Template talk:str index
The following discussion has been moved from Wiktionary:Requests 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.
Ridiculously heavy templates imported from Wikipedia that attempt to manipulate strings by using padright tens of thousands of times. --Yair rand (talk) 22:08, 7 February 2011 (UTC)
- I'd rather keep them, and use them for templates of categories. --Daniel. 22:18, 7 February 2011 (UTC)
- As a side note, I would also be happy if alternatively mw:Extension:StringFunctions could be installed here. --Daniel. 22:20, 7 February 2011 (UTC)
- The devs have made it pretty clear that none of the Wikimedia wikis are ever going to get string functions. --Yair rand (talk) 22:24, 7 February 2011 (UTC)
- By my understanding, that's only partly true. The devs take issue with some specific aspects of that extension, but they're not opposed to the entire concept of string-functions. IINM, it's considered likely that a significant subset of its functions will eventually be incorporated into the ParserFunctions extension (which is installed here). —RuakhTALK 22:30, 7 February 2011 (UTC)
- Delete
{{str find}},{{str index}}, and{{str sub}}. (I'd be fine with deleting the other two as well, but they actually aren't very expensive, so I don't terribly mind if we keep them.) —RuakhTALK 22:21, 7 February 2011 (UTC)
- Incidentally, because of the way they're implemented, these templates are also quite limited:
{{str find}}can only search the first 50 characters (which limit can be raised, but only by making the template commensurately more expensive), and{{str index}}and{{str sub}}only support a very limited character set (which set can be made less limited, but only by making the template commensurately more expensive). —RuakhTALK 23:47, 7 February 2011 (UTC)
- Incidentally, because of the way they're implemented, these templates are also quite limited:
- Strong support of what Ruakh said. DAVilla 19:52, 22 February 2011 (UTC)
Keep. I've been using this extensively to add inflection tables. It removes the last letters of a word to produce a stem for inflection. For example, in the word hevonen, the following:
{{fi-decl-nainen|{{subst:padright:|{{subst:#expr:{{subst:str len|{{subst:PAGENAME}}}}-3}}|{{subst:PAGENAME}}}}}}
Produces: {{fi-decl-nainen|hevo}}
I had valuable help on this from Yair rand. ~ heyzeuss 22:15, 8 February 2011 (UTC)
- Well, if we can guarantee any of these templates only be used substed, I wouldn't mind keeping it. I suppose using the subst: trick that ensures that, protecting the page, and adding notes to the talkpage and documentation indicating that the templates must remain nontranscludable will do.—msh210℠ (talk) 22:22, 8 February 2011 (UTC)
- I'd be O.K. with that as well. —RuakhTALK 23:12, 8 February 2011 (UTC)
- Thank you kindly. That would have pulled the rug out from under me! :) ~ heyzeuss 06:12, 9 February 2011 (UTC)
- BTW, also delete
{{str len}}and delete or fix{{wide image/sandbox}}. —RuakhTALK 22:29, 7 February 2011 (UTC)- I've added those headers, above, now.—msh210℠ (talk) 22:34, 7 February 2011 (UTC)
- Delete str find, str index, str sub, and str len, per others' comments.—msh210℠ (talk) 22:34, 7 February 2011 (UTC) ← This opinion has been qualified; see my comments, above, of 22:22, 8 February 2011 (UTC).—msh210℠ (talk)
- Keep Template:wide image/sandbox, but do not use in NS:0. It's a sandbox. Mglovesfun (talk) 23:07, 13 February 2011 (UTC)
- I have now deleted
{{str find}},{{str index}}, and{{str sub}}; kept{{str left}}and{{str ≥ len}}; kept{{str len}}but rendered it unusable except for subst:ing, and kept{{wide image/sandbox}}per Mglovesfun, but noting that it's no longer used anywhere in mainspace and will no longer work now that{{str sub}}and{{str find}}are gone. —RuakhTALK 20:23, 22 February 2011 (UTC)
Resurrection?
This template can allow a kind of ultimate declination template, which could:
- Display the orthographies of the most part of the plurals (and feminine), only from their singular pages names (with {PAGENAME} on them).
- Idem with their pronunciations, and eventually categorized into "missing pronunciation", and maybe their transcriptions and transliterations.
So when putting a nature and a language on a page, anyone would see the normal declinations of this orthography into this language. JackPotte 23:09, 8 October 2011 (UTC)
For some examples, please see Wiktionary:Beer_parlour#fr:Template:en-nom-rég2. JackPotte 10:29, 30 October 2011 (UTC)
Deletion
- See also Wiktionary:Beer parlour archive/2011/November#For the lack of a "Wiktionary:Requests for deletion review" page, Wiktionary:Requests for deletion/Others#Template:py-to-ipa.
I am very much disappointed with this deletion. These are potentially very useful templates. There's no need to manually transliterate a lot of scripts if these are allowed (eg. Template:bo-transli for Tibetan, or convert Pinyin into IPA with Template:py-to-ipa). Now this is just exponentially increasing the amount of work needed and inducing human errors. Hbrug 22:13, 29 October 2011 (UTC)
- French Wiktionary uses it to form Spanish plurals (fr:Modèle:es-rég), Chinese Wiktionary uses it to transliterate Uyghur (zh:Template:维吾尔转), and Vietnamese Wiktionary uses it to auto-generate Vietnamese pronunciations in six dialects (vi:Bản mẫu:vie-pron). The benefits of using this obviously outweigh the "expensiveness". Hbrug 22:19, 29 October 2011 (UTC)
- Anything that could be done by slicing a template-parameter using {{str index}}, could be done by accepting a sequence of parameters. —RuakhTALK 22:39, 29 October 2011 (UTC)
- Yes, but that is manually expensive, isn't it? Say I want to transliterate Tibetan རྒྱལ་མཚན using the Wylie scheme, since Tibetan script uses stacking, that would mean I have to find the base characters and diacritics from elsewhere when doing this, i.e. {{bo-transli|ར|ྒ|ྱ|ལ|་|མ|ཚ|ན}} (instead of {{bo-transli|རྒྱལ|མཚན}}), which is very inefficient and prone to errors, yes? Hbrug 22:46, 29 October 2011 (UTC)
- Another option is to have the template exist, but require substitution. —RuakhTALK 23:07, 29 October 2011 (UTC)
- But the disadvantage of that, in case some feature of the transliteration or convertion needs to be systematically changed, for example in Template:py-to-ipa (say the final -i, after zh/ch/sh/r in Pinyin, would like to be indicated in IPA by the Sinologist convention /ʅ/ instead of the obscure /ʐ̩/, just an example), would be the need to utilise a bot (or equivalent) to go through all past substitutions and do replacements (with the possibility of false-identification), yes? Hbrug 23:13, 29 October 2011 (UTC)
- Also, the code would be much cleaner without having to substitute beforehand. Compare {{ug-transli|ئ|ۇ|ن|ى|ۋ|ې|ر|س|ى|ت|ې|ت}} and simply {{ug-transli}} at the Uyghur entry ئۇنىۋېرسىتېت ("university"), the latter is undoubtedly much easier to produce and use. Hbrug 23:29, 29 October 2011 (UTC)
- This sort of thing would be much better done using the Transliterator extension; see https://bugzilla.wikimedia.org/show_bug.cgi?id=20246. In the meantime, I think you're best off just choosing a transliteration scheme and sticking with it, rather than trying to implement it via templates. —RuakhTALK 11:42, 31 October 2011 (UTC)
- I don't think the Transliterator extension can handle the transliteration of Tibetan (please see code for Template:bo transli). This is how many transliteration errors in Greek existed on en.wikt in 2010. Is it really hard to implement changes here, even if the benefits will be enormous? Or is it just people (esp. admins) being indifferent and nonempathetic to the convenience of other working users? After all, the templates pose no visible harm to editing, loading, or viewing whatsoever. Hbrug 23:21, 31 October 2011 (UTC)
- You're right, the Transliterator cannot handle Brahmi derived scripts like Tibetan. I am all for restoring str index, but do not want to overturn my fellow admin's decision all by myself. -- Liliana • 23:28, 31 October 2011 (UTC)
At this point, I wish there was a way to somehow resurrect Robert Ullmann from his grave. -- Liliana • 11:49, 1 November 2011 (UTC)
The following information has failed Wiktionary's deletion process.
It should not be re-entered without careful consideration.
These have no transclusions and aren't really needed anymore because we have Lua to do this now. There's also Template:str left but that's still used on a few pages. —CodeCat 23:10, 26 December 2013 (UTC)
- They're also very expensive. So, yes, now that we have Lua, we should orphan and delete these templates. We should make sure User:Heyzeuss isn't still using these via subst: (as they were the last time these were RFDed), though, and make sure they know what Lua code to use to get the same effect. - -sche (discuss) 23:22, 26 December 2013 (UTC)
- I am not currently using any of these, even with subst. As long as there is some alternative, I have no objections. Lua looks interesting, and I suppose that I will have need of it sooner than later. ~ heyzeuss 13:33, 23 January 2014 (UTC)