Wiktionary:Votes/2017-03/Deprecating 4=, 5= and gloss= parameters in favor of t=

Deprecating 4&#61;, 5&#61; and gloss&#61; parameters in favor of t&#61;
Voting on: Deprecating the positional parameters 4 in templates, , and and 5 in , , and , and the named parameters gloss in all the aforementioned templates and glossN in  and its related templates, in favor of the named parameters t or tN. For flexibility, there will be two separate questions in this vote:
 * Question 1: Deprecate the positional parameters 4/5
 * Question 2: Deprecate the named parameters gloss/glossN

Rationale:
 * To increase consistency between templates.
 * To increase flexibility in parameter ordering, while maintaining readability and brevity.

Schedule:
 * Vote starts: 00:00, 5 April 2017 (UTC)
 * Vote ends: 23:59, 4 May 2017 (UTC)
 * Vote created: WikiTiki89 18:31, 29 March 2017 (UTC)

Discussion:
 * [[Image:Wikt rei-artur3.svg|20px]] Beer parlour/2016/October
 * [[Image:Wikt rei-artur3.svg|20px]] Beer parlour/2017/March
 * [[Image:Wikt rei-artur3.svg|20px]] Wiktionary talk:Votes/2017-03/Deprecating 4&

Support

 * 1)  --Daniel Carrero (talk) 03:31, 5 April 2017 (UTC)
 * 2) --Dixtosa (talk) 08:40, 5 April 2017 (UTC)
 * 3)  --WikiTiki89 15:29, 5 April 2017 (UTC)
 * 4)  -- Crom daba (talk) 08:25, 6 April 2017 (UTC)
 * Germyb (talk) 07:25, 19 April 2017 (UTC)
 * User not qualified to vote. --WikiTiki89 15:23, 19 April 2017 (UTC)
 * 1)  — I see a small benefit in eliminating blank parameters generally. (For clarity, I consider  preferable to, but I consider  slightly preferable to .) — I.S.M.E.T.A. 23:02, 21 April 2017 (UTC)
 * 2)  (I explain in the next section why I have changed my vote.) — SMUconlaw (talk) 12:10, 30 April 2017 (UTC)

Oppose

 * 1)  It is widely in use and cuts down on a character or two. I don't see any reason why people shouldn't be allowed to use 4 and t interchangeably. --Victar (talk) 05:43, 5 April 2017 (UTC)
 * I would support, . --Victar (talk) 16:28, 26 April 2017 (UTC)
 * for the reason given by Victar. — SMUconlaw (talk) 13:03, 11 April 2017 (UTC)
 * After having read the discussion below, I have decided to change my "oppose" vote to "support" on this issue as I can see some benefits in doing away with a positional parameter that is empty most of the time and thus could be confusing to new editors. — SMUconlaw (talk) 12:09, 30 April 2017 (UTC)
 * I disagree with that supposition because a) if you keep both options available for use, new editors can use t if they find it easier, just as if you prefer to use positional parameters instead of labeled parameters in, you have that option available to you, and b) if the issue is having a blank parameter 90% of the time, perhaps we should actually remove the blank parameter and replace it with a labeled one, i.e. alt. --Victar (talk) 07:21, 2 May 2017 (UTC)
 * 1)  I think the existing convention is too widely embedded for it to be reasonable to deprecate it. Benwing2 (talk) 02:30, 18 April 2017 (UTC)
 * 2)  Increasing flexibility by reducing options? Why would we make things more difficult? Korn &#91;kʰũːɘ̃n&#93; (talk) 09:51, 20 April 2017 (UTC)
 * 3)  per Victar and Benwing2. --Droigheann (talk) 11:37, 22 April 2017 (UTC)
 * 4)  —Aryamanarora (मुझसे बात करो) 19:11, 27 April 2017 (UTC)
 * 5)  as per Victar (am I allowed to vote?) —Julien D. (talk) 01:39, 30 April 2017 (UTC)
 * You need a certain number of edits. I don't know how many, but you have over 500, which is almost certainly enough. Benwing2 (talk) 05:25, 30 April 2017 (UTC)
 * See here. --Droigheann (talk) 11:06, 30 April 2017 (UTC)
 * 1)  I think the fourth positional arg is ok and does not need deprecation. --Dan Polansky (talk) 11:03, 30 April 2017 (UTC)
 * 2)  per Victar (with the proviso that it saves just one character). Lingo Bingo Dingo (talk) 13:45, 3 May 2017 (UTC)

Abstain

 * 1)  This is a matter of absolutely no importance whatsoever. —Aɴɢʀ (talk) 11:33, 5 April 2017 (UTC)
 * And yet you're willing to edit-war someone over it. --Victar (talk) 22:29, 6 April 2017 (UTC)
 * That was when I thought it was already well established that 4=, 5=, and gloss= had been deprecated. —Aɴɢʀ (talk) 13:55, 25 April 2017 (UTC)
 * 1)  -Xbony2 (talk) 22:26, 6 April 2017 (UTC)
 * 2) . I don't care. It would just change the way I've been doing things with no advantage/disadvantage. I do, however, see a potential benefit for newer users, as the empty parameter can be confusing at first. Andrew Sheedy (talk) 05:50, 10 April 2017 (UTC)

Support

 * 1)  --Daniel Carrero (talk) 03:31, 5 April 2017 (UTC)
 * 2)  --Victar (talk) 05:43, 5 April 2017 (UTC)
 * , although I think this is less important than the positional parameters. --WikiTiki89 15:30, 5 April 2017 (UTC)
 * 1)  -- Crom daba (talk) 08:25, 6 April 2017 (UTC)
 * 2)  Seems too long, not matching the recent trend seen in Votes/2016-06/label → lb. Furthermore, I think that, in l, the parameter should not be there at all. --Dan Polansky (talk) 11:06, 30 April 2017 (UTC)
 * 3)  — Eru·tuon 23:39, 4 May 2017 (UTC)

Oppose

 * 1)  clarity over hmm what? Additional 4 letters?. Dixtosa (talk) 08:39, 5 April 2017 (UTC)
 * 2)  in favour of allowing both t and gloss to be used as synonyms . — SMUconlaw (talk) 13:01, 11 April 2017 (UTC)
 * 3)  per SMUconlaw. Ƿidsiþ 07:51, 19 April 2017 (UTC)
 * 4)  — If anything, deprecating t and tN in favour of gloss and glossN (or abbreviations thereof) makes more sense (cf. SMUconlaw’s points in the  section below). — I.S.M.E.T.A. 23:11, 21 April 2017 (UTC)
 * Just pointing out that most people would complain that gloss is too long, and the potential abbreviation g, which has been proposed in the past, is already taken as the gender parameter. --WikiTiki89 02:38, 23 April 2017 (UTC)
 * What about gl and glN? — I.S.M.E.T.A. 10:24, 25 April 2017 (UTC)
 * Well some people are already complaining that t is too long. I think t it's just short enough. Plus it was already in use for this purpose. --WikiTiki89 14:12, 25 April 2017 (UTC)
 * Using gl and glN for glosses requires just as much typing as using tr and trN for transliterations. It's not a problem, in my opinion. — I.S.M.E.T.A. 21:59, 25 April 2017 (UTC)
 * Human perception is relative. If you shorten tr to just = and people get used to that, when you change it back people will complain it's too long. --WikiTiki89 14:20, 26 April 2017 (UTC)
 * I can't say I have much sympathy with such a complaint. — I.S.M.E.T.A. 23:34, 4 May 2017 (UTC)
 * 1)  per SMUconlaw. Lingo Bingo Dingo (talk) 13:45, 3 May 2017 (UTC)
 * --Hekaheka (talk) 11:08, 20 May 2017 (UTC)

Abstain

 * 1)  This is a matter of absolutely no importance whatsoever. —Aɴɢʀ (talk) 11:33, 5 April 2017 (UTC)
 * 2)  -Xbony2 (talk) 22:26, 6 April 2017 (UTC)
 * 3)  I am leaning towards oppose but I will abstain on this for the moment. Benwing2 (talk) 02:30, 18 April 2017 (UTC)
 * 4)  Didn't even know gloss(N) can be used in these templates. --Droigheann (talk) 11:40, 22 April 2017 (UTC)
 * 5)  —Aryamanarora (मुझसे बात करो) 19:11, 27 April 2017 (UTC)

Comment
What does the t actually stand for? — SMUconlaw (talk) 12:01, 11 April 2017 (UTC)
 * Translation? --Giorgi Eufshi (talk) 12:36, 11 April 2017 (UTC)
 * If so, it seems rather odd to use t since much of the time the parameter is not used to indicate a translation but a gloss. — SMUconlaw (talk) 13:00, 11 April 2017 (UTC)
 * Yes, it means translation. What's the difference between a translation and a gloss? Also note that tN has been in use since forever in some templates, it wasn't made up just for this. --WikiTiki89 14:16, 11 April 2017 (UTC)
 * I was under the impression that a translation meant an explanation of a term in a different language, whereas a gloss was simply an explanation of the term. I note that our entry defines the word thus: "A brief explanatory note or translation of a foreign, archaic, technical, difficult, complex, or uncommon expression, inserted after the original." Therefore, it seems all translations are glosses, but not all glosses are translations. — SMUconlaw (talk) 14:40, 11 April 2017 (UTC)
 * You're right, I didn't think of that, because most of our glosses are for non-English terms. But technically a translation doesn't have to be to another language, but can be to another register, and so you can say that any English term that needs a gloss, is in an incomprehensible register and therefore requires a translation. Anyway, I had proposed g for this purpose before (I still can't find that discussion), but it was already taken for gender. --WikiTiki89 14:46, 11 April 2017 (UTC)
 * m or d would have made sense as "meaning" or "definition", respectively. Oh well. --Victar (talk) 17:09, 12 April 2017 (UTC)
 * But again, we were already using t for this purpose in some templates. --WikiTiki89 17:19, 12 April 2017 (UTC)
 * Which is why my comment was both retrospective and capitulative. --Victar (talk) 02:08, 13 April 2017 (UTC)


 * As we have a module which autostrips diacritica, wouldn't it be more reasonable to remove parameter 4 from inh/der in favour of a named/optional one? Korn &#91;kʰũːɘ̃n&#93; (talk) 09:54, 20 April 2017 (UTC)
 * True. — SMUconlaw (talk) 12:59, 20 April 2017 (UTC)
 * Because then would have to be replaced with, which is much more cumbersome. --WikiTiki89 13:39, 20 April 2017 (UTC)
 * Ah, I see, though arguably it's not really that much more cumbersome. — SMUconlaw (talk) 17:48, 20 April 2017 (UTC)
 * Much more cumbersome than replacing with . --WikiTiki89 20:09, 20 April 2017 (UTC)
 * How is that an optional named parameter? – Done. Korn &#91;kʰũːɘ̃n&#93; (talk) 22:36, 20 April 2017 (UTC)
 * Sorry, didn't see that part of the comment. The problem is our standard name for such a parameter is alt which may be considered too long (I've always hated using it at least), and using anything shorter would be introducing a new parameter that no one's heard of before. That said, it's not out of the question. --WikiTiki89 22:41, 20 April 2017 (UTC)
 * I suppose longer names are generally acceptable by the reasoning acceptable length of name is proportional to obscurity of parameter is inversely proportional to frequency of usage. On the other hand we have somewhat a tendency of shortening to single letters, which is debated, though. There is no immaculate solution, but any solution is preferable to the template having an empty parameter in its majority/default case. I add that a strict policy of standardisation of abbreviations across all templates will make any of them more acceptable. Korn &#91;kʰũːɘ̃n&#93; (talk)
 * You sound like someone who should be voting "support". Glosses have a high frequency of usage, and this vote will also standardize across templates the parameter used for them. --WikiTiki89 23:24, 20 April 2017 (UTC)
 * The issue of whether to add t= to all templates (which I agree with) is independent of whether to deprecate gloss= or 4/5=. Benwing2 (talk) 00:31, 21 April 2017 (UTC)
 * A "strict policy of standardization" would mean getting rid of 4/5= because they don't (and can't) exist in many other templates. --WikiTiki89 10:09, 21 April 2017 (UTC)
 * I did not speak about "a strict policy of standardisation", I spoke of "a strict policy of standardisation of abbreviations". I stand by my vote as before: There's no benefit in removing options for nothing in return. Korn &#91;kʰũːɘ̃n&#93; (talk) 10:33, 21 April 2017 (UTC)
 * But you also said "any solution is preferable to the template having an empty parameter in its majority/default case". --WikiTiki89 10:35, 21 April 2017 (UTC)
 * If the last parameter is empty, it is absent and thus no waste of typing. Or is there a misunderstanding and we're talking about different things? Korn &#91;kʰũːɘ̃n&#93; (talk) 11:35, 21 April 2017 (UTC)
 * The only case I see here with an empty parameter is . I'm not sure what you were thinking of. --WikiTiki89 11:51, 21 April 2017 (UTC)

Wait, I got confused, this is what I'm for: Deprecate the unnamed alt form (parameters 3 in m/l and 4 in der/inh) in favour of alt= (or a=) and then move up the gloss from parameter 4/5 to 3/4. Reasoning being as before: The alt form is now mostly handled by a module and used much more rarely than the translation and its implementation as a non‐final parameter leads to what feels like 90% of our templates having empty bars. Korn &#91;kʰũːɘ̃n&#93; (talk) 10:31, 22 April 2017 (UTC)
 * So, you mean ? --Victar (talk) 16:00, 22 April 2017 (UTC)
 * Yes. Korn &#91;kʰũːɘ̃n&#93; (talk) 21:10, 22 April 2017 (UTC)
 * That's probably better formatted as . But still, you said "There is no immaculate solution, but any solution is preferable to the template having an empty parameter in its majority/default case." which implies that you would support either solution, as long as it gets rid of the empty parameter. --WikiTiki89 02:37, 23 April 2017 (UTC)
 * If 'either' refers to 'removing the alt form or removing the gloss', then no. The gloss/translation is normally used, the alt form is normally empty. So I'm for saving two characters (t=) when entering the gloss, I'm for removing the empty parameter of the alt form. (If this vote passes, it has a similarly beneficial effect, but it's not what I'd choose given the option.) Korn &#91;kʰũːɘ̃n&#93; (talk) 09:11, 24 April 2017 (UTC)
 * I kind of like that proposal, aside from the fact that you have to put the parameters in an odd order when there's display text (as says), for the code to make sense. There's a tradeoff when optimizing the code for one case; it causes the other cases to be worse off.  looks good, but so does . Confusing. — Eru·tuon 10:47, 26 April 2017 (UTC)
 * If you have more than one named parameter in the template, it's better to use the equals workaround that way the displayed order is the same as the entered order, I fixed that for you, but it's now unclear which order you intended so if you intended the other order, please switch it.  is much better especially when you compare to the old dilemma of  vs . When you take into account that code is dynamic and people are always adding and removing parameters, the named parameter t makes life a heck of a lot easier. --WikiTiki89 14:25, 26 April 2017 (UTC)
 * Right: what I was saying is that the t parameter is great when there's transliteration, but without transliteration it's not as elegant as 's proposal above, of using 2 for translation. — Eru·tuon 07:49, 3 May 2017 (UTC)
 * Actually Korn's suggestion works fine even with transliterations, but it's a more radical change, and also would not work for templates like . I think that the consistency of having one single way to provide a gloss across all these related templates is more important than saving one or two characters. --WikiTiki89 14:57, 3 May 2017 (UTC)
 * Ever since and  started ignoring diacritical marks for Latin, etc., 3 has been mostly wasted space, so I would also support . --Victar (talk) 15:57, 26 April 2017 (UTC)

Decision
Result: No consensus.

Overview: --Daniel Carrero (talk) 00:06, 5 May 2017 (UTC)
 * Question 1
 * No consensus: 6-8-3 (42.86%-57.14%)
 * Question 2
 * No consensus: 6-5-5 (54.54...%-45.45...%)
 * An aside: I don't think it a good practice to state two percentages. In my view, only the percentage of support / support + oppose should be stated. Furthermore, 43%, less than 50%, is what we call "fail" rather than "no consensus". --Dan Polansky (talk) 07:53, 16 May 2017 (UTC)
 * Actually, I've been wondering what the numbers and percentages mean too. I suppose the numbers are those in support, those opposing, and those abstaining? What about the percentages then? Does a proposal pass when those in support count for more than 50% of the total number of votes, including abstentions? (I'm guessing this is the rule.) "Voting policy" is silent on this issue. — SMUconlaw (talk) 14:01, 16 May 2017 (UTC)
 * : I absolutely think that Voting policy should mention how many votes are needed for a proposal to pass.
 * In reality, a vote passes when this happens: Abstain votes don't count. The proposal needs at least 2/3 (66.6666...%) of support votes. I believe that's the general consensus, but ironically we don't have a voted rule saying that this is true. I seem to recall that very few votes were closed as "failed" above that threshold or "passed" below that threshold, but apparently that general rule still stands.
 * The percentages above mean: How much support, and how much oppose. Abstain votes are ignored. I'm basically always using this system, but I'm pretty sure I never invented it. I probably copied what other people were doing. I don't mind using only one percentage if people think that's better.
 * But I disagree about the "fail". 43-ish percent support means that the number of supports and opposes are very similar. That's "no consensus" by definition. (If we get technical, everything from 0% until the passing threshold can be interchangeably described as either "no consensus" or "fail", but when the supports and opposes are of a similar number, "no consensus" is better.) When I completely revised and edited Votes/Timeline, I used "No consensus" for fails above 40%. I even said so at the introduction of the page. --Daniel Carrero (talk) 17:47, 16 May 2017 (UTC)