Template talk:derv

Other languages

How will this work on words derived from other languages, or what template will be used for those? — lexicógrafo | háblame — 00:54, 29 September 2010 (UTC)

In principle, it would work. As Raif'har had another approach and some languages have their own (especially agglutinative ones), I am not seeking to apply it outside the most obvious English cases. There are some fundamental differences between the one component-one template approach and that taken by Raif'har and by the authors of {{confix}}, {{prefix}}, temp|suffix}}, and {{compound}}. Their approach saves typing, but doesn't give much flexibility. The lack of flexibility limits their applicability. I would like to demonstrate the value of the single-component approach in areas where I have a good grasp of the issues, ie, English terms derived from English terms. BTW, I have suggested it to an Esperanto contributor. DCDuring TALK 17:16, 29 September 2010 (UTC)

I think you may have misunderstood me — I was asking about words in English derived from other languages, e.g. French or Latin. — lexicógrafo | háblame — 17:43, 29 September 2010 (UTC)

Raif'har was proposing two templates, one for native (same-language) derivations and another for cross-lingual ones. I have applied this template to, eg, derivations from (deprecated template usage) refer, not, I hope, to terms that are borrowed/derived from other languages in which "refer" is also a base, for which {{base}} would be and has been used in English etymology sections. DCDuring TALK 18:07, 29 September 2010 (UTC)

head= parameter

A parameter indicating that a given component is the "head" of a compound (open or closed) would enable one to distinguish among types of related terms. For example, "load leveler" does not seem nearly as "related" to "carload" as "truckload", "boatload", etc, do. Marking all the latter as "headed" by (deprecated template usage) load could segregate them for distinct treatment (categorization). Values might be "1", "right" (a good default), "left", and "0" (equivalent to null). The one-component approach means that it is possible for more than one component to be marked as a "head", though this seems uncommon, even impossible, and undesirable. DCDuring TALK 17:36, 29 September 2010 (UTC)

sense= parameter

For really productive components a breakdown of a given Etymology/Pos may need to go further to distinguish certain senses or ad hoc groupings of senses. DCDuring TALK 17:36, 29 September 2010 (UTC)

type= parameter

In some cases, it may be desirable to include open-compound derivatives or affixed derivatives separate from others. Possible values could be "open", "suffix", "prefix", "hyphen". DCDuring TALK 17:36, 29 September 2010 (UTC)

Deleting the template

I for one would like to see this template deleted. Its use instead of the prefix template is curious, as in "{{derv|en|odor}} {{suffix||ous}}" instead of "{{suffix|odor|ous}}". If the community actually agreed on having categories for derived terms, the way to go would be to modify prefix template to place entries in these categories, instead of violating in this roundabout manner the way in which the template suffix is designed to be used. I have started a BP discussion here: Wiktionary:BP#Deleting_categories_for_derived_terms, February 2011. --Dan Polansky 08:20, 16 March 2011 (UTC)

Deletion debate

The following information passed a request for deletion.

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

---

Template:derv

Seems to me the best way to have a conclusion to this 'debate' is to nominate this for deletion. I haven't made up my mind yet; my instinct is to delete it and just list derived terms in entries; a category is superfluous. We also have {{der-top}} (mid and bottom) for long lists of derived terms. Mglovesfun (talk) 13:00, 17 March 2011 (UTC)

delete. The idea of a category just doesn't strike me as good. One disadvantage a category has is that it doesn't allow for listing derived terms which haven't been created yet. -- Prince Kassad 14:49, 17 March 2011 (UTC)

Would there be a way to somehow combine both? So that it lists both the words in the category and the words that have been added to the entry in one list? —CodeCat 14:53, 17 March 2011 (UTC)

Have a bot follow the recent changes, that's all I can things of

Oh, weak delete. The categories seem redundant to just listing the entries as red or blue links. Mglovesfun (talk) 14:55, 17 March 2011 (UTC)

There is nothing mutually exclusive about the category and a manual listing. It is a simple, but not essential, maintenance task to derv-ify the blue links in a manually created list. The SoP red-links should probably be turned black as they are to be kept as derived terms at all. They are possible raw material for some list of collocations, which might be useful to assess the completeness of coverage of our definitions.

Delete. If anyone intends in all earnest to use "derv", let them explain how this is going to play together with "suffix", "prefix", "affix", "compound", "term" and other templates already in use. Furthermore, the derv-procedure presupposes that there are going to be as many derivation categories as there are terms that have at least one derived term, including a category for every adjective from which a "-ness"-noun is derived, and a category for every verb that is -able (see this revision). See also Wiktionary:BP#Deleting_categories_for_derived_terms, February 2011. --Dan Polansky 15:03, 17 March 2011 (UTC)

{{term}} is marginally relevant. In a long discussion in which few participated, the experimental rollout scheme was discussed and the desirability of deferring integration with existing templates was discussed. See Wiktionary:Beer_parlour_archive/2010/September#.22Synchronic.22_and_.22diachronic.22_etymologies.

Some numbers: "Category:English words suffixed with -ness" has 1,789 members; "Category:English words suffixed with -able" has 779 members; these are at least 2568 words and thus 2568 derivation categories to be created. We could adopt a policy that not all of these derivation categories should be created, but that is not what the creator of "derv" template had in mind, as he has created "Category:English words derived from: austenite", a category that can feature only a few members such as "austenitize". --Dan Polansky 15:46, 17 March 2011 (UTC)

As I've mentioned elsewhere, IMO we do not want a "derived from" category for every term from which something is derived. (Consider Middle English withdrawen, for example, from which modern English withdraw is derived.) Thus, I do/would not like template:derv to be used in all generality. Does that mean it should be deleted? Not necessarily, but if not then its use should be restricted. (Or, to state the contrapositive, if its use is not going to be restricted, then it should be deleted.)​—msh210℠ (talk) 15:34, 17 March 2011 (UTC)

Given that this will be used only a handful of times on any given entry, we could use #ifexist: to have it categorize an entry if and only if the category already exists. I agree with above commenters that we don't want "Category:English words derived from: austenite" and "Category:English words derived from: withdrawen". —Ruakh_TALK 16:13, 17 March 2011 (UTC)

I agree with msh210 here. We ought only to use {{derv}} for those terms for which we want derived-terms autocategories (like partite, which functions much like a Latinate equivalent of the Grecian -merous); in other cases, we can just keep using {{term}}. On that basis, I vote we keep this template. However, the format of its categories should be standardised with the format of the categories created by {{prefix}}, {{suffix}}, &c. BTW, I've not had the Beer parlour on my watchlist for a long time, so I'm a bit out of the loop; could someone please give me a summary of developments and/or links to prior discussions? Thanks. — Raifʻhār Doremítzwr ~ (U · T · C) ~ 17:51, 17 March 2011 (UTC)

I recommend to your attention such categories as Category:Mandarin words derived from: yuánshēngtài, Category:Mandarin words derived from: shōuyīnjī, Category:Mandarin words derived from: bànzhǒng, or Category:Mandarin words derived from: jīngshén. How is an added #ifexists to the template going to prevent creation of such categories? Once they are created, the template will place entries to them. What is the proposed regulation that will prevent creation of categories that contain only few members, if there is such a regulation at all? Is it okay to have a mixture of use of categories and non-categories in the DT sections, which will necessarily follow from some lists of derived terms not having their derivation categories? Looking at partite, I see a nice section for derived terms containing 20 terms, some of them redlinked; how are the categories going to improve the entry? The list was created with a single edit of a person who knew what he was doing: see this diff. Where is the evidence that the categories, a technique that makes things more complex, are actually going to simplify the work? --Dan Polansky 19:25, 17 March 2011 (UTC)

Also worthy of a look is Category:Mandarin derived terms, a category that now contains 698 subcategories. You can try to randomly click into the subcategories and see how many of these categories contain one or very few items. --Dan Polansky 19:28, 17 March 2011 (UTC)

A small report: I have taken the list of the first 200 subcategories from Category:Mandarin derived terms, and selected only those that have zero further subcategories. Here is the list of the selection, with the number of elements in brackets: biàn (verb) Etymology 变 (1 e), biànzhèng (1 e), biéchū (2 e), biān (affix) Etymology 边 (1 e), biānjí (1 e), biānzhī (1 e), biāozhǔn (1 e), biǎo (affix) Etymology 表 (1 e), biǎojué (1 e), bu (infix) Etymology 不 (3 e), bà (noun) Etymology 爸 (1 e), bàn (noun) Etymology 瓣 (1 e), bàn (number) Etymology 半 (3 e), bàn (verb) Etymology 伴 (1 e), bàn (verb) Etymology 拌 (1 e), bàndǎotǐ (1 e), bàng (noun) Etymology 棒 (2 e), bànggǎn (1 e), bànshì (1 e), bànzhǒng (1 e), bào (verb) Etymology 报 (1 e), bào (verb) Etymology 抱 (1 e), bào (verb) Etymology 爆 (1 e), bàofā (1 e), bàolì (1 e), bàopò (2 e), bái (adjective) (3 e), bái (affix) (7 e), báifèi (1 e), báihuà (1 e), básī (1 e), bèi (affix) (1 e), bèiwàng (1 e), bèiyòng (1 e), bì (affix) (1 e), bìhé (1 e), bìng (noun) (3 e), bìnyí (1 e), bìsè (1 e), bí (affix) Etymology 鼻 (3 e), bíhuà (1 e), bíkē (2 e), bù (adverb) (1 e), bùdé (1 e), bā (noun) (1 e), bā (suffix) (1 e), bāo (affix) (1 e), bāo (noun) (1 e), bēi (affix) (1 e), bēi (noun) (1 e), běi (adjective) (3 e), Běijīng (2 e), bīng (adjective) (1 e), bīng (noun) (1 e), bǎihé (1 e), bǎihéhuā (1 e), bǎn (adjective) Etymology 板 (1 e), bǔ (verb) (1 e), bǔxié (1 e), chuán (noun) (1 e), chuāng (noun) Etymology 疮 (1 e), chàng (verb) (3 e), chàngfǎ (1 e), cháng (adjective) (1 e), chén (affix) (1 e), chéng (affix) (2 e), chéng (noun) (1 e), chún (noun) (1 e), chāshāo (1 e), chē (noun) (2 e), chī (verb) (1 e), chīfàn (1 e), chǎng (noun) Etymology 厂 (1 e), chǎng (noun) (2 e), chǐyóu (1 e), chǔlǐ (1 e), chǔn (adjective) (1 e), cài (noun) (2 e), cào (affix) (1 e), càodàn (1 e), cí (affix) (1 e), cí (noun) (2 e), cídiǎn (1 e), císù (1 e), cā (verb) (1 e), cāomǐ (1 e), cāxié (2 e), de (affix) (3 e), diàn (noun) Etymology 电 (1 e), diàn (noun) (2 e), diànshì (3 e), diàntī (1 e), diànzǐ (1 e), diào (noun) Etymology 调 (1 e), diào (noun) (1 e), diǎn (affix) (1 e), diǎnxin (1 e), duān (affix) Etymology 端 (1 e), dà (adjective) (1 e), dàfǎ (1 e), dàlù (1 e), dàn (noun) (1 e), dàng (affix) Etymology 档 (1 e), dào (affix) Etymology 悼 (1 e), dào (affix) Etymology 道 (1 e), Dào (noun) (3 e), dì (noun) Etymology 地 (1 e), dì (prefix) (4 e), dì-sì (1 e), dì-sān (1 e), dì-yī (1 e), dì-èr (1 e), dìlǐ (1 e), dìlǐxué (1 e), dìzhì (1 e), dìzhìxué (1 e), dòng (verb) Etymology 动 (1 e), dòngwù (1 e), dòngwùxué (1 e), dòu (noun) (2 e), dòubàn (1 e), dāi (adjective) Etymology 呆 (1 e), dǎn (noun) (1 e), dǎotǐ (1 e), dǎpò (2 e), e- (1 e), fàn (noun) (7 e), fànbāo (1 e), fànguō (1 e), fán (affix) (1 e), fántǐ (1 e), fèn (affix) (2 e), fén (noun) (1 e), Fó (noun) (3 e), fùyìn (1 e), fú (affix) (1 e), fúshǒu (1 e), fútī (1 e), fācái (1 e), fān (affix) Etymology 番 (1 e), fāng (affix) Etymology 方 (1 e). --Dan Polansky 19:34, 17 March 2011 (UTC)

These categories reduce human-maintained redundancy, which is A Good Thing™ in principle; in concrete terms relating to this case, it means that whenever someone creates an entry which derives from some other term for which we also have an entry, he doesn't have to go to that entry to add that new entry to the list. The contrary also holds, in that if a term ever fails RFV or RFD and is deleted, no one has to go to another entry to delete the red link from its Derived terms section. The advantage of using these categories is the same as the advantage of using the categories created by {{prefix}}, {{suffix}}, &c.; and accordingly, those advantages are most accentuated in the cases of terms which are highly productive in forming new derivations.

As for the specific regulation for restricting category creation, why not a minimum number of members? Setting the minimum as low as four would lead to the deletion of all but three of the categories in your above selection. What would you consider a reasonable regulation? — Raifʻhār Doremítzwr ~ (U · T · C) ~ 19:52, 17 March 2011 (UTC)

So it means that you are okay with there being a laerge number of categories with 5-10 entries, right? I am not sure what I would consider a reasonable regulation, but having the limit of minimum of 4 entries is already better than nothing, although I dislike the idea of there being thousands of small categories with less than ten members. I have no problem with what you call "human-maintained redundancy" when it comes to lists of derived terms. I have created such lists myself. I for one just want to see all these categories gone, but I admit that having a fixed limit on the minimum number of entries is an interesting proposition that reduce the number of tiny categories.

Coming back to {{derv}}, however, I do not see what job it does that cannot be done by {{prefix}}, {{suffix}}, {{compound}}, {{term}}, etc. If the community agrees to have such derivation categories, these templates can be adjusted to automatically file the entries into the categories. From what I can see, {{derv}} was created to circumvent the necessity of community support for the modification of the mentioned templates. --Dan Polansky 20:14, 17 March 2011 (UTC)

Nice rhetorical miscategorization of an experiment as circumvention. There was a full discussion of the (manual) experiment at Wiktionary:Beer_parlour_archive/2010/September#.22Synchronic.22_and_.22diachronic.22_etymologies. The results of said experiment would be visible and useful if you hadn't mass-deleted them. It would be a courtesy for you to mass-revert the depopulation of the English categories. The experiment was done mostly at the bushy end of the tree to keep it small, to determine the level of effort required for mass implementation.

As you are seeking to delete all derivation-based categories (except suffix and prefix related ones, which are curiously absent from your complaints about small-sized categories), it seems your advancing the empty proposal to use the existing templates for the purpose is also mere rhetoric.

You had the chance to participate in the discussion of this

{{term}} is not only for etyma.​—msh210℠ (talk) 20:18, 17 March 2011 (UTC)

For that reason alone I would prefer that this template is kept. It doesn't have to do anything that {{term}} doesn't, in principle. But it allows us to keep track of which links are derivations and which aren't. —CodeCat 20:29, 17 March 2011 (UTC)

(unindent, @msh210) Okay, you mean that {{term}} is also used to format terms from which the term of the entry is not derived such as cognates, right? But this should be easy to fix using a new parameter of (deprecated template usage) term, right? In any case, derv was used by its creator to create such markups as "{{derv|adsorb}} {{suffix||able}}", circumventing the native function of {{suffix}} called by "{{suffix|adsorb|able}}". --Dan Polansky 20:33, 17 March 2011 (UTC)

Nice rhetorical use of the pejorative "circumvent", Dan. Also nice rhetorical mischaracterization of the rationale for not proposing a change to {{suffix}}. I was interested in exploring the possibilities of categorization for creating derived terms lists without requiring changes to existing templates such as {{suffix}}, {{prefix}} and {{compound}}. Those template might be worth altering in such a way as to work more tightly with {{derv}} and its relatives, but not until the overall desirability of the concept was established.

About small categories: We have many. The categories that would be introduced by universal use of {{derv}} would have a size distribution that would include many small categories. The number could be reduced without great loss by deferring creation of PoS-specific derivation categories. Small derivation categories need not actually be created if there is reason to believe that they cost more in editor, resource, and download-lag terms than the manually created derived-terms lists. I am not aware of any evidence or arguments to that effect. DCDuring TALK 17:52, 24 March 2011 (UTC)

{{term}} is also used outside of etymology sections. I know I've seen it in 'manually'-written form-of sense lines (using {{non-gloss definition}}); probably elsewhere, too. (Otoh, a form-of sense line's referents are also etyma.) But yes, this should be doable with a parameter. AfaIct, that's a good way to go: get rid of {{derv}} and add a, say, derv parameter to {{term}}, {{suffix}}, etc. which takes values 1, indicating that the entry should be categorized in an etymon category based on the {{{1}}} parameter of those templates (or {{{2}}} parameter of {{prefix}}), and foo, indicating that the entry should be categorized in an etymon category based on foo.​—msh210℠ (talk) 20:56, 17 March 2011 (UTC)

• @Dan Polansky: I am not circumventing the native function of {{suffix}}/{{prefix}} called by "{{suffix|x|y}}/{{prefix}}|x|y}}", the problem is they don't suitable for Mandarin entries. Ddpy 08:49, 18 March 2011 (UTC)

@Doremítzwr: Are you planning to start a discussion and a vote or something on the limit of 4 entries for the derivation categories? Or is Ddpy (talk • contribs) to be allowed to create further 1000 tiny categories? Do you at least plan to tell the user that you disagree with his creating the tiny categories, if you actually disagree? --Dan Polansky 20:33, 17 March 2011 (UTC)

"Tiny category" can grow up though it is tiny now, such as the above "tiny category" example you mentioned (Category:Mandarin words derived from: jīngshén) is growing up. Ddpy 01:44, 18 March 2011 (UTC)

@ Dan Polansky (20:14, 17 March 2011): I'm fine with the minimum limit being ten instead, if you'd prefer. The point is that the autocategory is more useful in the case of a term currently with six derivations, but which sees a lot of protologistic coining, than it is for a term with sixty derivations, but whose productive days have long since passed; again, the same applies to affixes. IMO, it would be a more rational template structure for {{prefix}}, {{suffix}}, &c. to be subsumed into {{derv}} &c.; {{derv}} thereby becomes, in effect, the "{{infl}}" of the world of autocategorising derivation templates.

@ Dan Polansky (20:33, 17 March 2011): I believe this is the start of the discussion on what the minimum number of members of a category ought to be to justify its existence. As I said, I'm not too bothered if the number is four, or ten, or whatever; once we hit upon a number that is agreeable to what looks like a consensus, then we can go ahead and formulate that newly-established convention into policy. BTW, these categories only give us half their potential utility ATM; where are the templates herefor equivalent to {{prefixsee}}, {{suffixsee}}, &c.?

@ Ddpy (01:44, 18 March 2011): Those tiny categories aren't very useful as they are. I suggest that, in future, you create entries for several terms that will populate such a category before creating the category itself. Your position will be strengthened if you no longer need to defend the existence of innumerable absurdly small categories. Where a term has too few derivations, please satisfy yourself with manually-generated Derived terms lists.

@Ddpy. That way you could enjoy the experience of having your work mass depopulated without explanation. DCDuring TALK 17:52, 24 March 2011 (UTC)

 — Raifʻhār Doremítzwr ~ (U · T · C) ~ 15:26, 19 March 2011 (UTC)

• Delete. It's a total waste of time. ---> Tooironic 10:14, 18 March 2011 (UTC)

  Are you saying as implemented in Mandarin or are you speaking about all languages and all implementations? I have no knowledge of how it would or would not work as a practical matter in Mandarin or whether it is to the taste of the majority of active contributors or users in Mandarin. DCDuring TALK 18:18, 24 March 2011 (UTC)

  I can only speak about my speciality, Mandarin. If we implement this categorisation for this language we will have a limitless number of tiny categories for EVERY SINGLE character in Chinese. Can you imagine the chaos that would create? Even if this could be done automatically (which doesn't appear to be the case at the moment) it's going to get awfully messy, especially considering we have to synchronise three different scripts (simplified and traditional, and pinyin - the latter of which has homophones galore!). ---> Tooironic 20:45, 1 April 2011 (UTC)

  I don't know anything about Mandarin, so I defer to whatever contributors' group keeps things from running off the rails there. If that group is you, de facto, I defer to your judgment. I never intended the {{derv}} experiment to occur in any language besides English without the support of a consensus of contributors in that language unless there were an all-Wiktionary consensus. I had never intended the English implementation to proceed much beyond what I had done. Others may be interested in exploring other small-scale experimental implementations, including some across languages. Once there are a few folks who have experimented with it, then it might be appropriate determine whether and how to deploy it (or make other changes). I would be happy to add yet another banner to the template to strongly discourage use in Mandarin and discourage use in CJKV generally. Also any other language for which someone might be in a position to represent an effective consensus of contributors in the language. DCDuring TALK 23:09, 1 April 2011 (UTC)

Keep It may be a waste of time for those not interested in derived and related terms, but it offers a system for making sure that we have complete lists of these without cluttering up entries and taking less time to complete than the tedious process of hand-constructing such lists. It is especially true for multi-word entries, for which one could populate multiple derivation categories while opening the etymology section but once. Alternative means for finding derived and related terms using search are quite tedious especially when the term does not begin the entry name or for compounds if there is no etymology. I wish that I could point to the examples of the deployment of these on a test basis, but that experiment has been eradicated by Dan Polansky. If you would like to see the experiment as applied to a series of English terms, please ask him to revert the changes. DCDuring TALK 17:52, 24 March 2011 (UTC)

I have left the following English categories non-empty, so people can get an impression of your experiment: Category:English_words_derived_from:_load_(verb) and Category:English words derived from: load (noun). (See also Special:WhatLinksHere/Template:derv to find out.) I do not see how, if this is an experient and a failed one by your own admission, it needs a host of other example categories. You have not protested at my talk page after I have completed the removal of {{derv}} from most English entries, neither have you protested in Beer parlour, nor has anyone else protested AFAIK. In retrospect, I am hesitant whether my having removed {{derv}} using the sort of unbureaucratic method typical of the local oligarchy was a good thing. However, it still feels wrong that a small minority can introduce a bad practice without a consensus, but a supermajoritarian consensus is needed to remove the bad practice. In any case, if your remark about a failed experiment was meant it earnest, it was above all your business to clean up the mess that you have left behind. --Dan Polansky 23:21, 24 March 2011 (UTC)

The mere fact that you think it to be a mess is immaterial. Just as my similar evaluation of the failed Wikisaurus experiment is immaterial to its deletion. I did not immediately protest the mass deletion out of a desire to avoid a massive display of hostility when, after all, you might have just been off your meds or otherwise temporarily indisposed. Now that it seems to be a more chronic condition, I also might ask why you did so little work to explore the background, why you looked at my effort to throttle back the user who was deploying {{derv}} in Mandarin as a license to mass-delete, and why you didn't take the trouble to confirm with me, with CodeCat, with Doremitzwr, or with Ruakh, all of whom had contributed in some way to the experiment.

One advantage of delaying my response is that I can now see that I should not take all this too personally. Judging from a larger pattern of hectoring behavior, you must be feeling oppressed by the oligarchy. If you view this collegiality as an oligarchy, that is a shame. I don't feel like an oligarch, personally, so if feeling like one would be a good experience, I would welcome some specific information that would make me feel like one.

IMHO, large-scale deployment of {{derv}} would unquestionably need consensus or a vote. I have tried to anticipate objections and have run a mostly inconspicuous experiment. Part of the purpose of the experiment was to get a preliminary estimate of the number of small categories that would be created. As I am not a technological adept, I do not view myself as likely to anticipate all possible technical issues or solutions (such as Ruakh's suggestion about testing for prior existence of a category to avoid the horror or category redlinks). As I am not Kreskin, I also could not anticipate what would cause strong visceral reactions.

We have many bad practices, some of great age. Some of my favorite are: providing translations for definitions with obsolete wording, using obsolete and archaic wording in translations taken from obsolete dictionaries, providing encyclopedic definitions, using slang senses of highly polysemic words as definiens, and building Wikisaurus on words instead of concepts. Most of those can't be said to be experimental. DCDuring TALK 01:46, 25 March 2011 (UTC)

I am responding to a few points; drop me a note on my talk page if you want a full response, which I think is better suited for your talk page.

I do not believe you wanted to send derv and the categories to a vote: you have created not a single vote AFAIK; you have opposed more votes than is reasonable; and you have proposed not a single improvement of wording of a vote AFAIR, but here my memory can easily fail me. I have not much reason to believe you are sincere in your talk of consensus and votes; your past RFV and RFD behavior points otherwise. You have a track record of pushing unvoted-on controversial parts of CFI. You have claimed "we don't do names of specific entities" long after a vote has shown otherwise. You have created derv out of a Beer parlour discussion whose title did not in any way suggest that people were about to begin replacing "Derived terms" sections with categories, a discussion on whose first part I took part, and which had a rather different subject than later parts of the discussion. Replacement of DT sections with categories is a big thing; I do not see why heading towards it requires an experiment in mainspace rather than in userspace. The alleged experiment, if left unchecked, was about to grow into yet another disunity that can no longer be removed by a supermajoritarian vote, as it has become entrenched.

My having removed derv from most English categories was, admittedly, a rash response to the creation of a large number of tiny Mandarin categories, and to vote-haters ruining a vote with arguments that things should be handled informally without votes in the mainspace, under the motto not pushed by me that "a mass change that is left unopposed is okay". Note, however, that the scale of the mass change was rather small: it involved around 450 entries.

If people want me to roll back the mass removal of derv, I can do it; nonetheless, I have no principle at hand that tells me that the use of derv is the status quo rather than the absence of derv being the status quo. In any case, I am willing to give in to a plain majority in an executive vote that wants me to revert the mass removal. --Dan Polansky 09:00, 25 March 2011 (UTC)

Fwiw, I agree with Dan re status quo and have no wish for him to revert his reversion at this time.​—msh210℠ (talk) 15:51, 25 March 2011 (UTC)

Oh, and another impulse for my rash removal was that the Chinese editor Ddpy started to recklessly create tiny English categories, such as Category:English words derived from: neuroma which has one member: "acoustic neuroma". See also #Category:English words derived from: neuroma. --Dan Polansky 09:32, 25 March 2011 (UTC)

I have found a discussion that shows that the use of derv and moving derived terms outside of entries has already been controversial: User talk:DCDuring#Replacing derived term lists with categories, October 2010, 3300 words. --Dan Polansky 11:54, 25 March 2011 (UTC)

Ignoring all the bitching and bluster, the main complaint against {{derv}} seems to be its use in creating tiny categories. A solution to that has already been suggested, viz. that we require a minimum number of members to justify the existence of a derivation category. Now, {{derv}} behaves in a manner materially identical to {{prefix}}, {{suffix}}, and {{confix}} — all three of which can also create very small categories. I have two questions for contributors to this discussion: 1. Does the "minimum number of members" regulation dispel your opposition to {{derv}}, and if not, why not? 2. On what basis might {{prefix}}, {{suffix}}, and {{confix}} be accepted whilst {{derv}} is nevertheless rejected? Answers please. — Raifʻhār Doremítzwr ~ (U · T · C) ~ 13:30, 25 March 2011 (UTC)

My objections against {{derv}}:

• o1: It introduces the practice of removing derived terms from the entries in the main namespace, and moving them into categories.
• o2: Its unregulated use leads to a creation of great many tiny categories. This you have addressed by setting a minimum limit. But this creates a subobjection:
  • o2a: If a limit is set, some derived terms land in categories while other derived terms land in lists of derived terms, creating less unified user experience.
• o3: The way it was deployed, it replaces the use of {{suffix|adsorb|able}} with {{derv|en|adsorb}} {{suffix||able}}

Unlike the words typically tagged using "derv", prefixes and suffixes do not tend to create tiny categories, so here is one significant difference.

A seriously meant attempt to make a useful deployment of derv should clarify above all what happens to the sections for derived terms, and it should first show that at least a plain majority of editors is okay with whatever is planned to happen with the section. The burden of demonstrating consensus or at least a plain-majority support for this radical deviation from the common practice lies on the shoulders of the supporters of this template.

What I find annoying about the whole method of procedure around "derv" is that I am forced to discuss and analyze something that I do not want to discuss and analyze at all: I just want to oppose it in a poll or a vote. The deployment of "derv" was set up in such a way that unless someone use their resources to set up a conflict and look like an argumentative moron, it becomes entrenched. I find this way of procedure really unfair. --Dan Polansky 13:50, 25 March 2011 (UTC)

Replies to objections:

1. {{prefix}}, {{suffix}}, and {{confix}} already do that. As I asked in my second question, why are they OK but is {{derv}} not?
2. This is already the case with {{prefix}}, {{suffix}}, and {{confix}} when used for affixes with a small number of derived terms (e.g., semper-, which has five members).
3. AFAIK, that is not a practice which is being defended, in that specific case because we don't want a derivations category for (deprecated template usage) absorb.

Derived terms sections for ordinary terms will come to resemble those for affixes. Have I answered all your questions? — Raifʻhār Doremítzwr ~ (U · T · C) ~ 17:13, 25 March 2011 (UTC)

I am sorry that you are annoyed by being forced to discuss and analyze something you do not want to. You do have the option of not participating in the discussion. Is it uncertainty or ambiguity that you don't like? Perhaps the consensus process itself? You certainly seem to be getting quite emotional about the matter.

I find your substantive contributions to the discussion useful, such as your mention of the consistency issue, which is why I would like to understand more about the actual bad consequences of small categories.

I don't understand the substance of your category-size objections in light of Ruakh's suggestion that the category link not be created by {{derv}} unless the category exists beforehand. DCDuring TALK 16:28, 25 March 2011 (UTC)

A minimum-number rule sounds okay. I guess the number should be such that more than that number is unwieldy to put in a "Derived terms" (or, for another language, "Descendants") section, perhaps 60 or 80. That's also a good measure of the important of the etymon in language: if a Latin word has 60 or 80 descendants (across all languages) or an English word has 60 or 80 derived terms (in English), then it's a pretty important word, so people may be interested in such a category (qua category, i.e. even if they're not otherwise interested in the entry).​—msh210℠ (talk) 15:51, 25 March 2011 (UTC)

I think a threshold of 60–80 is far too high; the vast majority of our pronouns categories couldn't achieve a threshold that high. The most important issue, I think, is productivity — Is the term in question the subject of new formations? — for productive terms, the lack of autocategorisation requires that their entries' Derived terms section be continuously updated. — Raifʻhār Doremítzwr ~ (U · T · C) ~ 17:26, 25 March 2011 (UTC)

I am unclear as the actual bad consequences of having small categories. Is there a performance problem? Does it interfere with some important aspect of category structure that has been sanctioned? DCDuring TALK 16:28, 25 March 2011 (UTC)

I also am unclear as the actual bad consequences of having small categories. Actually, we have many small categories that no one (as far as I know and remember) wants to delete. Here is a list of big categories and tiny counterparts, for comparison.

• Category:English nouns (134,805 entries) and Category:English particles (8 entries)
• Category:Latin script languages (577 categories) and Category:Syriac script languages (2 categories)
• Category:Faroese conjugation-table templates (127 templates) and Category:Georgian conjugation-table templates (3 templates)
• Category:English language (105 categories) and Category:Celtiberian language (1 category)

--Daniel. 10:23, 26 March 2011 (UTC)

Delete. Derived terms sections should not be split into categories except for affixes. --Yair rand (talk) 19:41, 27 March 2011 (UTC)

Why should we have categories for terms derived from individual affixes, and affixes only? These categories rarely have more than 50 entries anyway. (See the list of subcategories of Category:English suffixes.) --Daniel. 19:55, 27 March 2011 (UTC)

Yes, Yair rand, what possible reason could you have for that opinion? — Raifʻhār Doremítzwr ~ (U · T · C) ~ 23:07, 27 March 2011 (UTC)

I think it works well as a general rule because non-collapsed derived terms sections will almost always be sufficient for non-affixes, since they usually don't have all that many derived terms, and affixes always have a large amount of derived terms, otherwise they wouldn't be considered to be actual affixes (or if that's not the case, they shouldn't be considered affixes unless they have many derived terms in my opinion). (BTW, we should really have a bot populate derived terms/descendants sections automatically, or otherwise have some tool for etymology editing that automatically updates other entries.) --Yair rand (talk) 23:50, 27 March 2011 (UTC)

Compare semper- with wort. — Raifʻhār Doremítzwr ~ (U · T · C) ~ 00:00, 28 March 2011 (UTC)

Yes, there are situations where non-affixes will have many more derived terms than affixes, but that's usually not the case, afaict. (Are there any non-affixes with more than 500 derived terms, let alone numbers like those of Category:English words suffixed with -ly or Category:English words suffixed with -like?) The case of wort, specifically, would not work as a category since most of the terms are redlinks. --Yair rand (talk) 00:09, 28 March 2011 (UTC)

Shoot first and ask questions later. Please forgive me for not counting, but I would have to manually count whereas you have the advantage of having the category-membership count appearing automagically. But time has some 1400+ total derived terms. Many are redlinks, but we may be missing items as well. For example, using OneLook's wildcard search, a search for "*time*" generated more than the maximum of 2000 that they accommodate. The last one was still an "e". Limiting the search to "common" terms including "time" generated 488, of which perhaps 25% were spurious (eg, sentiment). As we would include proverbs, phrasebook entries, and various largely SoP phrases which are translation targets, the number of items we should have would far exceed the "common words and phrase" that OneLook has. DCDuring TALK 00:55, 28 March 2011 (UTC)

(edit conflict; replying to Yair and possibly repeating part of DCDuring's reply above)

If Wiktionary had mature and reliable categories for all derivations of each individual word, then we could easily search for the quantities of derivations of affixes and nonaffixes as well, thus your question would have been more easily answered.

Nonetheless, from our current coverage, I can see that:

• Most affixes and nonaffixes in English have less than 50 derivations.
• Various affixes and nonaffixes in English have more than 50 derivations.

Most importantly, the number of nonaffixes with more than 50 derivations is higher than the number of affixes with more than 50 derivations:

Many or most prepositions, conjunctions, articles and numbers are expected to have an increasingly enormous number of derivations. This number is virtually multiplied by all languages with parts of speech very much used in idioms like these.

In addition, common words such as names of basic colors and concepts of basic anatomy, basic physics, etc., commonly have at least 100 derivations; see black and eye.

See also the "Derived terms" section of the; it is painfully incomplete. On the other hand, this other list of derivations of that word is more abrangent. --Daniel. 01:30, 28 March 2011 (UTC)

Looks like the whole situation is a mess. time has a huge amount of derived terms, making it the second-largest page in Wiktionary after water, but many of them are redlinks, so making a category for them wouldn't work well. I really don't think manually creating categories for every word that has derived terms is a good idea, and creating categories just for those that pass a certain minimum number of derived terms would cause a lot of inconsistency. I think the best course of action is to keep the setup of simple manually-filled derived terms sections for non-affixes and {{prefix}}/{{suffix}}-populated categories for affixes, until we have a real solution, hopefully involving bots and/or javascript tools, and some new templates. Creating hundreds of new derived terms categories isn't helpful. --Yair rand (talk) 01:46, 28 March 2011 (UTC)

One example of helpfulness of having many categories of derived terms is their convenience to be edited and kept up-to-date.

If we categorize every item from time#Derived terms into categories of each individual root, then we would not only populate Category:English terms derived from: time, but also Category:English terms derived from: from, Category:English terms derived from: in, Category:English terms derived from: travel, and many others.

The existence of redlinks is a separate issue. They could still be listed under time#Derived terms, together with the categorized blue links.

Granted, bots and JavaScript probably could hypothetically be convenient and up-to-date too, but it's better to stick with ideas that already exist and work. It would probably be very, very much easier to accept and support the use of relevant bots and JavaScript codes if they were already created. --Daniel. 03:33, 28 March 2011 (UTC)

I wholly support Daniel.'s reply. I would further be perfectly happy to have us suspend use of {{derv}} for some months to allow time to give some reality to the alternatives referenced by the hand-waving above about "real" [read "purely hypothetical"] solutions based on bots or dump-processing. I don't see a vast core of technical adepts willing to take such a task on. We have lost a great deal of our maintenance infrastructure with the apparent departure of Robert.Ullman and the diminished presence of Conrad.Irwin. Who will undertake the task? I doubt if we can count on any complex bot lasting more than five years, if we can get it developed and implemented at all. DCDuring TALK 05:09, 28 March 2011 (UTC)

We seem to be discussing two related yet distinct subjects, which get confounded or forgotten. To clarify the option structure as I see it, I have created the following outline, which you will see when you expand the following section. (I have placed it to a collapsible section, as it is quite long.)

Management of derived terms--options

Derived terms and their management (s stands for subject and o stands for option):

• s1: Derived terms--lists vs category
  • o11: Place derived terms manually into lists in "Derived terms" section
  • o12: Create categories for each set of derived terms, and put the content of the categories into the DT section using some sort of automatic magic, as seen in pre-
  • o13: Create categories for each set of derived terms, and only link to the category from the DT section
• s2: Derived terms--derv vs compound etc.
  • o21: Do without derv, given o12 or o13 is chosen (if o11 is chosen, derv seems outright superfluous)
    • Such as, let {{compound}} default to considering the components to be derivation sources for the compound. But will this work across languages? In case of doubt, let it default to non-derivation and switch to derivation using {{compound|head|ache|derv=1}}, or, even better, deriv=1.
    • (This option or its variant has been mentioned by msh210 in his post from 20:56, 17 March 2011, in this thread.)
  • o22: Use derv instead of or in combination with other morphology templates:
    • For closed compounds, replace the uses of the template "compound" with the use of derv: instead of {{compound|head|ache}}, write {{derv|head}} + {{derv|ache}} or the like.
    • For suffixes, instead of {{suffix|do|able}}, write {{derv|do}} + {{suffix||able}} or the like
    • Ditto for prefixes
    • For open compounds, instead of {{term|x}} + {{term|y}}, write {{derv|x}} + {{derv|y}} or the like

Now my stance toward this option structure is this. I am uncertain about o11 vs o12 vs o13 (I lean toward o11 and dislike o13), and I am unable to say what is wrong with having hugely many tiny categories other than that I dislike the idea. But I very strongly support o21 over o22. Thus, I still maintain that derv should better be deleted. --Dan Polansky 13:49, 28 March 2011 (UTC)

As an aside, a category should not be named 'English words derived from: load (noun)' but rather 'English terms derived from "load" (noun)'. The used placement of colon is discouraged by various styleguides; the categories are also for multi-word terms including open compounds. Whether quotation marks are needed is unclear; there is "Category:English words prefixed with pre-" but 'English nouns ending in "-ism"'. --Dan Polansky 14:26, 28 March 2011 (UTC)

Listing the entries in DT using collapsible magic: I have checked -ism#Derived terms terms and realized that only first 200 entries are shown by the magical expandable and collapsible "+". (Must have been discussed before, I know.) It follows that there is no clear way how to show, say, 2000 entries from a category in the DT section. Thus, it is far from clear how the requirement that all the derived terms should actually be visible in the DT section can be satisfied if the option o12 is chosen. --Dan Polansky 16:06, 28 March 2011 (UTC)

Thanks for taking the time to get into these details. By "option structure", do you mean "structure for (an) option", "optional structure", or something else? DCDuring TALK 17:25, 4 April 2011 (UTC)

I'm closing this discussion after 1½ month without arguments. The result is:

• No consensus. For this very controversial subject, especially, this probably means don't delete the template and don't use it either. (or don't use it very much) --Daniel. 06:11, 21 May 2011 (UTC)

Deletion debate 2

The following information has failed Wiktionary's deletion process.

It should not be re-entered without careful consideration.

---

These are part of an old experiment that was rather hotly disputed. Since they never caught on, we might as well clear them out again? —CodeCat 12:13, 5 August 2013 (UTC)

Support, I never liked these to be honest. Mglovesfun (talk) 07:54, 11 August 2013 (UTC)

• Delete. For a previous discussion, see Template_talk:derv#Deletion_debate from March 2011, where deletion was supported by Prince Kassad, weakly Mglovesfun, Dan Polansky, Tooironic, and Yair rand, while keeping was boldly supported by Raifʻhār Doremítzwr, and DCDuring, and non-boldly by Ddpy (who seemed to like derv very much), and quite possibly by Daniel Carrero, who closed the discussion as "no consensus"; the positions of Ruakh and msh210 in that discussion can be assessed as abstaining. --Dan Polansky (talk) 19:31, 16 August 2013 (UTC)