Template talk:hu-suffix

RFD discussion: June 2016–March 2019

The following information has failed Wiktionary's deletion process (permalink).

It should not be re-entered without careful consideration.

---

Template:hu-suffix

• Keep. This template has additional functionality that the {{suffix}} template does not have:
  1. It categorizes by the last suffix only, e.g. {{hu-suffix|egyetem|leg|es|pos=adj}}
  2. It is able to categorize by the suffix even when the stem is left blank, e.g. {{hu-suffix||izmus|pos=n}}. This is used in many entries.
  3. It has a pos2 parameter for placing a word into two PoS categories. --Panda10 (talk) 12:59, 5 August 2016 (UTC)

Template:hu-prefix

In the 2013-2014 RFDO discussion Template talk:hu-compound (to delete: {{hu-compound}}, {{hu-suffix}} and {{hu-prefix}}), most people voted "Delete" but only hu-compound was deleted and Wonderfool closed the RFDO as "Kept as nobody's deleted it", which I find unconvincing as a reason for keeping the templates. It should have been: "RFD failed, waiting for all templates to be orphaned".

I did not participate in that discussion; I'd vote delete too.

Rationale:

• Use {{suffix}} or {{prefix}} (better yet, in my opinion: {{affix}} or the shortcut {{af}}) in all cases.

--Daniel Carrero (talk) 23:38, 28 June 2016 (UTC)

Delete. I'd also note that this template is used to provide etymologies for inflected forms, such uses should be removed. See for example kapok. —CodeCat 23:43, 28 June 2016 (UTC)

Keep. What's the reason {{fi-suffix}}, {{fi-prefix}}, {{hi-suffix}}, {{hi-prefix}} etc. are allowed but the Hungarian templates are not? Also, I want to keep using hu-suffix for inflected forms, I think the etymology section is useful for agglutinative languages, users might want to click on the suffix to check it out. I do not understand what's the harm in this. --Panda10 (talk) 13:38, 30 June 2016 (UTC)

The Finnish and Hindi templates you gave are headword-line templates for prefix/suffix entries. Not at all equivalent. —CodeCat 14:04, 30 June 2016 (UTC)

How about {{fa-prefix}}, {{fa-suffix}}, and {{rom-suffix}}? --Panda10 (talk) 14:24, 30 June 2016 (UTC)

I speedied {{rom-suffix}} (it was created in 2011 and was being used only in 6 entries: melalo, samtalo, baxtalo, shilalo, rupuno and rupalo)

Correct me if I'm wrong: it seems Persian affixes don't have a hyphen and we can't use {{affix}} for this language (except for compounds: سیلاب); so, there's an actual need for {{fa-suffix}} and {{fa-prefix}}. If we can replace these two by {{affix}}, I'm for deleting them.

For the record, a cursory search reveals that some (or all?) entries for Persian suffixes like گاه and prefixes like بی are using {{head|fa|suffix}} and {{head|fa|prefix}}, respectively. It's a bit confusing if some languages have the format Template:xx-suffix for actual suffix entries and other use that format for etymologies involving suffixes. --Daniel Carrero (talk) 22:32, 30 June 2016 (UTC)

Concerning inflected forms with etymologies: do you mean something like adding "lion +‎ -s" in the entry lions? I don't think we should do that, but in any case {{affix}} can do this as well, so that's not a good reason for keeping {{hu-suffix}}, in my opinion. --Daniel Carrero (talk) 22:37, 30 June 2016 (UTC)

Ok, I will delete them after I reworked the entries. I have already started using {{af}}. This may take some time, though. {{hu-prefix}} has about 900 entries, {{hu-suffix}} has about 17,000. Re inflected forms: Adding an etymology section to English forms may have less practical value, since English is not an agglutinative language. It is different in Hungarian where nouns, adjectives, pronouns, and numerals can have more than 34 different inflected forms, verbs much much more. If a Hungarian editor is willing to add this information using {{affix}}, what's the harm in it? --Panda10 (talk) 23:23, 30 June 2016 (UTC)

Thank you. Fair enough, I don't mind Hungarian entries having the etymologies for inflected forms, per the reasons you mentioned.

Maybe the change from hu-suffix/hu-prefix to affix can be done by bot? I believe these would be the search+replace rules:

• "{{hu-suffix|" → "{{af|hu|"; and adding the "-" in all unnamed parameters starting from 2= (in diff, parameters 2= and 3= have the hyphen added)
• "{{hu-prefix|" → "{{af|hu|"; and adding the "-" in all unnamed parameters except the last

@CodeCat: Would you be so kind as to use MewBot to do that?

@Panda10: Did I make any mistake in the rules above? --Daniel Carrero (talk) 00:01, 1 July 2016 (UTC)

What about the categorization? The two templates don't work the same in that regard, {{hu-suffix}} ignores all suffixes except for the last one. Also, {{hu-suffix}} has a pos2= parameter that {{suffix}} lacks (well, it's there, but it does something else). Finally, I disagree with putting etymologies in inflections. We don't do it for other languages, agglutinative or not. Compare Finnish. In fact, a non-agglutinative language has more use for such etymologies than an agglutinative one, because the affixes are much less obvious and harder to tell apart in a word. —CodeCat 00:22, 1 July 2016 (UTC)

Let's separate the three issues:

• 1) {{hu-suffix}}: If {{suffix}} does not have the functionality of {{hu-suffix}}, then we will have to keep {{hu-suffix}}, just like we keep {{fa-prefix}} and {{fa-suffix}}.
• 2) {{hu-prefix}}: I can deal with it manually. Those entries need other standardization, so I'll do them at the same time and I will delete the orphaned template. No need for a bot.
• 3) Etymology in inflected forms: This needs a separate discussion. It is not part of this deletion request. I suspect that it would be useful for every language but it is time-consuming and editors simply do not want to spend time with it. --Panda10 (talk) 11:55, 1 July 2016 (UTC)

Hold on. Why does {{hu-suffix}} categorise that way? Surely if the word is suffixed with several suffixes, we'd want categories for all of them, not just the last? And I consider the pos= parameter to be deprecated, since we now have id2= that gives finer control and better disambiguation (i.e. by sense of the suffix rather than part of speech of the thus-formed word). The pos2= parameter is puzzling, as it essentially causes the entry to be categorised twice in two different suffix categories. —CodeCat 14:23, 1 July 2016 (UTC)

There is nothing puzzling about pos2. A word can be both adjective and noun, e.g. akadémikus. Pos adds the entry to the first category, while pos2 in the second. By the way, you say pos is deprecated, then how come it is used by {{affix}}? It makes sense to categorize Hungarian entries by part of speech. About categorizing by the last suffix: sometimes there is no Hungarian word with the middle suffix. E.g.: egyetemleges, there is no *egyetemleg. --Panda10 (talk) 14:51, 1 July 2016 (UTC)

One more thing: If the template name is not up to the current standards of this wiki, then why not just simply rename {{hu-suffix}} to something else, run the renaming bot, keep all the current functionality and be done with it. --Panda10 (talk) 15:03, 1 July 2016 (UTC)

What is the reason that {{hu-suffix}} only categorizes by the last suffix? Words like egyetemleges should probably be in both suffix categories (at least if we don't want to start creating entries for "compound" suffixes like -leges). Words derived from an intermediate derivation should by contrast IMO only show the immediate base word, e.g. kérdéses from kérdés, not from kérd or from kér (though those could be of course mentioned in prose in the etymology). --Tropylium (talk) 20:08, 26 August 2016 (UTC)

Please see if this thread answers your question: [1]. Also, how would you handle fickó, helyesbít, javasol, jövedelem, költözik, lassacskán, nyilatkozat, piszkozat, etc. --Panda10 (talk) 21:48, 26 August 2016 (UTC)

There are huge numbers of languages all over the world that have complex morphologies like this- not just suffixes, but prefixes, infixes and circumfixes. We really need to work on getting our general templates to be able to handle such issues, so someone working on, say, some obscure American Indian or African or Southeast Asian language doesn't have to come up with their own custom templates. Chuck Entz (talk) 05:18, 27 August 2016 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘: Maybe {{af}} could have built-in language-specific rules, like "for Hungarian, categorize only the last suffix". --Daniel Carrero (talk) 05:32, 27 August 2016 (UTC)

Like Tropylium (a linguist who works with agglutinative languages), I have my doubts that categorizing only the last suffix is a good idea, but it's an extremely complex issue. When you can have a single word that means "I saw those two women walk this way out of the water", you need a more systematic approach to categorization of the components. Chuck Entz (talk) 06:07, 27 August 2016 (UTC)

Deleted {{hu-prefix}}. --Panda10 (talk) 12:49, 5 August 2016 (UTC)

@Panda10, Daniel Carrero, Chuck Entz, Erutuon, Rua Can we finally figure out how to delete this? Benwing2 (talk) 22:17, 17 March 2019 (UTC)

@Benwing2: Here is what I think could be done in this order:

1. Run a bot in Category:Hungarian non-lemma forms based on Daniel Carrero's instructions above, copied here for convenience: "{{hu-suffix|" → "{{af|hu|"; and adding the "-" in all unnamed parameters starting from 2= (in diff, parameters 2= and 3= have the hyphen added). These entries already should contain nocat=y.
2. Run a bot in Category:Hungarian lemmas to find all instances using pos2 and list them in a category. I'd like to take a look. Example: {{hu-suffix|festék|es|pos=adj|pos2=n|t1=paint, colour|t2=adjective-forming suffix}}
3. Run a bot in Category:Hungarian lemmas to find all instances with more than one unnamed parameters and list them in a category. I'd like to take a look. Example for two unnamed parameters: e.g. {{hu-suffix|egyetem|leg|es|pos=adj}}
4. Run a bot in Category:Hungarian lemmas to change all other instances that are not in the above two cases. The lemmas are all categorized by part of speech. So pos=n should be changed to pos=noun, pos=adj to pos=adjective, pos=adv to pos=adverb, pos=v to pos=verb. There might be more things to pay attention to, and I'd like to think about it more tomorrow with a clear mind. Panda10 (talk) 01:21, 18 March 2019 (UTC)

While we're at it, let's get rid of Template:fa-prefix and Template:fa-suffix, now that {{affix}} supports Persian hyphens. Benwing2 (talk) 22:24, 17 March 2019 (UTC)

@Panda10 I can do those bot runs. Benwing2 (talk) 01:36, 18 March 2019 (UTC)

@Panda10 These categories already exist and have for years: Category:hu-suffix with pos2, Category:hu-suffix with multiple suffixes and categorization disabled, Category:hu-suffix with multiple suffixes and categorization enabled. Benwing2 (talk) 08:46, 18 March 2019 (UTC)

@Benwing2: Thanks, I completely forgot about the tracking categories. It might take a day or two for me to clean the two categories where categorization is enabled: 120 + 79 entries at this moment. Do you prefer to run your bot in two steps outlined above or only once on all items linked to {{hu-suffix}}? Panda10 (talk) 16:01, 18 March 2019 (UTC)

@Panda10 I can run the bot on the entries in cases #1 and #4, and handle the remaining words later, if that's OK with you. I've already written the code to do so in fact. We can add support for two parts of speech to {{affix}} if necessary. I'm not sure what to do with the issue of creating a category only for the last suffix; you can do that in {{affix}} by prefixing each suffix that shouldn't be categorized with a ^, but that seems an ugly solution to me. Maybe we should just categorize all suffixes? That's what we do for other languages if two suffixes are given (but the proper solution is to include the intermediate form as a full word if it exists, e.g. internationalization should be analyzed as internationalize + -ation, not as international + -ize + -ation). Benwing2 (talk) 19:21, 18 March 2019 (UTC)

BTW one of the things my code does is attempt to convert cases where t1, t2, t3 etc. have been coopted to indicate a part of speech (or similar) into pos1, pos2, pos3, etc. Currently I look for cases where t1 etc. ends in a space followed by any of the following words: marker, suffix, ending, vowel, plural. This should catch things like "plural marker", "translative suffix", "accusative case ending", "linking vowel", "possessive plural", etc. Benwing2 (talk) 19:26, 18 March 2019 (UTC)

Note that if the conversion of t1 into pos1 catches something it shouldn't or misses something it's not a huge issue, it just means that the quotes that indicate a gloss will wrongly be present or absent. Benwing2 (talk) 19:31, 18 March 2019 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘: @Benwing2: The t1 to pos1 etc. conversion is good. Thanks for thinking about it. Occasionally, you may find a tr1 and tr2, those were used by mistake, obviously. I went through Category:hu-suffix with pos2 and it is now empty. You said your code will not touch items in Category:hu-suffix with multiple suffixes and categorization enabled, and that's good because they are somewhat complicated. I'm sure you are aware that nocat=y can also be nocat=hu or nocat=1.

What would the bot do with these:

{{hu-suffix|-e-|d|nocat=y|t1=linking vowel|t2=possessive suffix}}

{{hu-suffix|-i|d|nocat=y|t1=possessive plural|t2=second-person singular personal suffix}}

{{hu-suffix|-l|et|nocat=y}}?

There was another functionality in {{hu-suffix}} where the first parameter, an actual word, is empty, and only a suffix is provided, e.g.: {{hu-suffix||n|pos=adv}}. In those cases, an additional + sign was displayed, most of the time unnecessary, but occasionally needed. I'm really worried that I forgot to think about stuff and there is no way to catch them because of the huge volume. I think if the bot encounters something unexpected, it should collect those items in a category, so later I can take a look. Panda10 (talk) 20:04, 18 March 2019 (UTC)

@Panda10 The first one maps to {{affix|hu|^-e-|-d|nocat=y|pos1=linking vowel|pos2=possessive suffix}}, the second and third similarly, the fourth maps to {{affix|hu|^|-n|pos=adverb}}. These two uses of ^ were recently added by me to ensure that all uses of {{suffix}} could be properly mapped to {{affix}}. Benwing2 (talk) 20:12, 18 March 2019 (UTC)

@Benwing2: Okay, thanks. Go ahead and run it. Panda10 (talk) 20:18, 18 March 2019 (UTC)

@Panda10 Cool. BTW my script will currently ignore Category:hu-suffix with multiple suffixes and categorization disabled as well but I can have it convert those (I guess it doesn't matter as there won't be any categorization). As for |nocat=hu, |nocat=1, etc., these will be left alone. This param is parsed by Module:parameters; the documentation says "No value, the empty string, and the strings "0", "no", "n" and "false" are treated as false, all other values are considered true." As for tr1=, tr2=, etc. I'll have the script convert those to t1= etc. Benwing2 (talk) 20:38, 18 March 2019 (UTC)

@Panda10 Results from the first 160 conversions are here: User:Benwing2/find-fix-hu-suffix-first-160. AFAICT all look fine but you might see something wrong that I didn't catch. Benwing2 (talk) 02:10, 19 March 2019 (UTC)

@Benwing2: I went through the displayed 100 items and they look really good! Thanks. Panda10 (talk) 16:33, 19 March 2019 (UTC)

@Panda10 Sure. I ran my bot and it converted about 21,300 cases. There are between 50 and 100 remaining pages that use {{hu-suffix}}; see Special:WhatLinksHere/Template:hu-suffix. These appear to be the same as the entries in Category:hu-suffix with multiple suffixes and categorization enabled plus a few scattered non-mainspace references, which you can mostly ignore (in particular, any page in userspace and any page that is for discussion should be left alone even after the template itself gets deleted). Let's see if we can get rid of the remaining cases. As I mentioned above, you can forcibly disable categorization of a given affix by preceding it with ^, but we should avoid doing this unless absolutely necessary (and even then, the practice in similar cases for other languages is to let all affixes be categorized). Benwing2 (talk) 02:48, 20 March 2019 (UTC)

@Benwing2: Great! Thanks again. This was very helpful. BTW, I emptied the above category and did not use the ^ sign. Can I delete the template and its documentation page? Panda10 (talk) 18:49, 20 March 2019 (UTC)

@Panda10 Yes, please do. Thanks! Benwing2 (talk) 18:50, 20 March 2019 (UTC)