Wiktionary talk:Votes/pl-2014-07/Redirecting attested romanizations

Romanization on native language pages
I have some questions about this: I hope that the answers would allow me to support this. DCDuring TALK 22:10, 14 August 2014 (UTC)
 * 1) Do all pages using non-Roman scripts have a romanization of the headword? (How do we know?)
 * 2) If not do we have a system for assuring that they will?
 * 3) In particular do we have a system for making sure that no romanization entry is rendered into a redirect to a page that does not have such romanization?
 * My intention is to create the redirects by hand. Initially, I would look for them by looking at the entries themselves, and then seeing if the romanization used in a particular entry is attested in print. If an entry in a non-Roman script is lacking romanization, then it will not be possible for a redirect to be made under this process. Moreover, an entry in a non-Roman script that is lacking romanization is a problem for us, because it is going to be less helpful to the English speaker who can't read that script, so a romanization should be created if one is lacking. This leaves the question, however, of what to do with an entry that has one romanization, where a different romanization for that entry is attested. I would think that attested variations in the romanization of a term would be information that we would want to include in an entry on that term, and should be added somewhere. In short, I can't see a circumstance where we would want an attested romanization redirecting to an entry and not also being noted in that entry. bd2412 T 23:25, 14 August 2014 (UTC)


 * There's no single system of assuring that all non-Roman scripts have a romanisation. For this, bots and complex modules and templates are written, which you dislike :) It's all in the hands of editors who look after languages or people who look after the state of Wiktionary in general. People who create redirects, should also ensure that those entries have romanisation.
 * What kind of redirects is this vote about - hard or soft? Hard redirects don't require citations. --Anatoli T. (обсудить/вклад) 01:50, 15 August 2014 (UTC)
 * This is about hard redirects - although they don't require citations, we generally don't use hard redirects at all. There is no rule, actually, requiring that citations be attached to an entry, and we have a number of citations pages for unattached terms to explain why they don't merit an entry. The imposition of a requirement here would allow us to create a record of usage for romanizations, and insure that editors do not begin making up an endless stream of their own otherwise unused romanizations. bd2412 T 02:15, 15 August 2014 (UTC)
 * I see. What happens if later, a romanisation is ALSO found to be a word in another language? Do you plan to address such situations at all (e.g. if it's already known that a romanisation is a word in another language)? Another point, what about similarly spelled terms, where the difference is only in diacritics (a word in one language vs a romanisation in another) used but same basic letters? --Anatoli T. (обсудить/вклад) 02:25, 15 August 2014 (UTC)
 * Interesting - I consider letters with diacritics to be different letters from those without, in the same way that "n" is different from "h" despite the only difference between them in Roman script is that the back line of the "h" extends a bit farther up. If conflicting uses are found later, then the redirect would need to be deleted, which I suppose would allow the search function to return both results. bd2412 T 02:34, 15 August 2014 (UTC)
 * OK, with diacritics, what I means is, it may be necessary to add on the terms, which are NOT redirects, that's all. And  would contain a link to the term, not the redirect, such as  . --Anatoli T. (обсудить/вклад) 02:46, 15 August 2014 (UTC)
 * I have no objection to that, although there is no difference in the outcome between using a piped link to the redirect target, and linking to the redirect that takes you to that target. It's merely an extra bit of syntax for the editors to add. bd2412 T 03:41, 15 August 2014 (UTC)
 * In addition to requiring extra typing, it requires all the links to be updated if someone finds mahā to be a word in (say) Hawai'ian — since at that point you want the link from [[maha]] to be to [[mahā]], not [[महा]]. (It's also just weird, in my opinion, for [[maha] to link to [[महा]], even if or perhaps especially if it appears to be linking somewhere else.) So, I don't think piping should be used; just use . - -sche (discuss) 04:03, 15 August 2014 (UTC)

Conflict with eventualism
How well are we to we ascertain that "the only use of the term at issue is as a romanization of the native script"? If it later turns out that a term that we thought exists only e.g. as a romanization of Japanese is also an independent lemma in Samoan or Maxakali or whatever, is the plan to then deconvert the Japanese term also to a soft-redirect or an independent entry? --Tropylium (talk) 17:06, 28 October 2015 (UTC)
 * I would think that a soft redirect would have to do in that case. bd2412 T 13:50, 29 October 2015 (UTC)