November 2018 (UTC)

Oppose

 * 1) . I support the creation of these alternative spelling entries only when that spelling is attested, just like in English. This bot would create entries en masse without checking whether they meet CFI. (I would be open to an alternate proposal where the bot uses corpus results to determine whether a spelling is attested, but that would require writing specific code to implement that.) —Μετάknowledge discuss/deeds 23:19, 24 October 2018 (UTC)
 * Those are not alternative spellings (like "color" and "colour"), those are alternative orthographies (like речник and rečnik, or like 複雑, ふくざつ and fukuzatsu). Does the same strict rules of CFI apply to those? In that case, somebody should take a look at the template, which doesn't provide for a way to suppress the alternative orthography if it's not attested. Also, some pages (like patriĉo) use Esperanto texts in an alternative orthography already to cite a word in the default orthography. 23:59, 24 October 2018 (UTC)
 * I think that the distinction between spellings and orthographies deserves greater attention. I would put your English and S-C examples in the same bucket as Esperanto (each spelling should be attested), whereas the Japanese example is more a way of redirecting readers to our entries as a result of the vagaries of Japanese script. —Μετάknowledge discuss/deeds 01:21, 25 October 2018 (UTC)
 * For the sake of argument, what about entries like OS that link to オーエス? —Suzukaze-c◇◇ 02:33, 26 October 2018 (UTC)
 * If I were dictator of Wiktionary, I'd say that unless attested, that link should be removed. —Μετάknowledge discuss/deeds 03:34, 26 October 2018 (UTC)
 * (honestly same —Suzukaze-c◇◇ 02:50, 31 October 2018 (UTC))
 * I would say that the links for predictable alternative writing systems were more akin to grammatical inflections. Having a link to dicerere, the alternative 2s imperfect subjunctive passive of Latin  is perfectly reasonable.  Having a link to a non-existent entry for such a form is useful; problems arise when there is an entry for such a form in another language.  A possibly relevant issue is inflected forms in the alternative writing systems; what page should list them? RichardW57 (talk) 17:48, 26 October 2018 (UTC)
 * 1) . The problem/use case described is understandable, but creating many new entries/pages (i.e., changing the content of the repository) is not a good solution for what is an indexing/resolution issue. I likewise wouldn't support the creation of a "corazon" entry simply because "corazón" is inconvenient to type with certain software. I agree with Metaknowledge that entries should be based upon attested forms. —User:miseDominic
 * : Per above. --Victar (talk) 18:38, 31 October 2018 (UTC)
 * : Allahverdi Verdizade (talk) 19:52, 3 November 2018 (UTC)

Abstain

 * 1)  It's an interesting conundrum. These aren't hypothetical alternate spellings, but, if I recall correctly, completely valid alternatives in standard Esperanto. I'd say they should be allowed if the standard Esperanto spelling is attestable, but I don't care enough to rock the boat in either direction. —  [ זכריה קהת ] Zack. — 17:13, 26 October 2018 (UTC)
 * 2)  I am inclined to oppose per Metaknowledge; let's document actuals and not plausible but unattested hypotheticals. The word "valid" as applied to these forms means nothing to me and should mean nothing to a descriptivist dictionary. What gives me pause is the argument that these entries would improve searchability of Esperanto entries, and that they should therefore be treated similarly to romanizations, as soft redirects. On the other hand, these are not romanizations since the original script is Latin. The question would then be whether vacice should be created as a diacritic-free entry for Czech vačice, to make things easier for people who do not know how to enter the diacritics on their keyboards. As for "orthography", I do not believe we want to make any statements as to what is "ortho-", right; in the mainspace, for reasonably well attested languages, we only report what we know to exist thanks to the observational evidence. --Dan Polansky (talk) 09:18, 4 November 2018 (UTC)
 * A good summary of the discussion so far. One small point I'd dispute, though: this is not quite the same thing as a hypothetical vacice for Czech. It's like if Czech, in addition to the letter č, also allowed the use of a digraph (say ch) for the same letter as equivalent in all contexts, and thus the entry would be vachice. That's what's happening here: it's not aca which is being proposed for aĉa, but acha and acxa. I'd therefore posit that romanization is a better analogy for this, as you wouldn't ever have a word with a diacritic but also a digraph - the systems are never mixed. ARR8 (talk) 01:22, 5 November 2018 (UTC)
 * If you enter "vacice" in the search bar, it automatically redirects to . This works for all the diacritics I've tried so I don't think diacritic-less redirects for user convenience are necessary; in fact they are inconvenient because they require one more click to reach the main entry than the search bar automatic redirect. —AryamanA (मुझसे बात करें • योगदान) 02:57, 8 November 2018 (UTC)
 * We are talking about Esperanto sunlevigxon to link to sunleviĝon. When a user enters "sunlevigon", diacriticless, they seem to be redirected to sunleviĝon; I tried that and was redirected. Therefore, the searchability for those who do not know how to enter diacritics seems really good. Sure, this will not work when you enter sunlevigxon, but you should not be entering that if it is not attested. --Dan Polansky (talk) 08:56, 10 November 2018 (UTC)
 * The trouble is with words like ĝoji (goji) whose diacritic-less forms are themselves words in other languages. Here, the Esperanto word is not even suggested in the search box. Also, attestation is not really meaningful here. Any Esperanto-speaker will correct levigxo to leviĝo in their heads, and vice-versa when writing; it is just a different way of writing a word. ARR8 (talk) 03:25, 11 November 2018 (UTC)
 * For goji -> ĝoji, we have also placed at the top of goji and producing "See also: ĝoji and go-ji".
 * As for "it is just a different way of writing a word": It is a way that all too few people use, or else it will get attested and then it can be created without any further ado. --Dan Polansky (talk) 10:24, 17 November 2018 (UTC)

Decision
No consensus. This, that and the other (talk) 10:28, 14 November 2018 (UTC)