User talk:Dan Polansky/2017

Pronunciation of 'w' in Czech
Hi Dan, How would you pronunce 'w' when written in Czech? (e.g. Wikipedie, Winchester) – AWESOME meeos ！ ＊ （「欺负」我） 12:51, 2 January 2017 (UTC)
 * @Awesomemeeos: It is pronounced like v. Thus, Wikipedie would be /vɪkɪpɛdɪjɛ/. --Dan Polansky (talk) 09:09, 7 January 2017 (UTC)
 * You can hear the Czech pronunciation of v at váza. --Dan Polansky (talk) 09:09, 7 January 2017 (UTC)

Czech lemmatizer
A Czech lemmatizer is at http://lindat.mff.cuni.cz/services/morphodita/

My favorite setting for the lemmatizer is as follows:
 * Task: Lemmatize
 * Tag set: Raw lemmas
 * Output: Plain

Example input: Komu není shůry dáno, v apatyce nekoupí. Komu se nelení, tomu se zelení.

Example output: kdo být shůry dát, v apatyka koupit. kdo se lenit, ten se zelený.

One use of this lemmatizer is that you pick a piece of Czech text, run it through the lemmatizer, wikify words and fill redlinks by creating Wiktionary entries. This redlink-filling activity was suggested by SemperBlotto some time ago, without the lemmatization part. Since I am interested in creating lemmas rather than inflected forms, I need a lemmatizer.

--Dan Polansky (talk) 18:39, 13 January 2017 (UTC)

The following Python script grabs clipboard content, wikifies words and puts the result back to clip: The regex may need finetuning. --Dan Polansky (talk) 18:54, 13 January 2017 (UTC)

CFI vote
I made the CFI vote start at 19:00 today (my local time is 18:49). I put back the end date as well. I removed the 'premature' tag.I hope that was the right thing to do.

John Cross (talk) 16:50, 17 January 2017 (UTC)
 * What you did in Votes/pl-2017-01/Trimming CFI for Wiktionary is not an encyclopedia 2 was fine, thank you. --Dan Polansky (talk) 18:51, 17 January 2017 (UTC)

Czech words for females
I hesitate how to mark up definition lines for Czech words for females such as učitelka, lékařka, ředitelka and prezidentka. The problem obviously applies to other languages as well, e.g. German Professorin.

One option that I have often used and that is quite possibly prevalent in the Czech entries is like this: A disadvantage of that is that the word "female" does not usually appear in translation; you do not say "she is a female teacher" but rather "she is a teacher".
 * 1) female teacher

Another option that I must have used at least once is this: What I do not like about this is that the disambiguator "female" appears only in the gloss, but maybe it's okay. Furthermore, I like the gloss to be an abbreviated definition, which "female" isn't; it would be "female teacher", which would lead to a repetition of "teacher" in:
 * 1) teacher female
 * 1) teacher female teacher

Another option that I must have seen somewhere is the use of a context label: That actualy looks okay since, in an English sentence like "she is a teacher", the subject of the sentence (she) is in the context of the predicate (be teacher).
 * 1)  teacher

Based on the above, I may stay with "female teacher", or I may switch to " teacher".

--Dan Polansky (talk) 11:49, 22 January 2017 (UTC)
 * It would be good if allowed you to put "m=učitel" of "f=učitelka" in the headword. But, anyway, I think "teacher (female)" or "teacher (male)" is the way to go. SemperBlotto (talk) 11:54, 22 January 2017 (UTC)


 * I don't see "female teacher" (or indeed "male teacher") being a problematic definition: they don't have to be an exact word-for-word phrase that you can insert into a translation without thinking. There are plenty of English entries of the same kind, like usherette. Equinox ◑ 20:54, 22 January 2017 (UTC)

(outdent) Someone likes to use feminine noun of, it seems, seen in Catalan psicòloga or French masseuse. Lehrerin uses that as well, but used to use feminine of:
 * , a female teacher person who teaches.

The above was converted to feminine noun of in July 2015 by MewBot. In more distant past, Lehrerin used to have the following plain markup:
 * (female) teacher (a person who teaches)

--Dan Polansky (talk) 17:57, 4 February 2017 (UTC)

No Babel/Language Categories
Hi Dan Polansky, just a general question, what do you feel when user pages do not have Babel and/or have categories derived from them? For me, I find it irritating. I wanted to find native French/Dutch/Japanese speakers, but they don't put Babel on their page, so I had to discover them from contributions from other pages – AWESOME meeos ！ *  (chōmtī hao /t͡ɕoːm˩˧.tiː˩˧ haw˦˥/) 22:55, 12 February 2017 (UTC)
 * I like Babel, which is why I often ask people to add it to their use pages. Babel is relatively important in a dictionary project, especially since we have seen multiple editors contribute in a plethora of languages they do not speak; with Babel, we know whether the person relied in part on their knowledge of the language or whether they had to go by sources alone. --Dan Polansky (talk) 10:18, 18 February 2017 (UTC)

Slovensko etymology
Hi Dan, I wonder why this slovak word did not have its etymology fixed up before? (I had to do it myself) — AWESOME meeos ！ *  (не нажима́йте сюда́ [nʲɪ‿nəʐɨˈmajtʲe sʲʊˈda]) 19:43, 10 March 2017 (UTC)
 * What makes you think it came from OCS rather than from an inherited form? --WikiTiki89 19:57, 10 March 2017 (UTC)
 * It was based off the original etymology. — AWESOME meeos ！ *  (не нажима́йте сюда́ [nʲɪ‿nəʐɨˈmajtʲe sʲʊˈda]) 19:59, 10 March 2017 (UTC)
 * I did not edit the etymology of the Slovak entry Slovensko, from what I remember and from what I can see by a quick glance at the revision history. I am not into Slavic etymology; what I did in etymology many years ago is source English etymologies from Century 1911, checking with modern sources. --Dan Polansky (talk) 07:52, 11 March 2017 (UTC)

External link templates and excessive detail
Keywords: reference templates, baroqueness, ornament.

I strongly prefer external link templates that are simple, simply formatted and contain minimum identification information.

In particular:
 * ISBN and OCLC are not necessary for unique identification and present visual noise. In general, I feel exposing numerical identifiers on user interface is rude to the user unless the user requested them or has special urgent need for them.
 * I see no need to render Dictionnaire Illustré Latin-Français into English as "[Illustrated Latin–French Dictionary]"; the translation should be pretty clear to anyone who understands English.
 * I see no need to spell author names in full. That is unnecessary for unique identification, and adds words for the eye to read.

There is cost to the skimming reader in being exposed to unnecessary detail for their eye to parse. Let us recall the busy search entry page of AltaVista search engine from the last century that was replaced by the beautiful minimalistic Google search entry page; the removal of items that were of no interest for the searching user was a major user experience improvement.

For those who feel a strong need for a lengthy, baroque identification, such identification could be placed to an appendix linked from the external link template. In such an appendix, there is a plenty of room for extraneous detail, and there, it does not disturb anyone's skimming.

For templates that are hyperlinking to online sources as opposed to merely referencing paper sources, the full identification is also present on the linked site. The key purpose of these templates is to take the reader to the target page to learn more about a particular word. Once the hyperlink is there, getting to the target information-bearing page is a matter of a single mouse click, unlike in the age of paper information resources when the resource identification was key for getting to the page where the referenced information was located.

For Wiktionary maintenance purposes, fuller identification can be placed on the template documentation page. Thus, even when a website goes down, the template documentation history shows details the resource once linked.

--Dan Polansky (talk) 07:40, 18 March 2017 (UTC)
 * To prevent further reverts, I create a vote: Votes/2017-03/Reference templates and OCLC. I remember a BP discussion on the subject where I argued extensive identification could be in appendices, but I cannot find the discussion. --Dan Polansky (talk) 08:16, 18 March 2017 (UTC)
 * You're definitely right that such a discussion about appendices happened, but I can't seem to find it either. I'll try and look around a bit more. — JohnC5 08:26, 18 March 2017 (UTC)
 * A little bit of discussion is now unfolding at Beer parlour/2017/March. --Dan Polansky (talk) 08:42, 18 March 2017 (UTC)

One work online that uses reasonably short link identification is Encyclopedia Britannica. To wit, its article on tiger has Additional Reading section containing K. Ullas Karanth, The Way of the Tiger: Natural History and Conservation of the Endangered Big Cat (2001) as an identification, with no ISBN and no place of publication. Its article on Australia contains similar style, e.g. Tony MacDougall (ed.), The Australian Encyclopaedia, 6th ed., 8 vol. (1996) and Australian Bureau of Statistics, Year Book, Australia (1978– ).

--Dan Polansky (talk) 12:15, 19 March 2017 (UTC)

Another work using reasonably short identification, of lexicographical quotations, is Oxford English Dictionary (oed.com). To wit, OED entry Popperian contains e.g. the following quotations, with the following approxiate appearance:
 * 1958  Brit. Jrnl. Philos. Sci. 9 228   On a Popperian theory of confirmation this hypothesis is confirmed by an observation-report of a black raven.
 * 1972  Nature 10 Nov. 110/1   The author then delineates in turn three theories of the logical structure of science—inductivism, Popperian falsificationism and positivism.
 * 1977  A. Giddens Stud. in Social & Polit. Theory i. 72   Lakatos's studies, although nominally directed at supporting main elements of the Popperian standpoint, show how wide the discrepancy is.

OED makes more detail available when you click on the link in the quotation, e.g. on "Brit. Jrnl. Philos. Sci.". The used format drives the reader's attention toward the quotations themselves, which I like.

--Dan Polansky (talk) 15:46, 10 June 2017 (UTC)

Much later: Polynomial, mathworld.wolfram.com, uses the following identification for a reference:
 * Barbeau, E. J. Polynomials. New York: Springer-Verlag, 1989.

Observations: 1) The first names are abbreviated. 2) There is no ISBN. --Dan Polansky (talk) 10:10, 30 March 2018 (UTC)

A discussion: Beer parlour/2018/May. --Dan Polansky (talk) 12:12, 27 May 2018 (UTC)

Condition
Hi, you have used the following condition in several admin votes:
 * on the condition that the editor will lose admin flag if, in future, someone creates a vote that seeks to confirm him in the adminship and the vote does not achieve consensus for keeping adminship; oppose to the extent the condition is not met. This is nothing personal; it is as a matter of general useful principle. A clarification: My position is that my condition only applies if passing of the vote depends on support of editors who used this condition.

Would you mind explaining the rationale underlying that? Lingo Bingo Dingo (talk) 13:56, 13 April 2017 (UTC)
 * @Lingo Bingo Dingo: The rationale is that you cannot know how good an admin someone is until they actually become one and start acting in that role. Furthermore, people tend to behave better before they get in the position of power. Therefore, it has to be reasonably easy to remove the admin flag, which it is not if 2/3-supermajority is required for removal. The event critical for control of admin power is not ensysopping but desysopping. It does not seem to make sense that entering a role requires surpassing 2/3-bar while keeping that role once it is known that you are not entirely fit for it requires surpassing 1/3-bar.
 * I admit that this issue is not so critical in larger wikis such as the English Wiktionary. But it is still a good principle. --Dan Polansky (talk) 11:12, 16 April 2017 (UTC)
 * I see, so the confirmation vote will require a ⅔ majority. Doesn't that also risk making it too easy for a group of troublesome users to block a well functioning admin? Wouldn't it make more sense to have more built-in checks? Lingo Bingo Dingo (talk) 13:17, 25 April 2017 (UTC)
 * @Lingo Bingo Dingo: There is such a risk, I admit, but it does not seem very serious to me, based on my wiki experience. Those who are bothered by that risk could be okay with our requiring 1/2 support instead of 2/3 support for a confirmation. From my perspective, even 1/2 begs the question of why 2/3 are required for someone who is not really known in their admin role while 1/2 are required for someone know is known. Still, even 1/2-confirmation would be a worthwhile improvement over what we have now. --Dan Polansky (talk) 17:27, 28 April 2017 (UTC)

Let us postpone the vote as much as discussion requires
I see that you like to say this when you create a new vote: "Let us postpone the vote as much as discussion requires."

I like that idea. I suggest adding it in Voting policy, so it will apply to all future votes. The policy says nothing about the 7-day waiting period in the first place, so I propose adding this item there, too.

Proposed wording:


 * 1) Once a vote is created, it should have a minimum 7-day waiting period before it starts. (except user right votes and bot votes)
 * 2) The start of a vote can be postponed as much as discussion requires.
 * 3) A vote for granting the rights of an administrator, checkuser or bureaucrat may start immediately after the recipient accepts.
 * 4) A vote for granting a bot flag may start immediately.

Is that OK for you? Would you change the wording? --Daniel Carrero (talk) 17:59, 13 April 2017 (UTC)
 * Votes are for concluding discussions, so if the discussion hasn't even started, it's not appropriate to create one. —CodeCat 18:01, 13 April 2017 (UTC)
 * I agree with CodeCat. The vote shouldn't have been created if the discussion is still going on (or worse, hasn't started yet). --WikiTiki89 18:07, 13 April 2017 (UTC)
 * @Daniel Carrero: I don't think we need more policy text in this case; we have common practices that serve us well. Other than that, I do not see a problem with the proposed text, except for having an except-clause in brackets in the first item; exceptions or at least their existence should ideally be indicated as part of the sentence to which they apply.
 * @CodeCat, Wikitiki89. I obviously diagree, as is apparent from my vote-creating practice. Requiring that discussion took place before a vote page is created is avoidable bureaucracy. A vote talk page page is a suitable place in which to conduct a discussion; it is well suited for multiple section headings pertaining to the vote subject. --Dan Polansky (talk) 10:54, 16 April 2017 (UTC)
 * In defense of having policy text: It would hopefully help newcomers. IMHO, having policy text is better than not having it, because to learn unwritten rules requires observing what's happening or asking people about it. Just reading about the rules should make things easier.
 * Concerning my proposed text, I agree with you about the except-clause in brackets in the first item. That part could be improved. --Daniel Carrero (talk) 11:05, 16 April 2017 (UTC)
 * I agree with Dan Polansky about this: "Requiring that discussion took place before a vote page is created is avoidable bureaucracy. A vote talk page page is a suitable place in which to conduct a discussion; it is well suited for multiple section headings pertaining to the vote subject." --Daniel Carrero (talk) 14:00, 18 April 2017 (UTC)

Abyssine
In reply to your old query at User_talk:Speednat, I can verify that the quotation is genuine and that the book in which it appears is on GBS. However, it lacked sufficient identifiers and long s had been replaced by normal s; as a result of the long s, the machine-read text that was quoted is quite a load of GBP: exprefsly forbidding the Abyffine clergy, monks, and priefls, to perform any prieftly functions. Lingo Bingo Dingo (talk) 13:11, 25 April 2017 (UTC)
 * Thank you. I think the point would have been, back then, where did the user find the quotation? If the user found the quotation on the Internet, as they claimed, they could have pointed to that location, the way you did, couldn't they?
 * does not find the location you pointed to, so in what manner did the user find the quotation? If you have access to The Shorter Oxford English Dictionary, 5th edition, Oxford University Press, 2003, you may check whether, incidentally, the quotation is also there. --Dan Polansky (talk) 16:02, 28 April 2017 (UTC)
 * Unfortunately I don't, though I can access the 6th edition. But I think the circumstantial evidence, including the unoriginal typography, already point toward copyright violation. Do you think that the entry in its current form is still problematic in the event we establish violation? Lingo Bingo Dingo (talk) 09:41, 29 April 2017 (UTC)
 * Copying a single quotation from a copyrighted dictionary to Wiktionary does not seem to warrant a removal, by my estimate. What is protected is the selection of quotations, not the quotations themselves. Therefore, there would be a problem only if the copying was systematic, on a larger scale. The goal of my posts to Speednat talk page was to prevent such a copying, if taking place, from becoming systematic. --Dan Polansky (talk) 09:50, 29 April 2017 (UTC)
 * I checked the 6th and 4th editions of The Shorter Oxford English Dictionary, both lack any quotation at Abyssine. If there was any copying, which I still suspect, it was likely from another source. Lingo Bingo Dingo (talk) 14:54, 3 May 2017 (UTC)

Noun or adverb - horko and allies
Some Czech -o words feel like a special kind of animal: mokro, sucho, dusno, horko, vedro, parno, studeno, ošklivo, mrazivo, slunečno, zamračeno, jasno, polojasno, veselo, smutno, teskno, pusto, prázdno but maybe even světlo, and ticho. As for their semantics, they seem to refer mostly to weather conditions and to human mood conditions. Their derivation consists in replacing the -ý of an adjective with -o. Compared to -ost yielding nouns referring to qualities, the -o derivation is rather unproductive.

They look quite a bit like nouns but SSJC and PSJC analyze multiple of them as both adverbs and nouns, depending on use. The uses that SSJC analyzes as adverbial are often translated into English using adjectives, as in Je mi smutno -> I am sad. As nouns, they could be thought of as referring to qualities. A question is why they don't end in -ost, which is very productive in generating names of qualities.

Some sentences in which the nounhood is not obvious:
 * Venku je horko.
 * Je mi vedro.
 * Je mi smutno.

The nounhood is obvious in the use of inflected forms:
 * Nastala vedra.
 * A je to v suchu.
 * Pojďme radši do tepla.

Adverbhood is suggested by there being words that have typically adverbial suffix and morphological appearance yet are used in a way that is similar to those -o words:
 * Je mi zle.
 * Venku je krásně.

Nounhood is further suggested by words that have non-adverbial suffix, have noun inflection, and are used in a similar way: However, the above uses are analyzed by SSJC as adverbial[REALLY?].
 * Venku je zima.
 * Venku je tma.

In Colloquial Czech, 2014, James Naughton mentions these -o words as follows:
 * "Some weather phrases use adverbs ending in -ě, but others use special forms ending in -o:"

SSJC has them often both as adverbs and nouns, even zima:
 * horko:
 * adverb: as in "je mi velmi h."
 * noun: 1. velké teplo; parno
 * smutno:
 * documented in smutný
 * noun: velké s. bylo po pohřbu v zahradě
 * adverb: bylo jí s.
 * zima:
 * adverb: chladno: není ti z.?
 * noun: ...; 2. velmi nízká teplota
 * tma:
 * adverb: temno; je úplně t.
 * noun: 1. naprostý n. částečný nedostatek světla; ...

What is it about the use like "je horko" that makes it necessarily adverbial? In "panuje horko", the horko is obviously the subject of the sentence and therefore not an adverb; and furthermore, in "panuje hrozné horko", horko is modified by an adjective. "je horko" can be analyzed as a short for "tam venku je horko" where there could equally be "tam venku je spousta lidí". In "je mi horko", the subject of the sentence seems to be "horko" (and not "já", inflected as "mi"), and therefore "horko" again would be a noun. However, in "je mi zle" and "je mi krásně", "zle" and "krásně" have adverbial suffixes, and it could be argued that the implied subject of the sentence is not "zle" but rather "to", that is, "to je mi krásně". Even then, that does not make "je horko" necessarily adverbial but merely ambiguous between nominal and adverbial. In "je hrozné horko", modification by an adjective suggests a noun; in "je mi hrozná zima", it is a noun again, but in "je mi hrozně smutno", modification by an adverb suggests an adverb; "je mi hrozně zima" can be found as well, whereas "je mi hrozné smutno" I do not recognize as Czech.

As per above, in some cases, the ambiguity is resolved by one of the following features: In the cases where the ambiguity is resolved using adverbial modification, it is hard to argue that the use is not adverbial, which I would very much like to do but do not know how.
 * Modification by an adjective vs. adverb ("je velké horko": noun, "je mi hrozně zima": adverb)
 * Being the subject of a non-auxiliary verb ("panuje horko": noun)
 * Being used in a nominal inflected form ("v tom horku nikam nejdu": noun)

Komično, tragično, krásno, fantastično, absurdno and absolutno also appear to be nouns, and are ending in -o, but they seem to be of a somewhat different sort.

In German, the ambiguity between noun on one hand and adverb or adjective on the other hand cannot so easily arise since German capitalizes nouns. So there is kalt and Kälte, and warm and Wärme. In Es ist mir kalt, kalt cannot be a noun. On the other hand, on the morphological surface, German shows ambiguity between an adverb and an adjective, so "Es ist mir kalt" is not unambiguously adverbial. The "Es ist mir x" pattern seems to be much less commonly used of human moods: Es ist mir traurig seems much rarer than Ich bin traurig, whereas in Czech, "je mi smutno" is very usual, no less so than "jsem smutný". On Es ist mir kalt, How does "Mir ist es kalt" make sense?, reddit.com, agrees with my understanding that it could be rendered as "It is cold to me", and argues that "Es ist mir kalt" makes actually more sense from a certain standpoint than "Ich bin kalt" and the English "I am cold". That was something of an aside, but anyway.

As a curiosity, the pattern was used to create the nonce word smažno to translate brillig of Jabberwocky, where 'Twas brillig was translated as Je smažno.

--Dan Polansky (talk) 10:10, 30 April 2017 (UTC)

Your not helpful advices

 * Your advice advice is not helpful, because you miss the point behind my notes: Being not in a position as admin, such words are the last means in cases where people do not react upon pledges for giving scientific sources for their claims, and even resort to nationalistic reasoning. And please note that your advice would be much more useful in those cases. I hope you see the point. HJJHolm (talk) 05:29, 2 May 2017 (UTC)
 * Dan, I apologize for the spam on your talk page, but I'd like to make some points to this user. To what "nationalistic reasoning" are you referring? Where are there "pledges for giving scientific sources" on this site? Also, Dan is an experienced user of this site whose opinion is very important to our discourse, and his advice to you was quite good. Also, what are you implying by the phrase "Being not in a position as admin"? How would your being an admin change these situations at all? Would you block editors with whom you disagree or protect pages without consensus? I'll warn you that is you continue with your current strain of rude behavior, you will lose what credibility and goodwill you currently retain. Again, sorry for the spam, Dan. — JohnC5 06:09, 2 May 2017 (UTC)
 * I think HJJHolm, not a native speaker, was trying to say that he, not me, is not an admin.
 * For reference, this thread is in response to, where I asked HJJHolm to stop using "!!!" and "???" and made some other requests. Meanwhile, HJJHolm has removed my request from his talk page. --Dan Polansky (talk) 18:50, 8 May 2017 (UTC)
 * Yeah, I understood all of that when I wrote this response, but maybe I was unclear that I was asking HJJHolm the quesitons. It's not important anyway. Thanks for your help, Dan. — JohnC5 19:23, 8 May 2017 (UTC)

Translation section problems
Recently, my translation sections stopped being collapsible and some bookmarklets stopped working. Using Firefox web console (Control + Shift + K), I realized there appeared the following error: This stemmed from User:Dan_Polansky/common.js, where I now commented out the following, and now it works again:
 * "ReferenceError: Tbot is not defined" [object Error]

importScript('User:ZBroz/Tbot.js'); addOnloadHook(function { Tbot.greenifyTranslinks('cs'); Tbot.greenifyTranslinks('sk'); });

Maybe there is something in User:ZBroz/Tbot.js that no longer works with Mediawiki. --Dan Polansky (talk) 17:53, 9 May 2017 (UTC)

Navajo
The utility of current CFI's approach to Navajo is being discussed e.g. in Requests_for_verification/Non-English (later at Talk:Sǫʼ Naalgeedí Gohwééh); two dictionaries are mentioned. Being a less documented language, Navajo can be verified by a single mention, as per WT:CFI. About Navajo currently does not mention inclusion criteria. Somewhat related is a failed vote Votes/2012-03/CFI for Endangered Languages. A relevant BP discussion is Beer parlour/2012/December. Talk:jádí dághaaʼígíí has a talk on a RFV-failed Navajo term.

Some Navajo dictionaries:
 * Navajo-English Dictionary by Wall and Morgan, 1958
 * The Navajo Language: A Grammar and Colloquial Dictionary by Young and Morgan, 1980
 * A Navajo/English Bilingual Dictionary by Alyse Neundorf, 1983
 * Analytical Lexicon of Navajo by Young and Morgan, 1992

A Navajo dict list: --Dan Polansky (talk) 13:30, 21 May 2017 (UTC)
 * Subject > "English language Dictionaries Navajo.", searchworks.stanford.edu

Renomination for admin?
I now realize that the reasons for my opposing your previous admin vote were silly. I'd like to renominate you. I realize that adminship on Wiktionary is only about proper usage of tools and nothing else, and has nothing to do with discussion etiquette or opinions against other people, and after thinking about it that way, I think you'd be fine with admin tools. After diverting my way of thinking adminship was almost completely a position of authority, I think anyone who knows how and when to use the tools should have them, regardless of my opinion of the person. You've been here for 10 years making nothing but great contributions, and occasionally flaming other users, like me, but, regardless of that, what I'm saying is that you know how to edit well, and that's what counts. (I know; two admin nominations in one day, but I just had adminship on my mind.) I feel other users may have had the same misperception that I did. Do you accept the nomination? PseudoSkull (talk) 19:28, 25 May 2017 (UTC)
 * Thank you. However, instead of starting a vote that is likely to fail (remember it needs 2/3 in support), let people who have changed their minds post support votes here, on this talk page. Also, let people consider a conditional support, such as "Support as long as the use of the blocking tool is forbidden" since that would help me advance RFV and RFD without having the block tool, or, more likely, without being allowed to use it. --Dan Polansky (talk) 16:41, 26 May 2017 (UTC)
 * For context, my admin vote failed in August 2016. --Dan Polansky (talk) 16:53, 26 May 2017 (UTC)
 * Note that the block tool is one of power over other people, and should not be awarded to people who we cannot trust, no matter how good editors they are. The use of the block tool does not require consensus, and blocks are rarely challenged. Multiple of current admins are not qualified to use the block tool, in my view. The deletion tool can be abused to hide trails of conversation; it was used in this way by an English Wiktionary admin who meanwhile ceased editing. --Dan Polansky (talk) 17:00, 26 May 2017 (UTC)
 * I feel that this is acceptable. Perhaps we could start a BP discussion on the compromise you just suggested? This talk page doesn't seem to be receiving much attention at the moment. PseudoSkull (talk) 00:13, 27 May 2017 (UTC)

Languages requiring 3 quotations
To address certain complaints, we could make the following change: --Dan Polansky (talk) 09:29, 3 June 2017 (UTC)
 * 1) Fold Criteria for inclusion/Well documented languages into the main WT:CFI page, making the page a redirect to the new section in CFI.
 * 2) Use heading "Languages requiring 3 quotations" for the new section in CFI rather than "Well documented languages".
 * 3) Change the intro sentence from
 * The languages well documented on the Internet as provided on the Criteria for inclusion page are:
 * to
 * The languages requiring 3 quotations in use for attestation are as follows:
 * 4) Updating WT:CFI by changing
 * For languages languages well documented on the Internet, three citations in which a term is used is the minimum number for inclusion in Wiktionary.
 * to
 * For languages requiring 3 quotations, three quotations in which a term is used is the minimum number for inclusion in Wiktionary.
 * 5) Removing as unnecessary the following: "This page may be modified through general consensus. To make a request to add or exclude a language, go to the Beer Parlour and click the "+" tab at top to input your request. When making a change to this list, consideration of how to handle existing entries should be taken into account."

Template:ws in entries
At the moment, exactly 55 pages in the main namespace use. Do you think that is acceptable? The template documentation states that it's for Wikisaurus pages only. --Daniel Carrero (talk) 08:12, 7 June 2017 (UTC)
 * Using ws in the mainspace does not make any see. The editor might have meant, in kulak, "See also Wikisaurus:employer". --Dan Polansky (talk) 10:15, 7 June 2017 (UTC)
 * Hah. It actually does something interesting, in kulak: it behaves exactly in Wikisaurus, and therefore, links to mainspace and is followed by "[WS]" linking to Wikisaurus. I cannot say the resulting appearance for the user in necessarily a bad thing. However, if this becomes widespread practice, it becomes harder to make changes to ws that have Wikisaurus in mind since these changes will also impact mainspace. --Dan Polansky (talk) 10:18, 7 June 2017 (UTC)

Wiktionary popularity among online dictionaries per Alexa rank
This is a follow up on User talk:Dan_Polansky/2013. Compared to the 2013 report, I kept in the table only sites that have one of the main three monitored Alexa ranks <= 1000, but made some exceptions based on feeling. I added some sites that were not there before. Where Alexa requires subscription to view a particular piece of data, I indicate "S/R" (subscription required).

Rank of dictionary web sites per number of visitors per Alexa.com, ordered by global rank:

See also http://www.alexa.com/topsites/category/Top/Reference/Dictionaries, a list of top sites in Alexa Dictionaries category. There, Wiktionary is 3rd; the score used may be the global Alexa rank or something else.

wikipedia.org has the global Alexa rank of 5; wikibooks.org: 2,006; wikisource.org: 4,134; britannica.com: 2,025 (U.S: 825).

ahdictionary.com has a terrible rank, whyever.

translate.google.com is a translation service that can be used as a dictionary which I do not know how to include; Alexa does not seem to provide ranks for subdomains, and this subdomain will not grab a large portion of the overall google.com traffic. Similarly for bing.com/translator.

hypestat.com (about) is an Alexa-like website that seems to use Alexa data. Its wiktionary.org entry has ranks for many countries; I wonder whether these are in fact Alexa ranks, maybe a bit outdated. They seem to show data from multiple sources, including Alexa, majestic.com, semrush.com, Google and Quantcast. The ranks for wiktionary.org:
 * Finland: 144
 * Sweden: 156
 * Greece: 169
 * France: 196
 * Poland: 211
 * Switzerland: 212
 * Czech Rep.: 235
 * Belgium: 248
 * Austria: 267
 * Russia: 272
 * Germany: 295
 * Norway: 357
 * Netherlands: 364
 * Israel: 365
 * Kazachstan: 374
 * Ukraine: 392
 * Azerbaijan: 400
 * U.K.: 507
 * Vietnam: 545
 * Canada: 561
 * Romania: 577
 * Italy: 621
 * South Korea: 667
 * Australia: 684
 * U.S.: 719
 * Spain: 1006
 * Taiwan: 1155
 * Turkey: 1271
 * Mexico: 1302
 * Iran: 1416
 * Brazil: 1684
 * India: 2254

Are these ranks accurate and relevant? They look pretty impressive, don't they?

A note on Czech Rep., of interest to me: per hypestat.com, Czech Rep. makes 1.6% of wiktionary.org pageviews, while accesses to cs.wiktionary.org make 0.75% of pageviews. If these numbers are comparable, we can calculate their difference, and obtain that at least 0.85% pageviews are traffic from Czech Rep. that does not go to cs.wiktionary.org. Possibly, a significant portion of these 0.85% pageviews is traffic from Czech Rep. that goes to en.wiktionary.org.

--Dan Polansky (talk) 18:19, 8 June 2017 (UTC)

Let's have a look at daily page views and unique visitors of some of the top dictionary sites we see above. The data is pulled from hypestat.com, whatever their ultimate origin: --Dan Polansky (talk) 12:30, 10 June 2017 (UTC)


 * Hi there. Did you think to include SemperBlotto (talk) 12:36, 10 June 2017 (UTC) p.s. You can use, as a login code, hertsb123456789 (or any similar string starting herts (for Herts Libraries))
 * oed.com Alexa global rank is 39,412; its rank in U.S. is 24,201; its rank in U.K. is 13,135. hyperstat.com / oed.com shows 12,226 unique visitors and 36,677 page views per day. --Dan Polansky (talk) 12:41, 10 June 2017 (UTC)
 * It's beautiful. It's got plentiful, beautifully formatted quotations. No boldface on the year; I dislike the boldface we use. No ISBNs and other identification nonsense in the quotation presentation. No boldface on the term documented; I dislike the boldface we use. --Dan Polansky (talk) 12:51, 10 June 2017 (UTC)
 * Yes, I use it all the time (even though I am no longer a Herts library user). Other goodies are available via . SemperBlotto (talk) 12:56, 10 June 2017 (UTC)
 * Hi there. How about . It's my third favourite dictionaryish site on the web. --Celui qui crée ébauches de football anglais (talk) 15:48, 10 June 2017 (UTC)
 * onelook.com Alexa global rank is 23,914; U.S.: 9,763; U.K.: 10,843. In 2013, onelook.com was in the table. --Dan Polansky (talk) 07:08, 11 June 2017 (UTC)

English entries lacking IPA pronunciation
In 2015, I provided such a list based on a request made at Beer parlour/2015/August; I placed the result at User:Ultimateria/en-needing-ipa. --Dan Polansky (talk) 16:55, 16 June 2017 (UTC)

This time in person
About this in your user page: "And, this time in person, I just love Wiktionary!"

This may be a silly question. Does it mean you've been saying you love Wiktionary but not in person before? --Daniel Carrero (talk) 08:33, 24 June 2017 (UTC)
 * I start the description in 3rd person, "Dan Polansky is ...". I follow in that tone in another sentence. The "this time in person" phrase was to indicate the switch to "I". I am not sure it is idiomatic English, though. --Dan Polansky (talk)

Czech words with accute u in the middle
Czech words with accute u (ú) in the middle, excluding the prefix-generated cases such as vyúčtování, are of interest since what is usually used in the middle of a word is ů, with the small circle above. These ú words often originate from non-Slavic words, and include the following: búr, difúze, drúza, frizúra, fúrie, fúze, iglú, infúze, kalambúr, kešú, konfúze, kúr, kúra, manikúra, medúza, mundúr, múza, ocún, pedikúra, petúnie, ragú, Rút, Saúd, skútr, súra, štrúdl, transfúze, trubadúr, túje, túra, and vúdú; proper names include Bejrút, Bhútán, Búr, Fúrie, Káthmándú, Núbie, Pú, Súdán, etc.; adjectives include difúzní, múzický, saúdský, súdánský, súfijský, etc.; inhabitant names include Núbijec, Súdánec, etc.

rúž does not seem to be used (has Slovak entry), and růž is attested, despite the apparent foreign origin from rouge; it may have been influenced by růže.

Unlike ú, ů is never written at the beginning, so we have úcta, účel, úprk, etc.

By some process, Slovaks managed to get rid of ů and only keep ú, where Slovak is a language very close to Czech, mutually intelligible. --Dan Polansky (talk) 07:49, 1 July 2017 (UTC)

My apologies
Dear Dan, of what I realized, you consider me as just a lexicographical contributor of Greek content to the English Wiktionary and not as a fellow wikipedian, which means you consider me as spot who is not eligible to have any opinion about what any WMF project must (or must not) do. Hence I want to apologize if I offended you in any way, either by the way I used the English language in the beer-parlour or even by me, contributing in en-wiktionary in a non-mynativelanguagehere content once a year. Be sure that if had seen your opinion in a parallel issue I would never answered in beer-parlour (or, at least, in the way I did it). Again, my apologies for any offending edit I did before or now. --Xoristzatziki (talk) 22:48, 1 July 2017 (UTC)
 * You are responding to Beer parlour/2017/June. When I wrote there "As a contributor of Greek content to the English Wiktionary, you do not need to fear the competition of further reading sources, I think", this was not meant to be demeaning at all; you do contribute Greek content, and you expressed a worry that your contributions will be less valuable, in '... I fear that my contributions are only for supporting these specific "further reading" sites ...'.
 * As for my English post to Slovak Wiktionary from April 2015 in which I opposed wholesale removal of images, I see nothing wrong about it; perhaps someone should elucidate to me what is wrong with it. There I wrote ".... above-voting User:Danny B. does not contribute any lexicography either (not here, and not anywhere else), nor does the above-voting User:JAn Dudík contribute any lexicography here, my vote stays." I think proper assessment of Danny B. and JAn Dudík was key there: these people think they have the right to propose and execute wholesale removal of images on a project on which they contribute almost no lexicographical content; by contrast, to me, the only people eligible to govern a wiki dictionary are those who contribute lexicography to it. An admin who only does adminship and no lexicography is something hard to imagine in the English Wiktionary. They removed images from the Czech Wiktionary wholesale and they started to do so in the Slovak Wiktionary as well, JAn Dudík with the help of a bot. --Dan Polansky (talk) 06:06, 2 July 2017 (UTC)
 * For the record, you are obviously welcome to discuss in the English Wiktionary Beer parlour, and I never indicated otherwise. --Dan Polansky (talk) 06:25, 2 July 2017 (UTC)
 * By "opinion" I mean exactly this. That for you, no one may have an opinion about anything, unless it contributes (with specific number of edits?) to that exact area. My opinion is far away from this. As an analyst and as a programmer for many years, I have found that the most valuable opinions is that of the ones that do not do the same thing as I do. I never told to a single client or observer of my work "You are not a programmer do not say what I can or cannot do". As a contributor to WMF projects I have seen that most contributors in a "specialized" area have difficulties to understand why other people, not expert in that area, may have opinion about that area (ex. the neutral point-of-view). But this is the magic of a collaborating system and especially the magic of WMF projects. Someone who never contributed to wiktionary, but is a constant reader (which we will never know...) may have a much better opinion of what this wictionary must or must not provide and how it must look. He sees things from the other side.
 * By the way, as a late crossword puzzle constructor have used (and possess) many dictionaries. Dictionaries do not include photos. Small encyclopedic dictionaries include, but are not for lexicographical reference, since they include less lexicographical content and specific to the photo. Also the existence of a photo greatly confuses the reader, but since wiktionaries in which I contribute decided to include them I do not interfere. At least I try to improve these wiktionaries and try to check with others (non contributors, but readers) if the photo is confusing or not (which means more of my voluntary time consumed in non "lexicographical content", less time contributing to "lexicographical content"). That said, I repeat that I would never answered the same way if I knew your opinion about the connection between contributions and the right to propose or the right to have any opinion. As I understand we are far far away about the objectives of the projects we all (voluntary) contribute. But no offense taken, I am here because I still imagine a world of free access to knowledge (as you also I believe). --Xoristzatziki (talk) 20:08, 2 July 2017 (UTC)
 * I object to your repeated misrepresentations of my positions. I can represent my positions myself, thank you.
 * My position is that anyone is entitled to an opinion, and that input from outsiders is welcome, but at the same time a wiki dictionary should only be governed by those who contribute lexicographical content to it. You are free to disagree. Note that the English Wiktionary has eligibility criteria for votes. Dictionaries do contain images, e.g. Century 1911 and Duden online. Historically, images took scarce space on the printed paper, which is why an American lexicographer drastically reduced their use in Webster's third, from what I remember; that made room for more words. We are not paper and we can afford to have a lot of images for the benefit of our readers. --Dan Polansky (talk) 20:37, 2 July 2017 (UTC)
 * Sorry for interfering once more, but I want to make clear that I was referring only to the part: "these people think they have the right to propose", not the rest. The right to vote is something I never mentioned. And of course I am aware about the existence of specific (and sometimes distinct) rules in all projects. But in beer parlour there was no voting, just conversation. And surely I will not discuss further anything about the lexicographical value of dictionaries with images, since I think such a discussion here is a waist of time (for all of us, not just for me). Sorry again if I offended you again. --Xoristzatziki (talk) 23:22, 3 July 2017 (UTC)
 * Let me point out that, on a project run as informally as Slovak Wiktonary, there is no significant distinction between a discussion and a vote. Danny B. and JAn Dudík oppose the idea of voting. In Slovak Wiktionary, there was no vote to remove all images; JAn Dudík let his bot remove images based on discussion alone, an inconclusive one at that. Therefore, Danny B. posts to discussion did act as votes, something to which, in my view, he had no right. The bot removal happened later; nonetheless, I saw in 2015 that this could happen since a similar thing happened in Czech Wiktionary. I made the April 2015 post to lend some support to the Slovaks opposing removal of images. I did not feel entirely comfortable doing that since I felt it odd that someone who does no lexicography in sk wikt such as myself should post opposes to that project; but I realized that since Danny B. does not do any lexicography either, I am no worse than him. You may feel that anyone should feel entirely free to post cross-wiki opposes; I am not entirely comfortable with that idea. --Dan Polansky (talk) 16:16, 5 July 2017 (UTC)

Inclusion of slogans
There is a deletion discussion WT:RFD (later at Talk:keep your rosaries off my ovaries) about a slogan.

We have workers of the world, unite, another slogan.

A deletion discussion of a state motto is at Talk:live free or die.

Wiktionary has proverbs.

--Dan Polansky (talk) 14:38, 6 August 2017 (UTC)


 * I have a dream failed. Equinox ◑ 16:37, 1 September 2017 (UTC)


 * Much later: A discussion will be at Talk:give me liberty or give me death. --Dan Polansky (talk) 12:04, 4 February 2018 (UTC)

1-0-0
I thought that 1 delete and 0 keep votes over a 1 month period or longer was grounds for the deletion vote to win? (referring to your comment at the full name discussion in RFD) For that discussion, no one even seemed to try to oppose and vote keep, so I guess one could assume everyone agreed it should be deleted. PseudoSkull (talk) 16:35, 1 September 2017 (UTC)
 * My position is that 1-0-0 is no consensus to delete. I don't know how many closures there are that show this to be a common practice, though. Consensus is a general agreement; only one pro-deletion vote does not signify enough of a general agreement to me. Sure, two pro-deletion votes are not all that much better but at least they provide elementary form of independence in that at least two minds expressly agree on the proposed action. --Dan Polansky (talk) 16:42, 1 September 2017 (UTC)

User Page
We have had this conversation before. It is good we are both consistent. - Amgine/t&middot;e

Babel being recommended for user pages
There seems to be a broad support for recommending Babel for user pages: Votes/pl-2007-08/Babel userboxes approved the following text relating to Babel:
 * Language-proficiency userboxes are encouraged, and may be added easily using .

Babel boxes are very widely used in the English Wiktionary.

Babel boxes are useful especially since Wiktionary is a multilingual dictionary project, in which people often contribute in languages which are not their native ones. When an editor sees something that looks like a mistake, checking the nativity of the contributor of the apparent mistake often helps.

The proficiencies indicated can be rather approximate, in my view; errors by one grade (e.g. en-3 vs. en-2) are not all that big of a deal. Even if the Babel boxes did nothing but indicate native tongues, that would already be worthwhile for a multilingual dictionary project.

Posting to user talk pages requests for Babel is something that some recipients do not like. Once recipient recently complained in Beer parlour/2017/August. In that discussion, the consensus seems to be that requests for Babel are okay.

--Dan Polansky (talk) 05:53, 22 September 2017 (UTC)

Editing archived records
@Dan Polansky Yes; I understand your point, even though it was my own paragraph and I wanted to retract my reaction to the annoying tampering with a recent edit then. It was absolutely vital that he reverted my edit in cog, which although mine was inadvertent, it was inexcusably careless, given the stringent guidelines on (my) user page. On another issue, I am very well aware now of the problems encountered whilst correcting inaccuracies in etymologies; since the nearest or most accurate source is usually imported for them as per Wikimedia rules; that is why I added "possibly" as to the questionable P.G. root of sulk and then added a derivative to support it; else I would have removed that root and kept the rest. I do not attempt to fabricate roots, but I have Aspergers and enhanced perception as to what is right and what is not correct in subjects of my interest. Most of the online etymology is inaccurate and Leasnam's is much better here! Andrew H. Gray 06:40, 22 September 2017 (UTC) Andrew
 * (I have edited away and  from the title of the post above.) I responded at User_talk:Werdna_Yrneh_Yarg, the original page of this conversation. If you choose to respond, please respond on your talk page. --Dan Polansky (talk) 07:06, 22 September 2017 (UTC)

Request for Look at
May I ask you to look at my entry at rozumět if it is correct? I would have added Akkusativ (like in German), but it seems that this word requires a Dative in the Slavic languages. --Rasmusklump (talk) 00:21, 30 September 2017 (UTC)
 * Indeed, rozumět komu čemu => dative.
 * The usage note you have added is less than ideal: It is rather wordy for what it does; on a minor point, English does not write "Dative" capitalized, and "du" is not an English word. I have removed the usage note, and instead added "with dative" to the labels, and added a usage example that shows the use of dative. --Dan Polansky (talk) 08:55, 30 September 2017 (UTC)
 * Thanks

"enhanced perception"
Regarding your final comment in this discussion, it's actually quite common for people with Aspergers to have a special, narrow interest in some topic and to be very attentive to details, sometimes to the point of fixation. I agree that Andrew's phrasing was unclear and Aspergers certainly gives no magical sense of what's correct or not, but I think that basically is what he/she meant. Lingo Bingo Dingo (talk) 11:21, 6 October 2017 (UTC)
 * While that might well be the case, and I would not have mocked him childishly, Andrew has such an extraordinary inability to distinguish correct from incorrect that his statement is absurd. —Μετάknowledge discuss/deeds 16:29, 6 October 2017 (UTC)
 * My post may have been less than ideal. Nonetheless, the substance of the matter is that anyone who thinks they have an "enhanced perception as to what is right and what is not correct in subjects of my [their] interest" has a belief that negatively impacts their ability to critically assess their other beliefs and ideas. Lexicographers need to be open to correction, to distrust their hunches, to double check in sources, to seek opportunity to refute what to them seems correct, and generally proceed as if their enterprise aspired to be a science. --Dan Polansky (talk) 17:03, 6 October 2017 (UTC)
 * I quite agree with your statement; and I realised the presumptiveness of mine - it is not what I may regard of my own abilities but what others find is all what counts! What Metaknowledge states is perfectly true as to my earliest edits.  I have found, however, that what my friend who has a post graduate in ancient languages stated as to the most accurate etymologies having been proved to exist in the multi-volume dictionary of Oxford that has been mentioned by others, as being true for myself; because they do not end up with wild connections and false assumptions!  This is why I use a 'point' system with most of my talk page edits, so as not to mislead users.  So, I shall follow through your statements before editing any further on the entry pages.  Being open to correction, whether in etymologies, harmony or any other human subject, is essential! Andrew H. Gray 07:58, 7 October 2017 (UTC)Andrew talk


 * Anyway, the reason I brought it up in the first place was that your previous efforts had me considering to support you for adminship in a future vote but this episode made me, at least for the moment, reconsider and prefer abstaining instead. Though I do not know whether a milder reaction would have got the point across to Andrew; honestly I don't know if anything does. Lingo Bingo Dingo (talk) 13:13, 14 November 2017 (UTC)

Czech participles
A couple of notes on what looks like Czech participles to me:
 * -ící: dělající (adjectival doing). Behavior: adj.: They take adjectival positions in sentences, inflect like adjectives and are modified by adverbs. Example: zpívající pták - singing bird
 * -ný: dělaný, dělávaný (done). Behavior: adj.: They take adjectival positions in sentences, inflect like adjectives and are modified by adverbs. Some sources seem to rank these as adjectives and I have entered them as such, but I have some doubts. One could argue that it is dělán that is the participle, but this does not seem all that convincing; dělán seems to be the predicative form while dělaný the attributive form, of the same participle, although dělaný may be used in the predicative as well.
 * -ní: dělání (nominal doing). Behavior: noun: They take noun positions in sentences, are modified by adjectives and inflect like nouns. These correspond to English -ing nouns/participles. Many of them are entered as nouns.
 * -l: dělal; curiously enough, some Czech grammars describe this as a kind of "participle", but it does not look like a participle to me.

Markup in Wiktionary: It is generally unclear how to enter participles, especially whether to enter them as lemmas with full translation or to enter them as "such and such participle of (Czech base lemma)". I am particularly uncomfortable entering -ící items as full lemmas. Latin participle entries in Wiktionary use Participle part of speech.

What is a participle: It is a derived form that is an amphibian, as it were; it has some characteristics of the base verb but behaves like another part of speech, usually an adjective or a noun. Thus, the pair of dělaný and dělávaný corresponds to the distinction between dělat and dělávat, and thus, a characteristic of the base verb is reflected in the derived adjectival participles. Similarly, usínání and usnutí show verb aspect, from usínat and usnout. Participles may be considered by some to be derivations and by some to be inflections. Either way, the process of their making is very productive and regular one. One thing supporting their being derivations of their base verbs is that they behave like a different part of speech.

Dictionaries: SSJC does not have dělající and dělání; it has dělaný as an adjective.

--Dan Polansky (talk) 10:10, 22 October 2017 (UTC)

A Couple of Module Errors
inflection of now requires specifying the grammatical identity of the inflected form, so hospodární and ztvární currently have module errors. Would you be so kind as to provide the missing parameter(s)? Thanks! Chuck Entz (talk) 01:49, 23 October 2017 (UTC)
 * The module errors are gone. I disagree with the change of the behavior of the template. --Dan Polansky (talk) 17:38, 23 October 2017 (UTC)
 * I have undone the change in Module:form of/templates. If the change can be traced to verifiable consensus, it can be returned back. --Dan Polansky (talk) 06:42, 31 October 2017 (UTC)

Some requests
Thanks --Rasmusklump (talk) 17:07, 26 October 2017 (UTC)
 * 1) oči  shouldn't the term plural be altered to dual? If not please change  očí
 * 2) nasnídat: Is this the pf form of snídat?
 * 3) když Is this term right: Nevidím nic, když nemám Brýle? If not please delete
 * Frankly, I am not really interested in correcting and verifying things after you; I am doing it at this point for the lack of better alternatives. I report things to you when I run into one of your mistakes, too many to my taste; I fear looking at your recent contributions. --Dan Polansky (talk) 11:42, 29 October 2017 (UTC)
 * Ok, but I think a example for if should be given. Does Czech differ between "when" and "if"? Also at pokud and jestli.  And the oči-problem  is a riddle that only can solve a native speaker.  --Rasmusklump (talk) 12:50, 29 October 2017 (UTC)
 * Sure, eventually, an example should be given, but I would like to do other things than add missing example sentences of your choice, and you do not really help anything by misinforming our readers. oči is probably dual (also in cs:oči), but I would have to research that and confirm what "dual" means; the non-dual would be oka. The distinction between "dual" and "plural" is not really important one, I think; "eyes" is "oči" in reference to animal eyes where "eye" is "oko", and the term "dual" is actually more confusing and specialist than "plural"; broadly speaking, "plural" is not a wrong description of "oči". --Dan Polansky (talk) 12:59, 29 October 2017 (UTC)

Indefinite or nominal Czech adjectival forms
The following Czech adjectival forms are distinct: mlád (vs. mladý), stár (vs. starý), zdráv, nemocen, chud, bohat, smuten, vesel, šťasten, dlužen, hoden, ženat, schopen, sláb, vinen.

They find above all literary use, as far as I can tell, but not all of them; vinen is an example of a form that seems to find a predominant use. SSJC says, in vinný, about vinen: "(v přísudku a v doplňku zprav. v tvaru jmenném vinen, vinna,...)"

Kunz 1873 calls them "neurčitá" (indefinite). Bečka 1945 calls them "tvary sklonění jmenného" and "přídavné jméno jmenné". Kopečný 1954 calls them "tvary [...] neurčité neboli jmenné". Czech Wikipedia calls them "jmenné tvary".

Dictionaries: smuten is neither in PSJC nor in SSJC.

For the section heading, I have translated "neurčitý" as "indefinite" and "jmenný" as "nominal".

Wiktionary markup: none so far. We could use "indefinite form of X" or "nominal form of X". vinen has "predicative form of X"; I am not sure that is correct or customary.

Links:
 * Česká přídavná jména
 * Česká přídavná jména
 * Česká přídavná jména

--Dan Polansky (talk) 17:36, 29 October 2017 (UTC)

Separate synonyms sections
Per long-standing practice and WT:ELE, synonyms should have a separate section, usually ====Synonyms====. A recent practice pushed by CodeCat/Rua is to place synonyms under definitions, similar to quotations of use. Simple English Wiktionary has a similar practice; there, the synonyms are hidden by default, just like en wikt quotations.

Discussions:
 * Beer_parlour/2017/February
 * Beer_parlour/2017/May

The poll above shows 2/3-consensus or preference for "placing various nyms directly under corresponding senses". We do not have a vote yet; a vote would enable broader participation by increasing the visibility of the proposal.

--Dan Polansky (talk) 07:46, 17 November 2017 (UTC)

Entry cat as it is now is an example of presentation mess that arises when the subsensing that I objected to is combined with synonyms directly under senses, which I objected to as well. When quotations are hidden (default), this is what I see:


 * 1) An animal of the family Felidae:
 * 2) A domesticated subspecies (Felis silvestris catus) of feline animal, commonly kept as a house pet.
 * 3) Any similar animal of the family Felidae, which includes lions, tigers, bobcats, etc.
 * 4) A person.
 * 5)  A spiteful or angry woman.
 * 6) An enthusiast or player of jazz.
 * 7)  A person (usually male).
 * 8)  A prostitute.
 * 1) An enthusiast or player of jazz.
 * 2)  A person (usually male).
 * 3)  A prostitute.
 * 1)  A prostitute.

Incredibly, "Synonyms" is boldface and is a link at the same time; making "Synonyms" non-boldface and without a link would be an improvement, but would still create pretty bad presentation. Finding definitions and staying focused on them in this mess is quite a challenge. Later: That's an overstatement: The definitions are right there, not hard to find. Something's very wrong with the presentation, but what is it?

--Dan Polansky (talk) 07:56, 19 November 2017 (UTC)
 * In my opinion, something that is very wrong with the presentation is this: If I care only about definitions and not synonyms, the synonyms are a nuisance because I have to make the effort (albeit small) to ignore them while looking for definitions. The same can be said if I care only for synonyms and not definitions. Incidentally, sometimes I like using senses and subsenses, which you said above you object to. --Daniel Carrero (talk) 05:25, 10 December 2017 (UTC)

Later: It's not only about synonyms but also hypernyms and hyponyms: blanquette is an example. There, each of five definitions has at least one synonym or hypernym sub-item. Unsightly. --Dan Polansky (talk) 07:08, 21 April 2018 (UTC)

Anon adding pronunciation in Czech entries
is adding pronunciation to Czech entries. From a random sample, they look fine. Sometimes, they add declension, as in. --Dan Polansky (talk) 12:46, 18 November 2017 (UTC)

Tbot - bot creating entries based on translation tables
User:Tbot was a bot run by Robert Ullmann that created entries based on translation tables, checking whether the entry exists in the non-English Wiktionary of the entry. The created entries got tbot entry badge of shame and were placed into a category, the root of which is Category:Tbot entries. The category tree is broken down by year and month (e.g. CAT:Tbot entries July 2009‎) and language (e.g. CAT:Tbot entries (Swahili)‎). The vote approving this kind of entry creation was Wiktionary:Votes/bt-2007-12/User:Tbot creating FL entries, passing 12:2.

Tbot also entered pronunciation, it seems, possibly from non-English entries.

Tbot source code in Python was originally in User:Tbot/code, and was later split into User:Tbot/code/createflw, User:Tbot/code/script, and User:Tbot/code/tbot.

The count of entries still carrying the badge of shame is 5044, per AWB "What transclued page" on tbot entry.

Some people complained of Tbot entries. Nonetheless, I have never seen any numbers of mistakes it created, nor a list of 10 such mistakes. In order for Tbot to insert a mistake (carrying a Tbot badge of shame), two errors must have combined: 1) wrong translation in the translation table, 2) wrong entry in the non-English Wiktionary. User:Tbot page has details on the checks. Complaints are e.g. in User talk:Tbot.

Some people liked Tbot: in 2009, they requested increased language coverage on User talk:Tbot.

For some contributors, Tbot entries would ease work: these contributors would only check the accuracy of those entries and then remove the Tbot badge of shame. That involves less typing or templating than creating a new entry from scratch, with no need to learn about the proper format: removing a badge of shame is a format-agnostic operation.

Since, for many languages, we now have templates linking to online monolingual dictionaries, a modern-day Tbot could make a third check, against the reference monolingual dictionary. A modern-day Tbot could automatically insert a Further reading section pointing to the monolingual dictionary, adding further value to the entry.

User:Ruakh/Tbot.js is a script of the same name, allowing contributors to click on a green item in translation table, which results in a new entry creation opening, filled with relevant headings, as if done by the Tbot. The difference against the original Tbot is that the entry is not created before the user manually clicks on the green link.

--Dan Polansky (talk) 14:48, 19 November 2017 (UTC)
 * It is difficult to find old deleted and cleaned up Tbot entries, and many were fixed early on, so one gets a biased view. Swahili Wiktionary is of extremely poor quality, and a very problematic editor had a habit of adding Swahili "translations" where they didn't belong. Our only Swahili editor at the time was Robert Ullmann, who ran Tbot and was not good enough at Swahili to notice the problems he was creating. Many entries that were created were unattestable, like Talk:Santalusia or Iraki; others were completely wrong, like Talk:sabibu and shabaha. —Μετάknowledge discuss/deeds 20:16, 9 December 2017 (UTC)
 * Thank you. Tbot could be enabled and disabled on a per-language basis. As said, "a modern-day Tbot could make a third check, against the reference monolingual dictionary"; this would reduce the error rate at least for term existence and spelling. --Dan Polansky (talk) 20:34, 9 December 2017 (UTC)

Musicians and bands
Musicians, bands and music groups currently included include AIH (Architecture in Helsinki), Apop (Apoptygma Berzerk), Beatles, COB (Clive's Original Band), Foos (Foo Fighters), Genesis, Led Zep, NKOTB (New Kids on the Block), Prodge (The Prodigy), RATM (Rage Against The Machine), RHCP (Red Hot Chili Peppers), Rolling Stones, Stones (Rolling Stones), Straits (Dire Straits), Zep (Led Zeppelin). --Dan Polansky (talk) 16:27, 19 November 2017 (UTC)


 * TAFKAP, Moz/Mozzer, arguably Madge. Equinox ◑ 18:38, 9 December 2017 (UTC)

Favicon
Our favicon is still based on the old logo. Do you have any interest in making a favicon based on the new logo? —Μετάknowledge discuss/deeds 23:49, 26 November 2017 (UTC)
 * @Μετάknowledge: A new favicon could be made using the same method I used to make the logo: remove the tiles. Thus, we would have to find the favicon used in fr wikt, and see whether by removing the tile we can get something that can be used as en wikt favicon. It may turn out that the result will be unsatisfactory. I am not enthusiastic about doing that myself, in part since I like the current favicon, but if no one steps up to do it, I may end up removing the tile and seeing whether the result is appealing. --Dan Polansky (talk) 12:28, 9 December 2017 (UTC)


 * Here's the red W at 16x16 and 32x32 in an ICO file, if that's any use: Equinox ◑ 12:33, 9 December 2017 (UTC)
 * Looks pretty fine, doesn't it? I have to note that the gray background of the browser tab handle in which I see the icon is of similar gray shade as the background used in the icon; it may looks quite different when a different background is used.. --Dan Polansky (talk) 12:37, 9 December 2017 (UTC)


 * I left the background in because I don't have the tools to do nice alpha blending at the edges! But our existing favicon also includes background colour (and I suppose removing it would make it disappear on a red background). Equinox ◑ 12:41, 9 December 2017 (UTC)
 * My guess would be that Gimp can do such alpha blending, maybe with a plugin, but just a guess. I do not know how to do it with Gimp, though; I would have to try to figure it out. https://docs.gimp.org/en/plug-in-colortoalpha.html looks promising. Either way, thanks! --Dan Polansky (talk) 12:51, 9 December 2017 (UTC)

Bolding
You have convinced me: henceforth, I shan't use bolding in the quotations I add. --Barytonesis (talk) 19:29, 8 December 2017 (UTC)