User talk:Rhododendrites/signature rfc drafting

Scope

There are other controversial elements of the signatures guideline which may be useful to revisit, but which are out of scope for this RfC IMO. These include what kinds of links are permitted (whether more or less than user/talk/contribs is ok), whether other text can be included in the signature (a message before, after, or in between the user/talk/contribs), what kinds of colors or effects are allowed in a signature, and so on. Let's keep this focused on username-displayed text correspondence for now. — Rhododendrites ^talk \\ 13:54, 25 May 2021 (UTC)

I'm not clear on the reason for the "Except in cases where a username uses non-Latin script" clause in the question. One of the issues with non-account sigs is pings; even if the name uses non-Latin script, it's still what must be typed for a ping. I think that clause should be removed from the question; instead, perhaps a new B: The username must appear in its entirety, without changes. If the user name is in non-Latin script, a Latinized version of the username can be listed after the actual username (or something like that). Schazjmd (talk) 15:12, 25 May 2021 (UTC)

I separated it for two reasons: first, because WP:NLS is a separate part of the guideline, but also I didn't think it was particularly controversial in that one instance. Perhaps I'm wrong. — Rhododendrites ^talk \\ 15:20, 25 May 2021 (UTC)

@Schazjmd: what do you think of this? (in the interest of modularity) — Rhododendrites ^talk \\ 15:27, 25 May 2021 (UTC)

That works, thanks! Schazjmd (talk) 15:41, 25 May 2021 (UTC)

If/when you decide to post the RfC, please mention the location here. I have to unwatch Wikipedia talk:Signatures, it's just a vicious battlefield now. Schazjmd (talk) 16:22, 26 May 2021 (UTC)

Structure

Multi-part questions can be messy, but given the variables being discussed it seems like breaking it up this way would be most likely to yield fruit. Happy to see other options, though. — Rhododendrites ^talk \\ 13:57, 25 May 2021 (UTC)

Suggest any RFC be on the fundamental questions or else offer just 2 choices. 3 or more choices would create a math problem. North8000 (talk) 15:39, 25 May 2021 (UTC)

@North8000: copied this here. I don't disagree with you. That's why there's first a yes/no, followed by the select all you agree with part. The problem is there are many interpretations here, and I don't think it's possible to distill them into a single question with two choices. Definitely open to it if you have an idea, though. — Rhododendrites ^talk \\ 15:56, 25 May 2021 (UTC)

Cool. If you ask people to weigh in on every choice, that also avoids the math problem. If you ask people to pick one "most preferred" item from a list of three or more (or if that ends up happening), then you can get distortions in the result.North8000 (talk) 16:13, 25 May 2021 (UTC)

Proposal to modify the main question and to prune some options

I wonder:

• Are a significant number of !voters going to argue 'no' on the main question? The guideline already says that a customised signature should make it easy to identify your username and that it is common practice for a signature to resemble to some degree the username it represents (btw: also note the British English here). The main question rather seems to be whether the guideline should be more specific in its requirements for how to make sure the username remains identifiable. Perhaps the main question should be something like Should the guideline on signatures contain more specific requirements for making usernames easy to identify within customised signatures? If yes, [...].
• For the same reasons, option D may be superfluous: that abbreviations or variations should make the username easily identifiable is already part of the current guideline, and those who !vote 'yes' on the main question could be expected to already agree with it.
• Is option E really needed here? This seems more like an extra requirement upon B/C that should really be the subject of a separate RfC if B and/or C are withheld.
• Isn't option G only relevant to those who chose C? A Latinization would be a variation, and only if variations are allowed would a Latinization instead of the username be acceptable. We could simplify by removing options F and G, by appending the F clause to options A and B, and by changing option C to read A signature can include a variation (including a Latinization) of the username instead of the full username.
• Is option H really relevant? The current guideline merely encourages us to include a user page link, but in any case what is relevant is always that it remains clear what the actual username is. I think that a more relevant last option would ask whether we should like to require that Any additional characters, text, links, or markup should not render it unclear which part of the signature is the username.

If you want, I could edit the draft to include all these proposed changes, just to show how it would look. Apaugasma (talk|contribs) 18:49, 25 May 2021 (UTC)

I see a difference between Latinizations and other variations. The wording of C implies that the variation should be, if not immediately recognizable, at least clear once you learn what it means. If I signed as T@/\/\≥!n, that connection would be clear enough; even if someone didn't make the connection to "Tamzin" their first time seeing it, they'd probably remember it in the future. But if my username were טמזין and I signed as Tamzin, the connection would only be clear to people who can read Hebrew, and they. Others wouldn't necessarily remember which editor with a Hebrew username is the one who goes by Tamzin. So I could see someone supporting C but not G.

I will add a nitpick on F, though: the clause but can additionally include a Latinized version seems superfluous in light of the disclaimer after the list of options. Every option except H (and in some cases E) allows for stuff like TZ, aka Tamzin or nizmaT (Tamzin) or unrelatedNickname (User:Tamzin). -- Tamzin (she/they, no pref.) | o toki tawa mi. 19:11, 25 May 2021 (UTC) c/e 19:50, 25 May 2021 (UTC)

You're right that someone could support C but not G. Perhaps we should keep F and G, but mark it as a separate decision by calling it X and Y. I'm mainly trying to simplify here, but you're right to point out the limits to that. I would keep F in its current longer version though: strictly taken, it's superfluous, but it avoids creating the wrong impression that F would escalate into disallowing the inclusion of a Latinized version. Apaugasma (talk|contribs) 19:38, 25 May 2021 (UTC)

Are a significant number of !voters going to argue 'no' - Not sure. It's worded to try to get as much agreement as possible, of course, but also so that consensus for "no" would have clear implications for the guideline. Namely, that it would mean those two bulletpoints cease to reflect consensus and be removed. Note that we don't have a language proposal here for an addition/change of bulletpoints but an idea which the bulletpoints would then need to reflect (and regardless of my spelling, I don't have a problem with the guideline being in British English :) ).

option D may be superfluous: that abbreviations or variations should make the username easily identifiable is already part of the current guideline - but we have a lot of people who just don't abide by that guideline. So the idea is to ask where consensus is so that we can modify (or leave alone) the guideline accordingly.

should really be the subject of a separate RfC if B and/or C are withheld. yeah that probably makes sense. Tamzin suggested this one, and I think it makes sense, but it might be a little of out scope.

Isn't option G only relevant to those who chose C? - yes. my assumption is that some people would be ok with a variation for a non-Latin username, but not a different version of a username that uses Latin script. Is that more unlikely than I imagine?

Is option H really relevant? - This was, as I understand it, part of BrownHairedGirl's primary goal. — Rhododendrites ^talk \\ 02:41, 26 May 2021 (UTC)

@Rhododendrites: I think it would be better to try and not make this RfC about to many things. If it really is a question in and of itself whether the current guideline should be enforced, I strongly suggest striking all the rest for the moment. Asking whether the guideline should be more strict, and in what ways, really just doesn't make any sense if there's even no consensus to enforce it. Better then to explicitly ask whether the current bullet points should be enforced (like every guideline, with common sense, and occasional exceptions may apply), or removed. When there's a consensus to enforce, we could perhaps have a second RfC on what extra requirements we want. Apaugasma (talk|☉) 03:44, 26 May 2021 (UTC)

Applicability option three

Frankly, I don't see the point of having that option available. What's the point of having an RfC that won't even have teeth in name? A purely advisory RfC feels like a waste of time, sorry. Sdrqaz (talk) 19:45, 25 May 2021 (UTC)

@Sdrqaz: I agree in principle, but at least one person in the other thread (and others elsewhere) did say that it's ok to have a guideline which suggests what to do without making anything mandatory. I think that causes too many problems, myself, but that's why it's there. Does that make sense, or do you think it's still pointless? — Rhododendrites ^talk \\ 02:26, 26 May 2021 (UTC)

@Rhododendrites: (Not sure why the ping didn't work, but replying regardless) I understand why it was included, but the default response to "you're violating a guideline" is often "guidelines are optional, also see BURO and IAR". Many editors view guidelines as being essentially optional. Watering them down further would not be helpful at all. The language of the first question could be altered to make it "must" instead of "should", if some teeth are to be added. Sdrqaz (talk) 08:23, 27 May 2021 (UTC)

• Amend 3 to read "Wikipedia:Signatures remains a guideline and not instructions". doktorb _words^deeds 20:04, 25 May 2021 (UTC)
  • Copied this here, Doktorbuk (and sorry, I removed your other comment which looked to be a response, but we're not settled on the questions yet :) ). — Rhododendrites ^talk \\ 02:26, 26 May 2021 (UTC)

New version

Considering some of the calls for a simpler proposal, I just added a different configuration with just three questions.

Pinging for thoughts: @Sdrqaz, Schazjmd, Apuagasma, Tamzin, and North8000: — Rhododendrites ^talk \\ 14:39, 29 May 2021 (UTC)

Oops. Fixing a ping, sorry: Apaugasma — Rhododendrites ^talk \\ 14:42, 29 May 2021 (UTC)

One more: Doktorbuk — Rhododendrites ^talk \\ 14:43, 29 May 2021 (UTC)

Thanks @Rhododendrites:. I want to focus - later, I'm pressed for time right now - on the second question. Over on the signature talk page @ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ: has posted something interesting about non-Latin characters and I think that's a direction we may need to take "variations" in. We have to be clear that the use of non-Latin characters is permitted with "variations". I'll have a think about this later on, be interesting to see what people think. doktorb _words^deeds 15:04, 29 May 2021 (UTC)

There is a note about this on question 3. The idea being to get a sense if variations/abbreviations are ok first, and then to see if they need to be recognizable (and whether non-latin script should be an exception). — Rhododendrites ^talk \\ 15:11, 29 May 2021 (UTC)

I'm not sure how that comment relates to question 2. The commenter wanted to ensure they could always use their full, actual user name. isaacl (talk) 15:27, 29 May 2021 (UTC)

I think the whole issue comes down to the new question #1. If consensus is "yes" on #1, then #2 and #3 must = "no", since they explicitly don't require the literal user name. (This question does not rule out stylized text or other text in the signature. in #1 allows for Latinized "translations" of non-Latin names.) If consensus on #1 is "no", then #2 must = "yes" since the community is saying we don't need the literal user name in the signature, so anything is acceptable. #3 just introduces a new point of contention where editors argue whether something is "easily recognizable" or not. Schazjmd (talk) 15:34, 29 May 2021 (UTC)

To editor Rhododendrites: I and a few others spent the last two months clearing out Special:LintErrors/obsolete-html-tag in Template namespace. There were about 18k errors mainly because of DYK nomination templates that had custom signatures. This was problematic because the presence of so many DYK templates drowned out the "real" templates that were important to fix. Most of the signatures needed <font>...</font> replaced with <span>...</span>. However your name is memorable because it was one of the few that needed <tt>...</tt> replaced with <samp>...</samp>. Inevitably at some point in future even span and samp tags may be made obsolete. At that point we will have to either replace the tags all over again or decide it is not worth it and just them let rot. Both of these are undesirable and such scenario wouldn't be present for plain old default signatures. This is what I meant by custom signatures creating maintenance overhead in future. However I also know that there is no chance of getting community concensus against using custom signatures. It is a Mediawiki issue and probably outside the power of English Wikipedia to disable the custom signature feature. ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ (talk) 16:02, 29 May 2021 (UTC)

@Rhododendrites: I believe this RfC captures the issues well. Perhaps questions 1 and 2 can indeed be collapsed into one question, since everyone will vote in one of two ways on the two of them (either 'yes' and 'no' or 'no' and 'yes', respectively). 'No' voters on question 1 will have a real choice to answer 'yes' or 'no' to question 3 though, so that one should stay as it is (answering 'yes' will indeed introduce a rather subjective point of contention, but to introduce such would be the voters' choice).

If there is a real interest in having a clearer and enforceable guideline, I think this RfC will be successful. However, I have come to believe that a majority among experienced and active editors doesn't care about the guideline because in no case would they want to see it actually enforced. If there's no will to have an enforceable guideline, there's little point in asking what the guideline should be. That's why I would recommend starting with an RfC that addresses that question first. But then perhaps the question whether a guideline should be enforced can't be the subject of a RfC, because that is normally decided by letting community practice speak for itself. That's why I have also come to the view that the one question we should ask first is whether we want to have a policy about the relation between customized signatures and usernames (e.g., a policy stipulating that customized signatures should display usernames in their entirety, or one that explicitly allows abbreviations or variations of the username, one that requires abbreviations and variations to be easily recognizable, etc.). I believe that a clear 'yes' to that question would make an RfC about what that policy should be much more productive, and a clear 'no' would serve as a good indication that we'd be better off directing our efforts elsewhere. ☿ Apaugasma (talk ☉) 19:00, 29 May 2021 (UTC)

I think if the guideline were clearer, it would be enforced. Also, I think some people assume it's out of step with practice/consensus, and that may well be true. My goal, as I've said, is more to ensure this reflects consensus than to enact new rules. If the rules as written aren't enforced because they are out of step with practice, we should just get rid of them. — Rhododendrites ^talk \\ 04:12, 30 May 2021 (UTC)

You're right; I overlooked that answering 'no' on question 2 (previously question 3) also allows for striking the whole guideline on usernames having to be identifiable, which asking whether there should be a policy about this wouldn't. That too would be a sensible outcome. If, however, we get 'no'-'yes' (as I kind of expect), it will have been a fruitless exercise. But then I do think it's worth trying to see whether there's an appetite for change, and this RfC does that well. ☿ Apaugasma (talk ☉) 05:41, 30 May 2021 (UTC)

Finally got around to replying to this; sorry for the delay. I'd say the new more streamlined RfC is an improvement on the previous iteration, mainly because it's an easier RfC to close and it's easier for a !voter's brain to wrap around. I do have concerns over gaming the "easily recognizable" clause, though. Looking through WP:ARBHIST, of our current members of the Committee, does David Fuchs' signature fulfil that requirement? Does Katie's? Does Worm's? I'd argue "maybe", "maybe/yes", and "definitely". How about the ever-changing signature of Praxidicae? Are the variations used by that user easily linked to the username? I'm not sure. I'd say it's not, but I think there will be debates down the line whether that is in line.

I'd also like to say that voting "yes" for Q1 should not mean abstention from Q2, and that misconception may need to be cleared up. Sdrqaz (talk) 16:26, 30 May 2021 (UTC)

@Sdrqaz: I don't anticipate there being a way to avoid edge cases that would need to be determined on a case-by-case basis. I would guess that if we have on the books that usernames do not have to correspond exactly, and that variations are allowed as long as they resemble in some way the username, then that would cut down on most complaints. The current problem is the existing guideline is easy to interpret as saying that those edge cases are simply disallowed. As I was tweaking the wording the other day, I started adding some examples of what would/wouldn't be allowed, but left them out because that seemed foolish. There are just too many possibilities, and beyond that, I don't want one of my examples to throw off a potentially useful RfC. — Rhododendrites ^talk \\ 16:44, 30 May 2021 (UTC)

Hmm. Thanks for the reply; I think you're right. I would still encourage all !voters to make their intentions with regards to Q2 clear: I have far stronger opinions on that question than on Q1 (and frankly doubt Q1's chances of passing). Sdrqaz (talk) 16:59, 30 May 2021 (UTC)

Based on some comments above, I've effectively combined 1 + 2. I've also restored the question about who this should apply to. I fully expect this to be shot down, but out of respect for the several people who have expressed support for some sort of "grandfathering," it seems like a good idea to include it. — Rhododendrites ^talk \\ 04:12, 30 May 2021 (UTC)

Good changes, Rhododendrites, this latest wording removes the logic problem I had with the previous version. (And thanks for all the work you're doing on this!) Schazjmd (talk) 15:19, 30 May 2021 (UTC)

I've added a background section to clarify the purpose. Curious about thoughts there. I'm wary of long RfC set-ups, but it also seems useful... — Rhododendrites ^talk \\ 17:07, 30 May 2021 (UTC)

Thanks for the ping. I've been off the grid for three days and only have a few minutes today and this is complicated. My quick comment today is to not be afraid of spending extra time to get the RFC put together well. If it's not too late I can do a better job tomorrow. Sincerely, North8000 (talk) 02:41, 1 June 2021 (UTC)

The one thing that i do have a strong opinion on is that, as you have done, you should ask for opinions on every idea, not just pick their most preferred. Regarding grandfathering, I suspect that the reality is that we have like 20 or 40 well established editors who's signature doesn't show their user name who are influential enough to stop this and where the situation is distortingly non-typical in that they are so prominent that lots of people know their user name from their sig. A pragmatic white flag on this would be to grandfather people who had the "violating" sig 5 years ago and had 10,000 edits 5 years ago. North8000 (talk) 01:36, 2 June 2021 (UTC)