Wikipedia:Village pump (proposals)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

• Table of contents
• First discussion
• End of page
• New post

New ideas and proposals are discussed here. Before submitting:

• Check to see whether your proposal is already described at Perennial proposals. You may also wish to search the FAQ.
• Consider developing your proposal at Village pump (idea lab).
• Proposed software changes should be filed at Phabricator (configuration changes should have gained a consensus).
• Proposed policy changes belong at Village pump (policy).
• Proposed WikiProjects or task forces may be submitted at Wikipedia:WikiProject Council/Proposals.
• Proposed new wikis belong at meta:Proposals for new projects.
• Proposed new articles belong at Wikipedia:Requested articles.

« Archives, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214

"Datebot" (limited scope)

This is a proposal to increase date consistency in citations, per MOS:DATEUNIFY. Specifically, a bot would look for templates like

• {{Use dmy dates}}
• {{Use mdy dates}}

and bring present dates in line with desired usage on that specific article. I would make a WP:BOTREQ for this, but I'd rather go in with a consensus that this is desired, and what scope is OK with the community. Note that the below is just the general idea/goal, there will be odd cases and kinks that need to be ironed out. That's what the bot trial would help with. If nothing reliable can be made, then the bot wouldn't get approved. Headbomb {t · c · p · b} 21:25, 17 December 2018 (UTC)

Option A: extreme conservativeness

Under this option (if option B is not accepted), the bot would not touch dates elsewhere in the article, quotes, or mess with valid format (21 Jan 2009 → 21 January 2009), it would simply those in |date=/|access-date=/similar parameters of the various {{cite xxx}} templates. For example, if {{Use dmy dates}} is used in an article, then the bot would convert

• Smith, J. (2006). "Article of things, part 1". Journal of Things. 1 (2): 3–4. Retrieved March 10, 2009.
• Smith, J. (March 23, 2007). "Article of things, part 2". Journal of Things. 2 (2): 3–4. Retrieved March 10, 2009.
• Smith, J. (2008). "Article of things, part 3". Journal of Things. 3 (2): 3–4. Retrieved 2009-10-29.
• Smith, J. (Jan 21, 2009). "Article of things, part 4". Journal of Things. 4 (2): 3–4. Retrieved March 29, 2017.

to

• Smith, J. (2006). "Article of things, part 1". Journal of Things. 1 (2): 3–4. Retrieved 10 March 2009.
• Smith, J. (23 March 2007). "Article of things, part 2". Journal of Things. 2 (2): 3–4. Retrieved 10 March 2009.
• Smith, J. (2008). "Article of things, part 3". Journal of Things. 3 (2): 3–4. Retrieved 2009-10-29.
• Smith, J. (21 Jan 2009). "Article of things, part 4". Journal of Things. 4 (2): 3–4. Retrieved 29 March 2017.

and if {{Use mdy dates}} is used, then it would convert to

• Smith, J. (2006). "Article of things, part 1". Journal of Things. 1 (2): 3–4. Retrieved March 10, 2009.
• Smith, J. (March 23, 2007). "Article of things, part 2". Journal of Things. 2 (2): 3–4. Retrieved March 10, 2009.
• Smith, J. (2008). "Article of things, part 3". Journal of Things. 3 (2): 3–4. Retrieved 2009-10-29.
• Smith, J. (Jan 21, 2009). "Article of things, part 4". Journal of Things. 4 (2): 3–4. Retrieved March 29, 2017.

If no {{Use mdy dates}}/{{Use dmy dates}} is found, the bot would do nothing.

Option B: enforce majority use

In addition to the above, the bot could also determine the majority use and further normalize the article. For instance, if 23 citation used something like |date=21 January 2009 /|access-date=25 February 2011, 2 citation used |date=21 Jan 2009, and 3 citations used |access-date=2013-02-25, then it would bring the minority cases in line with the majority. For example the above would be normalize to

• Smith, J. (2006). "Article of things, part 1". Journal of Things. 1 (2): 3–4. Retrieved 10 March 2009.
• Smith, J. (23 March 2007). "Article of things, part 2". Journal of Things. 2 (2): 3–4. Retrieved 10 March 2009.
• Smith, J. (2008). "Article of things, part 3". Journal of Things. 3 (2): 3–4. Retrieved 29 October 2009.
• Smith, J. (21 January 2009). "Article of things, part 4". Journal of Things. 4 (2): 3–4. Retrieved 29 March 2017.

Ties would result in no action since the bot could not determine which variant is preferred.

Option C: Status quo

Leave the mess to be dealt with by humans.

!Vote (Datebot)

• Support both A and B as proposer. Headbomb {t · c · p · b} 21:25, 17 December 2018 (UTC)
• Support both A and B - good idea. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:01, 17 December 2018 (UTC)
• Support both A and B – sounds like a good idea, little potential to go horribly wrong, hopefully "Datebot" would not be confused too often with DatBot. SemiHypercube 22:10, 17 December 2018 (UTC)
• C at least until questions below are gone through. — xaosflux ^Talk 03:29, 18 December 2018 (UTC)

Discussion (Datebot)

The description seems deficient. For option A, for the first list of changes, there is no indication whether one of the use xxx dates template is present. I oppose option B because the bot could encounter the article at a moment when it does not reflect the long-standing consensus. Also, it would be inappropriate to have different criteria for bot edits than for manual edits; if a bot were allowed to change on the basis of the majority of the dates in citations, it would be OK for manual edits too. But currently it not OK for manual edits. Jc3s5h (talk) 21:54, 17 December 2018 (UTC)

Jc3s5h (talk · contribs) I clarified a few things. Some words were missing. Also, it is not only ok, but encouraged for humans to do those sort of changes. Headbomb {t · c · p · b} 21:58, 17 December 2018 (UTC)

The expectations in MOS:DATEUNIFY would indicate an editor should not change an access-date from 2008-11-18 to November 18, 2008 just because the majority of all citation-related dates was mdy, so long as all or nearly all access-dates used the YYYY-MM-DD format. Also, if the article almost exclusively used mdy in the article overall, because most of the citations only gave a year, just counting a handful of citations that did give a month, date, and year could easily give a false result. Jc3s5h (talk) 22:21, 17 December 2018 (UTC)

If the citations don't match the body of the article, then the body didn't really make the article worse than it was before, since it was already inconsistent. If it normalized the citations the 'wrong way', just slap either {{Use dmy dates}} or {{Use mdy dates}} on it and you'll ensure the long-term stability of the article dates. Headbomb {t · c · p · b} 03:48, 18 December 2018 (UTC)

We have had a recent discussion which rejected the expectation that the access date be in the same format as any other dates (only that the access dates be unified among themselves was a consensus there). That's this discussion, which, probably should have been closed as "consensus against" the proposal rather than the "no consensus" close by Compassionate. --Izno (talk) 00:25, 18 December 2018 (UTC)

Uggh is this really a mess that needs to be dealt with at all? Citation parameters aren't "prose" they are "data" - they should just be in ISO format. Can you show some good examples of versions of an article that have this "problem"? — xaosflux ^Talk 03:30, 18 December 2018 (UTC)

ISO/YYYY-MM-DD are relatively rare in dates save for accessdates. But for a good example, see Cancer, with several inconsistent dates the bot could fix. Headbomb {t · c · p · b} 03:51, 18 December 2018 (UTC)

@Headbomb: OK, so in Cancer we have:

A few examples

Bast RC, Croe CM, Hait WN, Hong WK, Kufe DW, Piccart-Gebhart M, Pollock RE, Weichselbaum RR, Yang H, Holland JF (3 October 2016). Holland-Frei Cancer Medicine. Wiley. ISBN 978-1-118-93469-2. Kleinsmith LJ (2006). Principles of cancer biology. Pearson Benjamin Cummings. ISBN 978-0-8053-4003-7. Mukherjee, Siddhartha (16 November 2010). The Emperor of All Maladies: A Biography of Cancer. Simon & Schuster. ISBN 978-1-4391-0795-9. Retrieved August 7, 2013. Pazdur R, Camphausen KA, Wagman LD, Hoskins WJ (May 2009). Cancer Management: A Multidisciplinary Approach. Cmp United Business Media. ISBN 978-1-891483-62-2. Cancer at Google Books. Archived from the original on 15 May 2009. Tannock I (2005). The basic science of oncology. McGraw-Hill Professional. ISBN 978-0-07-138774-3. Schwab M (23 September 2008). Encyclopedia of Cancer. Springer Science & Business Media. ISBN 978-3-540-36847-2.

What will be changed there - the access date in #3? — xaosflux ^Talk 04:10, 18 December 2018 (UTC)

@Xaosflux: diff. Headbomb {t · c · p · b} 04:39, 18 December 2018 (UTC)

@Headbomb: thanks for the example, any idea what sort of volume you are trying to address? I'm only seeing this minimally beneficial to readers, with month names already spelled out it isn't fixing any ambiguity. — xaosflux ^Talk 14:13, 18 December 2018 (UTC)

@Xaosflux: honestly it's pretty hard to gauge. I'd guess that in the ballpark of 10% of articles have date inconsistancies, but I don't know how much of that could be fixed by bot before things are trialed.Headbomb {t · c · p · b} 16:25, 18 December 2018 (UTC)

10%! ~689376 edits just to tweak some date formatting in the references section seems quite excessive to me, are there good arguments for how this will improve usability for readers? — xaosflux ^Talk 17:24, 18 December 2018 (UTC)

Wouldn't be 10% of all articles edits, just ballpark 10% of articles with MOS:DATEUNIFY issues somewhere, only a fraction of which are in citation templates, and only a fraction of which could be addressed by bots. Headbomb {t · c · p · b} 17:43, 18 December 2018 (UTC)

It might be more than 10% if it is fixing mdy/dmy inconsistencies which are common. But if it wasn't a "fast bot" (immediate fix) rather the slow boat to China for a year to process all 5.5m articles, before starting over, changes are saved up and it doesn't thrash watchlists. I think the pro argument is people seem care about fixing date formats so it would save labor - the number of cites being added is exponential compared to number of qualified editors - it keeps getting worse over time. This proposal should be linked from the FA and GA forums as they tend to be most focused and might have concerns we don't know about. -- GreenC 18:10, 18 December 2018 (UTC)

• 2 Questions - 1) If option B was used, which states "in addition to the above", what happens if the DMY/MDY statement conflicts with majority usage? Which wins?

2) What do we do about date style squabbles? Some articles, usually those with both British and American topics involved, have some mind-numbing disputes over date styles. Would this bot freely replace each date set, each time the DMY/MDY requirement is edited? Nosebagbear (talk) 16:14, 28 December 2018 (UTC)

@Nosebagbear: could be several options, but the way I envision things I'd see {{Use dmy dates}}/{{Use mdy dates}} trumping everything else, since someone placed that on purpose. If there's a squabble, decide which of the two to use, and the bot will follow that decision. Headbomb {t · c · p · b} 21:22, 28 December 2018 (UTC)

"Watch this page" should be the default behavior when editing

the proposal

Let's say I edit a wikipedia page. A few month later, somebody modifies my version (not using "undo", just overwriting what I've done). By default, there will be no alert about this. The basic setting should be to alert you, that's it to have the checkbox "Watch this page" checked by default each time you edit a page.

• Comment "Watch this page" should give individual more selective options on what "what the user want to watch". Right now it does not have. Colton Meltzer (talk) 05:44, 28 December 2018 (UTC)

the cons

"Power" editors, editing thousand of pages, would of course get a lot of notifications. But firstly it's not sure that it's unpractical for every power editor and secondly this kind of editors knows how to find their way to disable this option globally (by Preferences>Watchlist>Add pages and files I edit to my watchlist). This option is a bit buried in the options so most users will probably never use this option it but that's probably not the case for "power" editors.

the pros

• It's Counterintuitive

That's what you'll find on most internet boards, blog comment sections and I guess that's the expecting behavior for any user on this kind of sites. You don't have to rummage through the settings to find the option to get notified.

• It's Harmful

It's harmful because it misses the opportunity to encourage "not regular editors" to become regular editors. I guess people editing page quite rarely will completely forget their modification. If they're not alerted for further modifications, most will probably never know that and never come back on the page.

Linuxo (talk) 12:14, 18 December 2018 (UTC)

Thanks for the suggestion, but I don't think it is needed. Some pages I edit, I check the little blue star and others I don't. You soon get into the habit. When your edit count grows you will no longer remember what you add to individual pages and you will learn to trust others to be sensible. There may be corners of WP where I don't tread where this is different- if you are working there check the star. ClemRutter (talk) 12:31, 18 December 2018 (UTC)

For yourself, go to Special:Preferences#mw-prefsection-watchlist and check the checkbox labeled Add pages and files I edit to my watchlist. If you're wanting to have that checkbox be checked by default for new users, it would need an RFC to show community consensus and consideration of the issues discussed in T38316. Anomie⚔ 03:00, 19 December 2018 (UTC)

Just would like to thank you for making me aware of this option. Thanks! WelpThatWorked (talk) 05:23, 20 December 2018 (UTC)

I have my preferences set to place every page I edit on my watchlist. That does have consequences: I now have more than 5,300 items on my watchlist (although I prune it occasionally, so that there are are over 7,000 pages I have edited that are not currently on my watchlist). I spend a good part of every day working through my watchlist. I'm sure there are a lot of editors that do not want, or have time, to deal with that. I think it would be best to leave this as an option. - Donald Albury 13:31, 19 December 2018 (UTC)

Watching every page you edit can easily become overwhelming. Only advanced users will like that. Therefore, I think it's better to keep the functionality opt-in. --NaBUru38 (talk) 14:26, 26 December 2018 (UTC)

If one wants to watch a page then click on the blue star. To set default on all pages which one edits would be overwhelming when edits on new articles increase. CASSIOPEIA^(talk) 14:37, 26 December 2018 (UTC)

"remind me" bot

Overview

Former discussion: Wikipedia:Bot_requests#Remind_me_bot.

The idea is to create a bot/template pair that sends on request a reminder to go back to a particular page. The case use is typically to check back on a discussion on high-traffic pages where a watchlist is inefficient or (on the contrary) when the discussion has very little participation and you want to make sure to come back one week afterwards to assess consensus.

As the reminder would likely take the form of an in-wp post, such a bot needs consensus per WP:BRFA. I believe VPP to be the best (or least worse) forum for that discussion, but do tell if I missed something obvious.

Envisaged technical implementation

A template (e.g. {{remind me}}) that takes one argument, either a fixed date or a duration of time, is left by a user on the page they wish to be reminded of. Template transclusions are watched by a Toolforge bot (running every hour or so?). When the time is up, the bot posts a notification to the requesting user's talk page.

Asking for a notification by this process is public (i.e. other editors know that you asked for an update, when you asked for it etc.).

Depending on the technical details, and if that template gets widespread use, it might become necessary to put limits on the use. I would suggest no more than X pending notifications per user (so that you can put hanging notifications for Christmas 2118 but it eats up one of your slots); I think X=10 is a reasonable starting value if it comes to that, but I do not think it is needed outright.

Possible objections

1. While the bot postings themselves are fairly innocuous (you get 'em only if you ask for 'em), there could still be some clutter in the reminder request templates: e.g. if 100 posters drop just a "remind me" template in a discussion where 5 people are actually participating. I believe (1) that is quite speculative and (2) if it ever becomes a problem, it might be solved by conduct guidelines (e.g. "avoid using this template outside an overwise meaningful post, as this can be considered disruptive").
2. As pointed in the previous discussion, that might duplicate a Community Wishlist item that was among the top 10 (and hence will be worked on by WMF staff). I believe the use case is slightly different (that one is targeted at talk pages, as opposed to articles), and even if it is indeed redundant, I do not see the harm. (As I will be the one handling the bot coding, time wasted duplicating a system can be ignored: I think it would be fun to code. The only question is whether having two systems might be harmful, not useless.)

Comments

I have a better design.

Make a bot that you place on your own talk page (or on a special page if that makes the bot easier to write) with the same fixed date / duration of time arguments.

If I put

 {{remindme|03:14, 19 January 2038 (UTC)|Check [[Talk:Cockcroft–Walton generator]] for replies}}

on the appropriate page It would wait until January of 2038 and then post the following to my talk page:

Remindme bot reminder notice

On 00:00 1 January, 1970 03:14 (UTC), you requested a 03:14, 19 January 2038 (UTC) reminder. The text of the reminder is:

Check Talk:Cockcroft–Walton generator for replies

-- [Bot signature]

That way you could be reminded of anything. I have a similar reminder set locally that I use to remind me to check the edit history of blocked IP vandals a week after the block expires. That way I can report them if they go right back to vandalizing.

To reduce abuse:

The reminder should happen even if some vandal deletes the request.

There should be a way to cancel a reminder.

Users or admins should be able to cancel reminders for that user.

Reminders should be limited to 255 characters.

Users should be limited to no more than 32 open reminder requests.

A FAQ should explain that this is for Wikipedia-related use only; no "don't forget to buy more Vegemite" reminder requests.

If possible, it should be impossible or at least difficult to set a reminder to be shown on someone else's talk page.

--Guy Macon (talk) 14:06, 18 December 2018 (UTC)

@Guy Macon: If this were done on a user "JSON" page (eg User:Example/remindme.json) it would be: simpler for bots to parse requests (I think), restricted to editing by the same user, or administrators (meaning that vandalism is not an issue, and if the reminder is only sent if the request is present at the time specified, allows cancelation), makes it easier to limit requests (only pay attention to the first 32 json objects). Thoughts? --DannyS712 (talk) 00:59, 27 December 2018 (UTC)

Community Wishlist

Hi guys, I might be misreading things, but does this not sound like one of the things in the top then of the community wishlist Article Reminders? Nosebagbear (talk) 19:55, 18 December 2018 (UTC)

It does. IMO people interested should get involved with the creation of that feature than rush to implement a bot to do mostly the same thing. Anomie⚔ 03:02, 19 December 2018 (UTC)

Indeed. I concur with Anomie. I wouldn't spend too much time on a bot/user script right now. The long-term solution might end up being a MediaWiki extension, which would address the possible objection about bot postings (since there'd be no bot involved). I suspect the implementation would allow for reminders about any page. To be clear, this project isn't "owned" by WMF, per se, rather it's now their responsibility. There have been top 10 wishes implemented by volunteers in the past. So, if you are eager to code you most certainly can be a part of it :) We just need to find the right solution first, one that accommodates the necessary workflows and will work for all wikis. — MusikAnimal ^talk 23:34, 30 December 2018 (UTC)

There are some XfD proposal

Everyone are welcomed to comment these RFCs:

1. WT:TfD#Proposal: move template and module rename nominations from RM to TFD
2. WT:AfD#Another "articles for discussion" RFC
3. WT:AfD#RFC on merging
4. WT:CfD#RFC: Move stubtypes (and categories populated by templates currently on TFD) to TFD

Thanks, Hhkohh (talk) 01:17, 23 December 2018 (UTC)

Some are closed. So strike now Hhkohh (talk) 20:01, 23 December 2018 (UTC)

Indefinitely semiprotecting the refdesk

These constant reverts, revdels, and 2-day protects are a bloody waste of resources. All done in the name of DENY, right? It is the opposite. The waste and disruption constitutes feeding the troll. Half the activity now seems to be vandalism and actions against it. In the name of DENY, I suggest some sort of indefinite protection.

And the refdesk is somewhat outside the mission anyhow. It is a bit of a public board and is being abused. So, if it is an essential resource to help improve the mainspace, make it a resource for the community. I know, I know, the IPs. Look at the history of one of the desks. Depriving a good IP once in a blue moon is a small price. Besides, we're here to build the mainspace. The refdesk is diverting our resources away from that. We're not here to sacrifice the mainspace to support what is mainly a public service. Enough is enough. Anna Frodesiak (talk) 22:43, 26 December 2018 (UTC)

Note: I have factored out discussion and !votes about completely shutting down the desks into a separate survey below. Please keep this initial survey focussed on the issue of whether or not to semiprotect them. Fut.Perf. ☼ 11:41, 6 January 2019 (UTC)

Survey

• Support and call for office actions: There's a reason why the main page and articles on contentious subjects are protected in one way or another. While one could argue that the troll (or perhaps his merry band of misfits) would simply divert his attention to other mainspace pages, I do agree that this long-term pattern of gross trolling, disruption and libel is more than crossing the line, and there's definitely something that should be done about it, but there should be more to this than just restricting RD access to trusted users. I believe that an action or two from the Wikimedia legal team should be done especially as there have been trolls who have carried out violent threats and libelous accusations against a number of users and administrators, which whilst they may be empty or done merely for the sake of schadenfreude/"dank" memes, is still by and large considered a felony. Blake Gripling (talk) 00:32, 27 December 2018 (UTC)
• Strong Oppose I understand (if grudgingly) why the Reference Desks are regularly protected. But what is currently happening is suboptimal. These are pages whose goals (at least in theory) is to help in the improvement of articles, and they are a good resource particularly for newcomers who want clarifications on certain article materials. I'd rather accept the RDs be shut down rather than having them restricted to few editors, as doing the latter would kind of defeat their purpose. With that said, I'd strongly support some sort of office action being done, and I'm surprised that it hasn't been done yet actually. Narutolovehinata5 ^tccsdnew 00:54, 27 December 2018 (UTC)
• Support long term protection. The Ref Desks are a poor substitute for Quora and Yahoo Answers or wherever people go for random people's opinions on the internet. We lack the crowd sourced scoring of answers that moves the good answers up and the crap ones down. Nothing would be lost in shutting down the ref desks or at least restricting them to registered autoconfirmed users. I see precious little evidence that anything happening at Ref Desks is going to improving articles. Legacypac (talk) 00:57, 27 December 2018 (UTC)

The options:

• Ref desk - no registration, middle quality answers, moderation: people criticize most bad answers
• Quora - privacy issues, very good and very, very bad answers, both with lots of upvotes
• Stack Exchange - easy registration, very good answers cause of very good moderation, but people are usually rude to new users
• Yahoo Answers - hard to register, absolute garbage answers, unmoderated spam

If anything these are all poor substitutes for the ref desks, not the other way. 78.0.232.155 (talk) 22:12, 30 December 2018 (UTC)

• Support. The refdesk should be denied to the vandal so that it can function normally for logged-in users. The refdesk was not intended to be a public forum supplying answers to questions. Binksternet (talk) 02:10, 27 December 2018 (UTC)
• Support I floated an idea along these lines a month ago. To clarify, my intention at that time was less ambitious, namely that the refdesks would be indefinitely semiprotected when needed, and that an informal arrangement among admins monitoring the area would result in someone unprotecting them at an unpredictable time, say after two or three days. However, the current farce is feeding the troll's habit so long-term protection is needed to resolve the issue. Chatting among editors with a track record of building the encyclopedia is helpful in that it also builds the community, but running an open-door chat forum is not helpful for Wikipedia. Johnuniq (talk) 06:43, 27 December 2018 (UTC)
• Strong Oppose - Being too lazy to clean up the mess is not a valid reason for admins to unilaterally disregard WP:PILLARS and turn a section of Wikipedia into a secret clubhouse that only they can participate in. 79.134.4.188 (talk) 09:19, 27 December 2018 (UTC)

Unsurprisingly, this IP comment was made from an open proxy. Favonian (talk) 09:35, 27 December 2018 (UTC)

Please WP:Assume Good Faith there are many reasons to use an open proxy. Zell Faze (talk) 19:18, 31 December 2018 (UTC)

...and many more reasons to block them as soon as we identify them. If you actually have a good reason to use a proxy, you can apply for Wikipedia:IP block exemption.

@Favonian: What the hell is that supposed to mean? 142.160.89.97 (talk) 05:20, 6 January 2019 (UTC)

• The Elephant In The Room: There is a cultural issue here. A small number of refdesk regulars act a certain way - a pattern of behavior that has been proven to attract trolls on a wide variety of online forums. The helpdesk regulars act in a different way - a pattern of behavior that has been proven to discourage trolls on a wide variety of online forums. The troll problem is a problem of our own making. --Guy Macon (talk) 09:33, 27 December 2018 (UTC)
• Support protection as the least of several evils. If protection refocuses the RefDesks, making them more useful to the encyclopedia, they may even be worth saving. Favonian (talk) 09:38, 27 December 2018 (UTC)
• Support To ind protect a page from being constant IPs vandalized edits is reasonable and the trade off is small. CASSIOPEIA^(talk) 12:45, 27 December 2018 (UTC)
• Support long-term semiprotection. No other page on this project has had such an enormous amount of disruption from anon IPs or new accounts (even before the recent pattern of vandalism), to be weighed against such a tiny amount of benefit from legitimate, genuine newbie edits, and there's no reason at all not to take the protective measures that would have been taken long ago had such problems occurred anywhere else on the project. There's been an irrational opposition to such protective measures from the side of some refdesk regulars, who tend to vastly over-estimate the value and importance of the desks in general and of their importance to anon editors in particular. The truth is: while the refdesks are a nice-to-have, they are of very minor importance in the overall mission of Wikipedia; the number of people who end up asking questions there is tiny compared to the number of visitors who use Wikipedia as a source of information in other ways; and the number of genuine, legitimate IP editors or newbies among these few is absolutely marginal. The vast majority of IP traffic on the desks, even before the activity of the current vandal(s), has always been from (a) trolls, (b) vandals, (c) a small number of serial long-term ban evaders, (d) two or three legitimate long-term IP users, who, while legitimate participants in principle, have known for years they'd only need to create an account to be immune from the protections, and if they still haven't done so they only have themselves to blame if they end up excluded. Finally, the current pattern of almost daily short-term protections simply isn't working. Protecting for 12 hours every day only gives the vandal his daily reward and another time to look forward to, when he can try the same thing again the next day. We need protection for a time long enough that there will at least be a realistic chance the vandals may get bored waiting in between and find something else to do with their sorry little lives. Fut.Perf. ☼ 14:22, 27 December 2018 (UTC)
• Support: Future Perfect at Sunrise's argument above has convinced me. --Guy Macon (talk) 17:32, 27 December 2018 (UTC)
• Shut Them down and let them move to Wikiversity (first choice) or Protect indefinitely (second choice) per Iridescent's compelling argument below. --Guy Macon (talk) 22:28, 5 January 2019 (UTC)
• not support the protection is crippling the ref desks, and few people ask questions when they are protected. Really we need another solution. Graeme Bartlett (talk) 00:32, 28 December 2018 (UTC)

• Moving the refdesks to Wikiversity would be a solution... I' just saying. --Guy Macon (talk) 22:28, 5 January 2019 (UTC)

• Oppose per pillars 1 and 3 and Jimbo's 2nd and 3rd principles. Wikipedia is an encyclopedia that anyone can edit. The refdesks are meant to operate like library reference desks, but Wikipedia is not a library. I think it, however, is a net positive in fielding questions from readers and new editors to help us improve coverage to meet their needs. Indefinitely removing the ability for unregistered users to participate eliminates that main benefit, and makes it hard for me to support a process that is both not encyclopedic and not freely editable. I do support office actions like a cease and desist letter to the ISP of the troll(s) Wugapodes [t^hɑk] [ˈkan.ˌʧɹɪbz] 02:05, 28 December 2018 (UTC)

Please face the reality of how ref desks really operate, not some wishful thinking. Why do we need a library ref desk for a place that is not a library? Legacypac (talk) 04:06, 28 December 2018 (UTC)

Exactly, why do we need a library ref desk for a place that is not a library? What encyclopedic purpose does the ref desk serve that the Teahouse, the list of portals at the top of the mainpage, the help desk, wikiprojects, and article talk pages don't already? The Ref Desk Guidelines even point anyone having problems with using wikipedia to the help desk, so it's not for problems with Wikipedia itself. It says that it won't do homework for posters, but says editors at the ref desk may help with understanding and pointing to information to help with homework: "The reference desk is not a service that will do homework for others. It should be made clear to questioners that we will give assistance in interpreting questions, help with ideas and concepts, and attempt to point them to resources that might help them to complete their tasks, but that in the end they should do the actual work themselves." That definitely is not the point of an encyclopedia or its editors. The main page lists the ref desk as "Serving as virtual librarians". This isn't a library. It's not a forum to answer general questions about topics. But let's face the reality of how ref desks really operate and look at some questions recently answered on the ref desk:

-"I've been browsing photos taken by some high-end smartphones (by Samsung Galaxy S8 and by Huawei P20 Pro which uses Leica) and at full size nearly all appear watercolorish and grainy - basically just like on their inexpensive counterparts, including mine. What causes it?" 15 Dec under Misc (answers seem to indicate this isn't a phenomenon to be acted upon in articles)

-"I want to write a letter to MaryEllen Elia, the current education commissioner, but I don't know her title. Is is Madam Commissioner or perhaps something else?" 19 Dec under Misc (subsequent changes to the article do not take this question into account)

-"This a real problem I will face soon. I have a pile of 20 banknotes. How many I need to check for counterfeit in order to get 95% that the entire pile is not a counterfeit? And a side question - how to calculate this for any n banknotes with given p percent certainty?" 24 Dec under Math (no linked article that would be improved by this question)

-"I wonder if there are any soccer statistics databases that chronice red cards? More specifically I seek information on the red cards given to John Arne Riise." 21 Dec under Entertainment (no resulting edit to the article)

-"There are 400GB MicroSD cards for US$45 (US$0.11/GB) but I'm hard-pressed to get more than 32GB in my desktop, and it cost of about US$5/GB. Now, I realize that the memory in a desktop is much faster than that in a MicroSD card, but why is there such a large disparity in the technology?" 25 Dec on Computing (no resulting edits to the SD card article)

To be clear, I don't mind the ref desks; as I stated in my original comment, the encyclopedic purpose I see and that is stated in the ref desk guidelines is that it allows editors to get a sense of issues readers have with our article coverage. I think that's actually a very compelling reason. Whether or not that happens is a different question, but that the possibility is there is enough for me. But If readers, many of whom are unregistered, cannot do so, that benefit is severely diminished and it turns into a registration required forum that just seems like a worse StackExchange. Wugapodes [t^hɑk] [ˈkan.ˌʧɹɪbz] 06:05, 28 December 2018 (UTC)

• Support per FPS ―Abelmoschus Esculentus ^{talk / contribs} 04:11, 28 December 2018 (UTC)
• Support per FPS. Enterprisey (talk!) 05:09, 28 December 2018 (UTC)
• Strong oppose - the ref desks are the worst pages in Wikipedia to require protection on, since they are so needed for help. It's not that I don't get the arguments being used, but there must be alternate steps that can be taken - SP risks the worst of both worlds. Nosebagbear (talk) 15:05, 28 December 2018 (UTC)
  • Hi User:Nosebagbear. You did see this, right? How many IPs are getting help there? Weigh that against the troll feeding and huge waste and....well, I just hope you read what Fut.Perf. wrote above and change your mind. Anna Frodesiak (talk) 20:25, 28 December 2018 (UTC)

  @Anna Frodesiak: - I am well aware of the deluge. I am also aware that the string of temp semi-protects aren't helping the IPs either. However, I'm just so reticent to accept it without any way to help those IPs and the (here) large group of non-autoconfirmed newly registered. I wouldn't say I'm qualified to judge whether the proposal offered below can mitigate on that count, so I'm interested in what others are saying. Nosebagbear (talk) 20:34, 28 December 2018 (UTC)

• Support: AGF is not a suicide pact. --K.e.coffman (talk) 19:44, 30 December 2018 (UTC)
• Strong oppose: I'm an IP who sometimes asks and answers questions there and frankly, I think there's more than two or three of us. If you check the history of ref desks, you'll see that the days when they're protected, the number of questions goes down to almost nothing. By doing this you're creating a barrier to entry and destroying the informal nature of the ref desks, and driving their users away. I'm kind of insulted by this talk of "having myself to blame". If I'm not gonna create an account to edit articles, I'm damn well not gonna create one to ask questions either. 78.0.232.155 (talk) 22:12, 30 December 2018 (UTC)
• Oppose: If the Refdesks are semiprotected indefinitely,the troll wins. — Preceding unsigned comment added by Gevaltio (talk • contribs) 09:13, 31 December 2018 (UTC) — Gevaltio (talk • contribs) has made few or no other edits outside this topic.
• Strong Oppose Per Wugapodes's arguments (particularly the loss of purpose if it is indef protected). Like him, I am down with office actions being taken as needed though. Zell Faze (talk) 19:17, 31 December 2018 (UTC)
• Oppose. Too great a barrier to entry just to ask a question. I don't really understand why, but a great many interested and thoughtful Wikipedia users will never take that step, easy as it is. --Trovatore (talk) 22:24, 31 December 2018 (UTC)
• Oppose. I still believe (AGF) that the amount of quality posts coming from IPs outweighs the bad ones and to lose those would not be a net benefit, we would lose out on so many potentially valuable posts from IPs. And the reference desks have been an integral part of enwiki for many years, "shutting them down" seems a drastic and shameful loss. -- œ^™ 06:17, 6 January 2019 (UTC)
  • You believe the good IP posts outweigh the bad ones? Well, sadly, your belief is mistaken. Have you counted? My estimate, after following the Hum and Lang desks for years, is that about one or two out of a hundred anon/non-confirmed posts are from genuine newcomers. Fut.Perf. ☼ 08:08, 6 January 2019 (UTC)
• Oppose per OlEnglish. The refdesk is a net benefit to the project. Balkywrest (talk) 06:26, 6 January 2019 (UTC)
  • Can you clarify please: is this an oppose to semiprotecting the desks, or an oppose to shutting them down, or both? Fut.Perf. ☼ 11:41, 6 January 2019 (UTC)

Discussion

• Query on Pending - Pending is obviously really tough to use on active pages, like these. But just how many posts do we get on one in an average day by non-EC individuals? If it is at all viable, then permanent Pending might be the way to go Nosebagbear (talk) 15:05, 28 December 2018 (UTC)
• Alternative proposal: Assuming that there is a consensus to semiprotect the refdesks, how about making subpages that are open to IP editors but are under pending changes protection, along with a strict rule that no answers are allowed on the subpages and that those who have the pending changes reviewer right are allowed to move the questions to the associated refdesk. Note: if you are not familiar with PC protection, please read Wikipedia:Pending changes before commenting on this proposal. --Guy Macon (talk) 17:03, 28 December 2018 (UTC)
  • For the current pattern of vandalism, this would not help at all – even on a PC subpage, the vandal edits would still all have to be reverted and revdeleted, so the amount of potential disruption and the workload required to deal with it would be exactly the same as before. (Same goes, of course, for the idea of PC-protecting the main boards themselves.) Fut.Perf. ☼ 17:17, 28 December 2018 (UTC)
    • If the creation of a page inevitably leads to the refdesk vandal using that page, why are the other 61,620,449 pages not filled with his vandalism? I fail to see how vandalizing a page where his vandalism would be invisible to the vast majority of Wikipedia readers would be attractive to a troll. Even if he did vandalize the subpage the only editors who would see it are the relatively small number who [A] have the pending changes reviewer right, and [B] watchlist the subpage. An admin could come by every few months and do a revdel so even PC reviewers could no longer see the vandalism in the history. --Guy Macon (talk) 18:25, 28 December 2018 (UTC)
      • All of the above would apply to PC protecting the refdesk itself, but doing that has a huge flaw: When editors who are not reviewers post to a PC protected refdesk that contains unreviewed pending changes, their edits are also marked as pending and are not visible to most readers. The subpage idea solves that problem. --Guy Macon (talk) 18:30, 28 December 2018 (UTC)
      • Correction: I just logged out cleared my browser history and looked at a page with PC protection. When reading the page I saw the latest accepted revision, but I could see the rejected pending changes in the history, so it looks like revdel would still be needed (but would be far easier to do on a subpage where nobody replies to a post). --Guy Macon (talk) 19:01, 28 December 2018 (UTC)

Proposal II: Shut down the Ref Desks

• Shut Them down and let them move to Wikiversity (first choice) or Protect indefinitely (second choice) per Iridescent's compelling argument below. --Guy Macon (talk) 22:28, 5 January 2019 (UTC)
• Close the refdesk Wikipedia is an encyclopedia, not a discussion forum or a reference desk. It has lived past its usefulness, as further evidenced by this situation. --Pudeo (talk) 23:51, 27 December 2018 (UTC)
• Shut them down - Bite the damn bullet and get rid of the friggin' ref desks, already Spin them off into another project, maybe, but they have little if anything to do with building an encyclopedia. Deep six 'em. Beyond My Ken (talk) 13:58, 5 January 2019 (UTC)
• Shut them down. The reference desks have almost no relevance to Wikipedia, haven't been used in the way they're meant to be used for well over a decade, and are a fertile breeding ground for problems that constantly spill over into the rest of the project. If the regulars really feel the desks are so damn important we can't do without them, take them over to Wikiversity where their mix of original research and vague speculation at least vaguely fits into the culture. There comes a point when the half-dozen people to whom it's a pet project can't keep expecting the rest of us to clean up their trash, and we passed that point long ago. ‑ Iridescent 14:52, 5 January 2019 (UTC)
• Oppose shutdown. While I can understand this reaction, I'd still consider this a regrettable outcome. The desks are potentially useful, even if on a rather modest scale. And the current issues aren't some home-made trash produced by the regulars (as some situations in the past were), but really just the problem of keeping a small number of vandals out. For that purpose, semi-protection works. In fact, as you may have noticed, somebody in the meantime went and did just that (without causing any outcry, surprisingly): semiprotected several of the desks for a few weeks. And see what happens: the desks are working fine now. Sensible questions coming in, reasonable answers being given. (I'm talking mainly about the Humanities and Language desks; I don't usually follow the others.) Fut.Perf. ☼ 17:25, 5 January 2019 (UTC)
• Shut them down - They are causing too many problems and are not improving the encyclopedia. I am not sure if any other foreign language Wikipedia has a reference desk but they are not helpful to the encyclopedia. Pkbwcgs (talk) 15:16, 5 January 2019 (UTC)
• Shut down per WP:NOT or, at least, semiprotect indefinitely. buidhe (formerly Catrìona) 19:32, 5 January 2019 (UTC)
• Shut them down and since WP:ABUSE has no teeth since ISPs will laugh off any reports from people without a @wikimedia email address, support WMF legal actions against the ISP, with detailed diffs of every misuse, with a preference for those which would violate criminal statutes against harassment, threatening statements, etc. (WP:OFFICE really has no use here other than to revdel/oversight/OFFICElock the RefDesks; thus asking for office actions to be taken on Wikipedia proper wouldn't particularly help. This is a fight where WMF's legal counsel would have to get involved to send out the C&Ds, and that isn't going to happen on Wikipedia itself.) —A little blue Bori v^_^v ^Bori! 22:42, 5 January 2019 (UTC)
• Shut them down since that appears to be on the table, again. Same arguments as last time, but they can be summarized as "outside our mission" and "cost exceeds benefit". ―Mandruss ☎ 22:52, 5 January 2019 (UTC)
• Shut them down I have long been of the opinion that the refdesks aren't really part of Wikipedia at all. They don't operate under the same rules and also don't operate like a proper reference desk, there a librarian will provide, you know, references, as opposed to just making something up. They do nothing to improve the encyclopedia and have long been a source of needless drama. If they were a user we would've banned that user years ago as a "net negative." Beeblebrox (talk) 04:07, 6 January 2019 (UTC)
• Shut them down Not only are they not relevant to our purpose, but apparently they have become a net negative to the project. In a time of declining editorship we need to use our resources more wisely. --Rschen7754 04:17, 6 January 2019 (UTC)
• Oppose and call for office actions -- the refdesks are a valuable resource, and the best way to deal with the sort of gross abuse which is the problem here is to start going after the trolls, so they would face real consequences (such as getting kicked off of Wikipedia) for their actions. 2601:646:8A00:A0B3:74C8:E1A4:A5D:7436 (talk) 07:43, 6 January 2019 (UTC)

• We have kicked the trolls off of Wikipedia. A few thousand times. They immediately come back through a new proxy that we never heard of before. I happen to be more technically sophisticated that most (but not all) trolls, but if I was a vandal or a troll I could, for the princely sum of $25, buy compete control of 1,000 computers with 1,000 IP addresses on a botnet, and use them to abuse Wikipedia until all 1,000 were blocked. $110 gets me 5,000 IPs and $200 gets me 10,000 IPs. --Guy Macon (talk) 08:01, 6 January 2019 (UTC)

• Strong Oppose - A human alternative when automated searches prove to be unsuccessful is useful to both editors and readers. As long as volunteers choose to allocate their time to that portion of the site, so be it. — Godsy ^(TALK_CONT) 07:53, 6 January 2019 (UTC)
• Oppose - I started a short discussion last Nov. about alternatives to repeated blocks or a permanent block, see Wikipedia:Village pump (policy)/Archive 147#Ref. desk protection. I can't speak for all the desks, but for one I work on regularly most of the question come from people who aren't regular editors, so blocks keep the main source of 'customers' away. To me the root cause here is that WP has not kept up with the technology available to vandals; when a bot can vandalize a page every minute, each time under a new IP or dummy account, it's not practical to try to counteract it manually. The Ref. Desks seem to attract a lot of attention from vandals, but as long as they have the technology there's no reason they can't move on to other pages, so a permanent block or removal would result in the problem being moved, not solved. I interact with several other wikis besides WP and I've noticed that they don't seem to have the same issues with vandalism, and I think that's mainly because they implement measures that WP isn't willing to. These measures seem innocuous to me: requiring a registration to edit, or using reverse Turing tests to weed out bots, and perhaps there were good reasons for not implementing them in the past, but the world has moved on and perhaps it's time to re-examine these reasons. To those who say the Ref. Desks are not encyclopedic, I think WP has a wider mission to disseminate reliable information, and the Ref. Desks fit in with that mission. Sometimes you don't know in which article to look for the information you need, or you don't know the relevant term enter into Google, and some kind of human guidance is needed. The proposal seems like cutting off your nose to get rid of a pimple, it does solve the problem but it's hardly the right answer. --RDBury (talk) 09:55, 6 January 2019 (UTC)
• Shut down the reference desk. It is not within the scope of our mission of writing an encyclopedia. Those who enjoy answering questions can migrate to Quora or fork the RD off into an offwiki project, where it does not produce maintenance overhead for Wikipedia. Sandstein 10:20, 6 January 2019 (UTC)

Reduce number of vandalism warning levels

It has long been my view that four warning levels is rather excessive. People really only should need to be told once to stop. The only caveat is whether they actually read the message, but those who clearly don't are also those for whom extra warnings are useless.

On nearly every other Wikimedia project I edit, usually only one or two warnings are given at the most. I think two levels would be a good compromise, retaining the softly-worded first level (in case of good-faith mistakes) and the third level while eliminating the second and fourth levels. Those who continue vandalizing after they know it's blockable don't need another warning (level 4), and level 2 warnings could easily be merged into level 1. This would decrease patrollers' workload while also decreasing the total amount of vandalism.--Jasper Deng (talk) 03:56, 28 December 2018 (UTC)

• Support too many levels. I don't warn when a level 4 is needed I report. Vandals know they are doing vandalism. Legacypac (talk) 04:04, 28 December 2018 (UTC)

• @Jasper Deng: Not sure if I'll end up actually participating in this as far as support/oppose goes, but are you proposing this for only the uw-vandalism series, for selected user talk warnings, or for all user talk warnings?--SkyGazer 512 ^{Oh no, what did I do this time?} 04:07, 28 December 2018 (UTC)

  I guess it's {{uw-vandalism}} only since that is only warning template series for vandalism ―Abelmoschus Esculentus ^{talk / contribs} 04:08, 28 December 2018 (UTC)

  For now, I'm just proposing it for the vandalism templates, but I do want to bring it up as a general point of discussion. Other templates for which we should be reducing the number of warning levels are those related to spamming (since spam often falls in the same realm as vandalism), introducing deliberate factual errors, creating inappropriate pages, and maybe even BLP violations.--Jasper Deng (talk) 04:18, 28 December 2018 (UTC)

  Sounds reasonable. I personally would probably support defining the other templates besides vandalism whose number of warning levels should be reduced as ones designed for any user whose is very likely to be acting in bad faith; imo, adding original research and expanding plot summaries unnecessarily should be treated very differently than blatant vandalism as far as warnings go.--SkyGazer 512 ^{Oh no, what did I do this time?} 04:51, 28 December 2018 (UTC)

• Support I don't think tolerating vandalism is a good idea. ―Abelmoschus Esculentus ^{talk / contribs} 04:09, 28 December 2018 (UTC)
• Is there any way to gather statistics on the effectiveness of the warnings? How many people get two warnings and then become productive Wikipedians? How many new editors change their ways after three or four warnings? Jack N. Stock (talk) 05:05, 28 December 2018 (UTC)
• Support provisionally, although like Jack just above I think it's better to make this sort of decision with more data. Enterprisey (talk!) 05:08, 28 December 2018 (UTC)
• Oppose. There's no requirement to go through all four levels. For egregious vandalism, I'll start with level 2, or I'll escalate 1 to 3. I also have no qualms blocking for clear vandalism after a level-3 warning has been given, since it mentions the possibility of a block. —C.Fred (talk) 05:09, 28 December 2018 (UTC)
  • I myself always start with at least level 2 unless I suspect it to be a genuine good-faith mistake. Long ago, I've had AIV requests declined after I did not go through all four levels. In the vast majority of vandalism cases, we really only need a single warning, not even two. So there is virtually no legitimate use case for all four levels.--Jasper Deng (talk) 05:36, 28 December 2018 (UTC)
• Two warnings - Mild and Serious - should be enough. Legacypac (talk) 05:58, 28 December 2018 (UTC)
• Please create a sandbox showing the proposed new warnings (could be fairly rough but enough to see likely wording). Johnuniq (talk) 06:23, 28 December 2018 (UTC)
  • @Johnuniq: I would keep uw-vandalism3 as-is, remove level 4 entirely, and pretty much keep level 1 with the modification that "did not appear to be constructive" is wikilinked to WP:VAND. This way, we can introduce the notion of vandalism without scaring the user by explicitly labelling them as such.--Jasper Deng (talk) 00:20, 29 December 2018 (UTC)
• Comment On the one hand, there are edge cases, where I might have to see several edits from a user before I'm sure they are vandalizing, so a series of warnings (of one type or another) is called for by AGF. On the other hand, I have blocked a user after as few as one warning when the user was rapidly vandalizing many articles. I no longer monitor AIV, because after carefully reviewing each request, I almost never did anything because the case did not require blocking (ie, not vandalism, stale, etc.). The few times I thought blocking was called for, another admin usually blocked while I was reviewing the report. - Donald Albury 14:36, 28 December 2018 (UTC)
• Unneeded - Like Jasper Doeng, I start with level 2 for vandalism. There's no need for a lvl 1, since a good faith vandalism mistake doesn't really exist - I tag them as lvl 1 disruptive instead. I wouldn't mind removing the bottom one, but I suspect people would then want to rephrase them and I think that would be counterproductive. I would definitely NOT want a reduction for the others warning templates. Nosebagbear (talk) 15:01, 28 December 2018 (UTC)
• Support - While it may be unneeded, the extra warnings are also unneeded. While some admins see the 3rd warning as enough warning, I imagine that not all do, and may not block until after the final warning. I also support the single warning idea, however, leave the 1st warning for WP:AGF. - ZLEA _Talk\^Contribs 16:47, 28 December 2018 (UTC)
• Unneeded As has already been said, there is no requirement to progress 1, 2, 3, 4 through the existing templates. Just give the appropriate level warning according to the situation, or no warning at all if that is what is warranted. GMG^talk 20:30, 28 December 2018 (UTC)
• Oppose we block as needed and when appropriate warning has been given or when no warning has been given at all if bad enough. This change would have no impact on admins. The advantage of the 4 stages of warning is mainly that it prevents people playing Huggle the Video Game® from reporting good faith IPs/school children and requesting indefinite blocks for stuff that isn't vandalism/doesn't need a block. Changing the warning levels would dramatically increase the bad reports at AIV and waste admin time in areas where it isn't needed. TonyBallioni (talk) 20:33, 28 December 2018 (UTC)
• Support Those suggesting that there is no need use all 4 levels in turn are correct. They are also fundamentally wrong about the lack of need to reduce the number of levels. As usual, they look at things narrowly, through their own myopic lens. The vast majority of non-admin vandal fighters will not come here. They will plod on with the 1-2-3-4 warnings in fear of ANI if an over-vigilant Admin. reports them or refuses to enact a block because 1-2-3-4 has not been followed. There should be an immediate reduction from 4 to 3. In 6 months, 3 to 2. Leaky Caldron 20:40, 28 December 2018 (UTC)

• Loving the personal attacks. Do you want 1, 2, 3 or 4 warnings for that ? Nick (talk) 20:43, 28 December 2018 (UTC)

• Oppose there is no requirement to go 1-4. Use common sense. But graduated warnings have a purpose, just not for obvious trolling and blatant vandalism. Praxidicae (talk) 20:43, 28 December 2018 (UTC)
• Oppose per Praxidicae. If you don't feel a certain warning level is appropriate, use the one you feel fits the situation. DonIago (talk) 20:53, 28 December 2018 (UTC)
• Can any opposer give a single case of nonaccidental vandalism where all four levels’ use has been useful? Because if we are always skipping levels, then there’s a redundancy, and it has always been the case in my experience that I am skipping level 1 and giving at most two warnings in total. Also, Huggle and ClueBot NG automatically escalate by just one level at a time in many situations where a quicker one is warranted. If those who use the tool are making frivolous reports after going through all four, then that quite frankly is abuse of the tool and does not address the broader problem. @Nosebagbear: my surname is not “Dong”.—Jasper Deng (talk) 22:02, 28 December 2018 (UTC)
  • Jasper Deng, the point is that what Leaky caldron is saying as a negative is actually a benefit: it really cuts down on bad AIV reports from people who are new and decided on their 3rd day on the project that they were going to dive into anti-vandal work without realizing how Wikipedia works. If you're an experienced user, use your judgement as to which warnings to give. Having the graduated warning system is very useful for new users who want to get involved with anti-vandal work, however. Basically the warnings exist equally to train them as much as warn vandals. TonyBallioni (talk) 22:12, 28 December 2018 (UTC)
    • I contend that the amount of vandalism eliminated will result in a net decrease in the amount of work. It might result in more work at AIV itself, but result in a net reduction of reverts, making vandalism-fighting less high-speed and less tempting to rush. Also, I seriously doubt it will result in a decrease in the amount of frivolous reports. Those who make them should, if anything, be educated earlier on NOTVAND rather than later, and we get a net reduction of frivolous warnings and consequently less BITE'y behavior overall. --Jasper Deng (talk) 00:27, 29 December 2018 (UTC)
• Oppose. Like others have stated, there is no requirement to go through all of the warning levels. But the warning levels do have their purpose, including for reports at WP:AIV. Some newbies just don't get the point and need at least two warnings. I don't see that four warnings are needed unless the editor missed the talk page messages somehow, but anyway. Flyer22 Reborn (talk) 22:16, 28 December 2018 (UTC)
• oppose unnecessary. If I only want 2 levels warning for serious people, I will start level 3 directly. If someone is a newcomer, we should remain level 1 per WP:BITE Hhkohh (talk) 22:17, 28 December 2018 (UTC)
  • @Hhkohh: I did not advocate removing level 1. See the second paragraph of my proposal.--Jasper Deng (talk) 00:27, 29 December 2018 (UTC)
    • Seen and known but I sometimes give them level either 1 or 2 depending on their edits Hhkohh (talk) 01:14, 30 December 2018 (UTC)
• Neutral per Johnuniq. Without a sandbox version of these new templates, I have no idea what I'm !voting on here. I completely sympathize with the idea; every time someone gives the goatse LTA the obligatory four warnings before reporting to AIV, I want to come out swinging with a clue-bat, but I can't imagine how these would be phrased so as not to be overly WP:BITEy when it comes to kids making test edits, or when the RC patroller made a mistake and flagged a good edit, etc. I'm open to ideas, though. Suffusion of Yellow (talk) 22:20, 28 December 2018 (UTC)

I've also had a request or two turned down at AIV. But if the editor is clearly vandalizing or clearly otherwise editing disruptively and I have warned the editor at least twice (via template warnings or non-template warnings), that is usually enough to get the editor blocked. Some admins will make a WP:NOTHERE block. Flyer22 Reborn (talk) 22:27, 28 December 2018 (UTC)

• Oppose - This adds unnecessary bureaucracy to the process and regarding an issue that hasn't shown a need for any kind of change. The number of warning levels exist in order to make the process easy for patrollers, accommodate for all situations that occur, and (most importantly) assume good faith by default with bots, editors, and automated tools. There's currently no requirement that users be warned with all four warning levels and have made a fifth disruptive edit before they're blocked. As stated above multiple times by other users: I warn users very differently depending on the severity of the vandalism or disruption, their past warnings, their block log, how quickly they're attempting to push disruptive edits, what pages they're making bad-faith disruptive edits to, if sock puppetry or abuse of multiple accounts or editing while logged out is suspected, and many other factors. I use this information to either start the user at a level 2 or 3 warning compared to a level 1 warning, start the user at a level 4im warning, block the user only after they've been given a level 3 warning as opposed to waiting until after they've been given a level 4 warning, or use a different method to warn the user alltogether. The warning level system, which warnings are left first, or when action is taken earlier as opposed to waiting until the user has been warned four times and has made a fifth disruptive edit - is widely practiced and used on a judgment-level basis, and administrators who regularly handle this disruption already know to review reports at AIV, ANI, or other places thoroughly, verify and determine if the edits constitute vandalism or bad-faith disruption or not, determine if the user was sufficiently warned given the disruptive edits made, and do this before they consider action. ~Oshwah~^{(talk) (contribs)} 22:34, 28 December 2018 (UTC)
• On the fence I usually start off at uw-2 and work up through uw-4 before blocking someone (unless it's pretty obvious they know what they're doing). This is not to say that uw-1 messages are useless: I use uw-test1 almost consistently in place of uw-vandalism1 when I feel that uw-vandalism2 might not yet be appropriate. There is no uw-test4 because, well, they're obviously engaging in vandalism -- that suggests uw-test and uw-vandalism need to be merged. That said, I manually post warnings just to be sure I'm using the right one. I've encountered plenty of middle-experience users using Twinkle who issue generic vandalism warnings for any revertable actions (instead of uw-delete, uw-npov, uw-notcensored, uw-agf, uw-npa, uw-fringe, uw-nor, uw-spam...), ignoring WP:NOTVAND to the point of failing WP:AGF. The reported user (who wanted to be productive) then gets it in their head that either:
  - anything that doesn't fit the mold is "vandalism" and sees no reason to try to cooperate
  - any idiot can leave whatever message they want (no matter how wrong) and so there's no point in paying attention to messages
  - because the reporting user was wrong to use the vandalism warning, the reported user's edits were somehow in the right
  Meanwhile, the reporting user gets frustrated because I have to stop and properly explain to someone who sincerely believes they're trying to help how their behavior wasn't helpful. Now, the counter example of LTAs is brought up, and for fuck's sake if it's an LTA just report them on sight. If someone is leaving automated warnings for an LTA to enable a report button, they should have their tools taken away so they can learn how warning and reporting users actually works.
  TL;DR: Three levels (maybe a mistake, warning, final warning) seem good for users who are manually warning and reporting but I feel like any automated warning tools should stick to four levels. Take away tools from people who use them to the exclusion of common sense. Ian.thomson (talk) 22:42, 28 December 2018 (UTC)

Ian.thomson, this is the best description of the issues surrounding AIV I have ever read. Also, yes, if its an LTA, we block on sight. TonyBallioni (talk) 23:33, 28 December 2018 (UTC)

I completely agree with Ian.thomson regarding how users tend to default their patrol actions to leaving a vandalism warning template compared to a more in-depth and explanatory warning (or God forbid, a custom warning... lol) for the actual problem - such as NPOV issues, unexplained content removal, adding content without a reference, test edits, or other actions that can be done and entirely in good faith. I consistently see users who patrol recent changes and do this, and it bothers me a lot when I see that. If anything, I think that certain templates shouldn't be automatically grouped into just warning the user for "vandalism" if a level 3 or 4 warning is left. I'll also add that automated tools should separate previous warnings left and know if it was for vandalism and blatant disruption vs the others, and not automatically stack and leave the next warning level template regardless of what the user was previously talked to about - especially if this or a similar proposal in the future should ever pass. Example: A user is warned twice (level 1 and 2) for content removal without an explanation, but a day later an editor leaves a warning for vandalism - in this situation, since the user wasn't warned for vandalism before, they shouldn't receive a level 3 warning for vandalism. ~Oshwah~^{(talk) (contribs)} 00:03, 29 December 2018 (UTC)

• Meh I tend to think a better idea would be to make it clear that you don't have to work through all four stages or even start at the "general note" stage when it comes to blatant and obvious vandalism like writing "poop" or something similar in an article. I, and I belive many others, choose the template they belive is right for the situation, which sometimes means going straight to level 3 or 4, but in less clear cases starting lower. Beeblebrox (talk) 23:00, 28 December 2018 (UTC)
• Support. Vandals know what they're doing. Having four levels makes editors feel they have to clunk through them in case they're later told they've jumped the gun. SarahSV ^(talk) 23:03, 28 December 2018 (UTC)

• Opposition This... can already been achieved. Yes, 1-4 are there and are meant to be used, however... there is warning 4im. If 1-4 doesn’t cut it, then you can use a 4im. In the rare event you need to give an instant only one warning to a vandal, just give a 4im. 4im exists, and “one warning only” is EXACTLY what 4im is. Essentially, you’re reinventing the wheel. This system has more flexibility with the ability for 1-4 and also in extreme cases skipping those. This’ll just add less flexibility and versatility to the system, you won’t put a littering person in jail for life, would you? Best regards, Redactyll Letsa taco 'bou it, son! 23:07, 28 December 2018 (UTC) 23:07, 28 December 2018 (UTC)
  • @Redactyll: This proposal is emphatically not about my preferences for giving warnings but about setting a project-wide precedent; I propose it because I see others wasting time and resources with giving all four warnings to blatant vandals. I might give that littering person a warning the first time if it were by accident (hence my proposal to keep two levels), but I would not hesitate to fine them if egregious. Also, since blocks are not punitive, this is not a proper analogy.--Jasper Deng (talk) 00:21, 29 December 2018 (UTC)
• Support CLCStudent (talk) 23:11, 28 December 2018 (UTC)
• Comment: The German Wikipedia, second largest after the English one, has only two warning levels for all cases, and a "-im" version if level 2 is applied directly. See de:Wikipedia:Huggle/Vorlagen. I personally never liked these templates; they instantly confront the editor with a big red stop icon and tell the user that their edits have been "unhelpful". However, it does seem to work there. Just for your consideration.
  ^{Disclosure: I have been invited to this discussion by a notice on my talk page, likely based on my Huggle activity.} ~ ToBeFree (talk) 23:14, 28 December 2018 (UTC)
• Oppose 2, Support 3; I use uw-vandal1 for cases of people trying to do something that they might think is right, clearly isn't, and could be construed as vandalism. I use uw-vandal2 as a first warning for obvious cases of vandalism. I think 3 and 4 could be merged. At play there is an issue that if it is rapid fire a person may commit more vandalism without having seen an incoming warning and then end up being blocked before they've had a chance to read a warning. 2 warnings alone might not allow for this. --Hammersoft (talk) 23:42, 28 December 2018 (UTC)
• Oppose - I've never interpreted the four warning levels as implying that all four need to be used in succession. In fact, I probably never have used all four in that way. I prefer level 1 for newbie edits likely made in good faith, as the text of the warning bears that out. If it appears deliberately-disruptive but is relatively benign, starting out with level 2 is perfectly reasonable and assumes no intent. With severely-damaging edits, there's no reason why one couldn't go straight from 2 to 4, or if continued disruption after level 1 belies bad faith then level 3 can be next.

I feel that the distinction between "probably good faith" and "probably bad faith" should be maintained; the "Hello, I'm Sable232" and direction to the Help Desk are good for the former, but sound laughably patronizing (or even condescending) to the latter and may prompt further disruption. --Sable232 (talk) 23:56, 28 December 2018 (UTC)

• Strong oppose: no-one is saying you have to use all four in order, as if the vandal is levelling up at each step and once they've surpassed level 4 it's time for the boss battle. The last thing we need is more BITE and merging level 2 and level 1 would do exactly that. Plenty of edits that I revert are unhelpful but done out of ignorance rather than malice, and I need a pre-written message to slap on their talk pages to explain why I undid it, but something kindly written enough to not scare them away. — Bilorv_(c)(talk) 00:26, 29 December 2018 (UTC)
• Weak Oppose I am a big fan of the three strikes rule and think a solid argument could be made for dropping the level 4 warning. Once you have been told to STOP a given pattern of behavior three times, a 4th warning should not be needed. However, dropping things down to two warnings is a bridge too far for me. And some of the comments above make good points by noting that multiple warnings helps keep overenthusiastic newbies from reporting people too quickly. Also correctly noted, there is no requirement to go through all 4 levels of warning. If the behavior in question is obviously bad faith start with a level 2 or 3 warning. There is actually no firm requirement to issue warnings at all when dealing with egregious behavior. I have issued more than a few no warning blocks. It all depends on the exigent circumstances. Editors should adjust their response to the nature of the problem. How serious is the disruption? What is the chance that it is good faith mistakes (and so on)? Dealing with disruptive behavior requires a lot of situational judgement. -Ad Orientem (talk) 00:46, 29 December 2018 (UTC)
• Oppose It's easy enough to issue a level 2, 3, or 4 warning "off the rip"; much harder to allow for good faith errors if the 2nd warning automatically triggers an AIV report. If anything, make sure it is abundantly clear in the instructions that it is acceptable to increase the automated level based on the patrolling editor's interpretation of severity and the clarity of a nefarious intent.--John Cline (talk) 00:50, 29 December 2018 (UTC)
• Oppose A better solution would be for people to stop defaulting to level 1 warnings for malicious vandalism. Choose whatever warning level is appropriate based on the edit. I recommend setting your twinkle default to level 2 if you are lazy like I am. Natureium (talk) 04:14, 29 December 2018 (UTC)
• Oppose Current system is a flexible tool to cater the appropriateness of warnings and levels as one see fits. I am an active vandal fighter and one of the handful counter vandalism program/course graduates. My vote based on my knowledge and experience as below.

1. New IP editors (most of the vandals) - WP:Assume good faith. Warning level start at 1 and increasing accordingly to 4 before reporting to AIV.
2. Choose the warning level appropriately to justify the nature of the edit or the behaviors of the vandalized editors. Any level warning can serve as the starting level and the warning level could skipped (such as from level 2 to level 4 for example) for clear malicious vandalized edits.
3. Level 4 or (im) warning. It could be given to a vandalized editor who performs egregious vandalized edits across multi-pages in a quick succession or inserts shocking inappropriate image(s).
4. an (im) warning. It could be given to the vandalized editor and report them to AIV at once if any abhorrent personal attack(s) such threat of legal action or threats of violence against an editor on their talk/user page.
5. If any of the template wording is not suitable or additional details need to be included, one CAN always write a personalized warning messages and give warning level on the vandalized editor talk page without using the warning templates. Thank you. CASSIOPEIA^(talk) 06:27, 29 December 2018 (UTC)

• Oppose If you don't think all 4 warnings are necessary, just skip to level 4 or 4im. SemiHypercube 14:50, 29 December 2018 (UTC)
• Comment: Perhaps an alternative is to keep the different warnings for different levels of disruption, but remove the presumption that there must/should be some sort of escalation through them before action must be taken (and ensure that those who act - especially at AIV and the like, understand this.Nigel Ish (talk) 15:24, 29 December 2018 (UTC)
• Oppose In my admittedly limited experience most vandals stop before I have to file an AIV report, but sometimes it takes multiple warnings to get their attention. Level 4im is available for editors who feel that a single report should be enough. Besides {{uw-vandalism}} I often use {{uw-delete}}, {{uw-unsourced}}, etc, which likewise have four levels. Strawberry4Ever (talk) 16:06, 29 December 2018 (UTC)
• Meh AFAIK, there's no requirement to escalate in the order 1, 2, 3, 4, report. I tend to use 4im or start on 3 if it's obvious vandalism. Probably best to keep it as it is but definitely support a measure to avoid instruction creep requiring anti-vandals to escalate in the aforementioned order. Leave it to the discretion of users. SITH (talk) 16:55, 29 December 2018 (UTC)

This thread doesn't seem to have any relationship with Middle Eastern military history task force... ―Abelmoschus Esculentus ^{talk / contribs} 08:02, 30 December 2018 (UTC)

• Oppose I'm probably one of the guilty ones who go through all the levels when probably not required, but I certainly don't see it as a requirement in the most egregious cases, and I've seen lots of IP edits who stop after the 2nd warning. I have seen plenty of individuals who get reported to AIV before the final warning if they are particularly egregious, and there are also no restrictions in placing a higher level warning, as outlined above. Agent00x (talk) 23:18, 29 December 2018 (UTC)
• Support. If one really does use all current levels, that amounts to five specific, distinct incidences of vandalism before a vandal is blocked, one for each level of warning plus the vandalism-after-final-warning that leads to a report. That's too many. So some say skip levels. Then what's the point of having so many levels? I've long thought we should reduce the number of warning levels, and I'm glad to see I'm not alone. As for how many levels, I'm fine with two. That way the number of incidents of vandalism before a block is three. And maybe it's just the baseball fan in me, but "three strikes you're out" is a fine paradigm. oknazevad (talk) 06:24, 30 December 2018 (UTC)
• Question for the supporters: My personal preference is to not use the template at Template:Uw-vandalism1 for obvious vandalism like "H1TLER WUZ R1GHT!" or "qwertyuiopasdfghjklzxcvbnm!" because I don't think that vandals should be sent to the help desk. I also don't like the redundant "I am X" (I already have a signature at the end) or the insincere "hello" (I am warning them, not greeting them). Instead I use Template:Uw-vandalism2 as the first warning. And no, I have never been criticized for that choice. Other editors clearly like to use Uw-vandalism1 as the first warning. So are you proposing that I cannot use the template that I prefer or are you proposing that those other editors cannot use the template that they prefer? --Guy Macon (talk) 09:16, 30 December 2018 (UTC)
  • (...Sound of Crickets...) --Guy Macon (talk) 04:23, 3 January 2019 (UTC)
• Oppose. I don't think I've ever used all four levels, but I think flexibility is beneficial. 331dot (talk) 09:29, 30 December 2018 (UTC)
• ’’’Meh’’’. I’m with those above who note that there is no requirement to follow the 1234 increasing level of naughtiness warnings. Roxy, the dog. wooF 16:39, 30 December 2018 (UTC)
• Support. As someone who handles problematic edits by new users, I believe that only two warnings are necessary. For those who are opposing this based upon flexibility reasons, I totally understand your point of view and it wouldn't be that bad if things were to stay as they are. I frequently accelerate the warnings that I give to users who make severe problematic edits so that they receive level 2 warnings to begin with rather than level 1 or even that they receive a level 3 warning, if they make a significantly problematic edit after receiving a level 1 warning. So why do I believe that only two warnings are necessary? Because in the multiple years that I've been handling problematic edits from users, I've never seen anyone redeem themselves on that specific account. I'm guessing if someone matures from making this kind of edit, it is likely to be 5 - 10 years from when they made the account that has problematic edits attached to it. So what happens then? I believe they make a new account and go on to be productive members of the community from their first edit without the stain of the problematic edits made in their youth being associated to that account. Often I've noticed that new users who are here to contribute positively to the Wikipedia make good edits right from the start or as an IP before they make an account. Assuming that most, if not all new accounts will make problematic edits within their first edits is a flawed assumption to make because it just doesn't seem to happen that way. Sure, there are users who make mistakes that aren't vandalism related, like not being aware of NPOV or making formatting errors but on the whole, there are little to none productive users who produce edits that seem like vandalism within their first few edits to Wikipedia. Practically every account that within their first few edits make problematic edits that look like vandalism are here to vandalise. So on that basis, clear vandalism should be given 1 - 2 warnings before that user is blocked from disrupting Wikipedia any further. Thank you. -=Troop=- (talk) 17:37, 30 December 2018 (UTC)
• Oppose per AGF. Four warnings has always been the norm. Skipping levels is occasionally appropriate for particularly egregious vandalism, and then it's up to the discretion of the responding admin whether to block. Reducing the number of warnings required would have no effect other than to increase the number of inappropriate blocks. Bradv🍁 17:42, 30 December 2018 (UTC)
• Oppose. However, it should be made clear that it is not required for anti-vandalism fighters to start at level 1 and work their way up. We also more test-or-vandalism templates at level 2-4; I can often say either it's a disruptive test or vandalism (especially with the Visual Editor, when the user may not even be aware of the problem), but neither the uw-test nor the uw-vand template series say what I'm trying to say. Also, WP:AIV should not require that a level 4 warning be issued. But that's a different proposal. — Arthur Rubin (talk) 09:42, 31 December 2018 (UTC)
• Eliminate level 3 I don't see a need to warn again after level 3, if someone vandalized after being told clearly that what they are doing is vandalism and may lead to a block, they should not be warned further, as this just gives them more chances to vandalize, so I would like to just get rid of the level 3 warning for vandalism and go straight from 2 to something like 4. I find it annoying when people(or bots) go 1 by 1 thought all 4 warnings for an account with no purpose other than adding racial slurs or profanity. We still must keep level 1 for unclear cases where a mistake could be good faith, I know some will say just leave a custom message, but this takes time and would make me much slower at removing vandalism. I could see merging uw-test1 and uw-vand1. Tornado chaser (talk) 20:09, 31 December 2018 (UTC)
• Oppose. The levels don’t have to go to 1–4 — you start at a level based on if the edit is good faith or not. — pythoncoder  (talk | contribs) 22:33, 31 December 2018 (UTC)
• Support. Excessive warnings are just a waste of time for all concerned--simplify, simplify, simplify. I've been a lot less active on anti-vandalism for the past few years, but there used to be some admins patrolling AIV who were very fussy about denying even mild blocks if some newbie IP or SPA wasn't given four full warnings in a very tight timeframe. This proposal seems to be heading towards "reject", but could closer please note that a major reason is that opposers don't think more than 1-2 warnings are necessary to justify a block? --Hobbes Goodyear (talk) 02:50, 2 January 2019 (UTC)
• Oppose 2, Support 3 per Hammersoft. Level 1 for possible good faith edits, specifying why the edit is not appropriate. For intentional violations, start with level 2. Eliminate level 3, level 4 remains the final warning. Jc3s5h (talk) 04:14, 2 January 2019 (UTC)
• Oppose The levels are really a check on automated tools and bots, there's no reason to jump through them all manually for blatant vandalism. Notices for good faith edits can be escalated by automated reverts, so fewer levels would result in more false positives in AIV. Qzd (talk) 01:41, 3 January 2019 (UTC)
• Oppose, but change the level 1 warning to wikilink "did not appear to be constructive" to WP:VAND, which should do to give the hint that blocks can come from this activity. bd2412 T 02:30, 3 January 2019 (UTC)
• Oppose. I mean you need these levels of nuance. I have seen editors who have been stepped up the ladder 1-2-3-4 and turned out to be useful editors. It's rare, but it doesn't happen never. I had one guy, like a year ago, who insisted on adding HTML tags. It took a while, stepped up the warning levels (also some actual person talk) and eventually even blocked for a bit, but he got it. And then he made good edits and AFAIK still is. And he wasn't summarily shunned from this human community, which don't we have enough of that in the world. So allow me this tool please. I have gone level 2-level 4 often enough, or even level 4 directly on rare occasions. But leave me the tools so I don't always have to do that. Herostratus (talk) 03:06, 3 January 2019 (UTC)
• Strongly oppose. Usually while recent-changes patrolling, I utilize all four warning levels whenever necessary, to their full potential. From experience, it appears that the current four-warning system works well; in my opinion, reducing the amount of warnings to two would give the warned too little leeway. Anon. U. 14:41, 3 January 2019 (UTC)
• Uneeded Warnings should be given out on a case by case basis. Sometimes 3 warnings are enough and sometimes edit warring noticeboard is the way to go. I guess in certain situations, four is enough. It all depends, I believe that showing specific examples in the help pages on when to report after 3 warnings and when to go straight to a level 2, 3 or only warning would also be helpful. JC7V (talk) 22:04, 4 January 2019 (UTC)
• Support I agree that not all of them are needed, but I think we should only go down to 3 --DannyS712 (talk) 22:09, 4 January 2019 (UTC)
• Oppose As others have pointed out there is no requirement that you go, 1, 2, 3, 4, 4in, then AIV, although personally I would always go up like that (as long as there is a 4in) without missing any out, I would not necessarily wait until after 4in to report/block. Crouch, Swale (talk) 22:12, 4 January 2019 (UTC)
• Oppose per WP:CREEP, in that doing so reinforces a view not supported by policy that an editor must be warned a certain number of times before blocking. Except for things like discretionary sanctions enforcement, warning a user is a courtesy, not a requirement. Ivanvector (^Talk/_Edits) 22:16, 4 January 2019 (UTC)
• Unnecessary. There already is an "only warning" template. And all the others can be used to suit any situation. No requirement to go through the levels sequentially. -- œ^™ 06:26, 6 January 2019 (UTC)

Good practices for graph by default

We are using a lot of graphs in Wikipedia, and that's a good things. Graph can really help understanding data.

Sadly, a lot of graph on Wikipedia are misleading because they do not follow basic good practices.

To name a few:

• Use of pie chart (almost never a good idea, bar chart should be preferred)
• Origin of the graph at some other value than zero
• No name to axis
• Use of percentage, but axes do not show the complete range from 0-100%

Most of the graph are done using https://en.wikipedia.org/wiki/Template:Graph:Chart. And I believe we can fix a lot of common issue by modifying this template. I opened the discussion on https://en.wikipedia.org/wiki/Template_talk:Graph:Chart#Convert_all_Pie_chart_to_Bar_chart but I think this go beyond the template page.

My propositions:

1. Use 0 as a default value for xAxisMin and yAxisMin.
2. If pie type is used for more that two values force the use of bar chart (rect)
3. if xAxisTitle and yAxisTitle are empty. Display a warning.
4. Be a bit smart about xAxisMax and yAxisMax. How do you detect their is percentage? We can't be sure, but we can use yAxisMax=100 by default when the sum of the value is equal to 100 for example.

Gagarine (talk) 17:47, 30 December 2018 (UTC)

Gagarine, While supporting the notion that many graphs in Wikipedia may fail to follow good practices, I suggest that mandating good practices is potentially problematic. At a minimum, we ought to start with some suggestions and over time, gradually determine whether some hard rules can be implemented. I suspect that might be impossible. While some of your rules sound plausible, it isn't hard to find exceptions. For example, mandating zero as a minimum value is not necessarily the right answer. I am quite aware of that the use of selective values can be misleading, but mandating zero isn't always the right answer. Temperature graphs are a good example in three ways. If the graph is in Fahrenheit, 0 degrees Fahrenheit is an arbitrary value. Similarly, if you adopt the scientific preference for centigrade or Celsius as some people like to say, zero is not the same as for Fahrenheit and is almost as arbitrary. If you then go on to note that those two values are arbitrary but a nonarbitrary value is absolute zero, I suggest that there are very few temperature graphs for which you would really want to use absolute zero as the zero point.

I'm going to push back on the notion that a pie chart is almost never a good idea. Perhaps there is some rationale I'm missing, but when talking about the makeup of some group where the entire group is viewed as 100%, I think pie charts can be appropriate.

It's hard to disagree that having no name on the access is a bad idea. I don't disagree that good practices are to include a name on the axis, but converting it to a rule could be a problem. Imagine that some artists is reproducing some famous graph which happened not to have a name on the axis. feel free to take pot shots at the originator of the graph, but we certainly shouldn't insist that a faithful reproduction of a (presumably public domain) graph should artificially include a name.

While not coming up with a great example off the top of my head, I'll be surprised if we can't come up with a percentage graph in which forcing the range to be 100% is always the best idea. (Not to mention the fact that percentages can exceed 100%, so forcing a rule would be problematic.

I apologize for only focusing on negatives, but we need to be careful when implementing rules that we don't artificially enforce a format that is not the best format. S Philbrick(Talk) 02:17, 1 January 2019 (UTC)

Sphilbrick I'm more about good default than imposing things. For example, I do not want to mandating zero as a minimum value, but if no value is provided then we should use zero.

The page about pie chart start with: "Pie charts are very widely used in the business world and the mass media. However, they have been criticized, and many experts recommend avoiding them, pointing out that research has shown it is difficult to compare different sections of a given pie chart, or to compare data across different pie charts. Pie charts can be replaced in most cases by other plots such as the bar chart, box plot or dot plots.". You can find the references on the pages.

I agree we need to be careful, but I think we can be smarter by using better default for the template and perhaps documentation about good practices. Gagarine (talk) 18:30, 1 January 2019 (UTC)

Gagarine, We are on the same page if your goal is to provide good advice as opposed to mandating rules. That seems to be the case in your recent post where you said "I'm more about good default than imposing things". however, your propositions include "If pie type is used for more that two values force the use of bar chart (rect)". that sounds more like a mandate than advice.

I'm a fan of Tufte. It's been quite some time since I read his notable work and I confess I hadn't recalled his distaste for pie charts, but I read him as being opposed to bad pie charts, and possibly carrying out a little too far by suggesting they are never appropriate. (I confess I haven't gone back to read the original but I'm picking up on some summaries.)

our article on the incoming Congress 115th United States Congress has some nice graphics showing the breakdown of the membership. The use of the individual dots is arguably better than a pie chart but I'll emphasize that it's arguable as opposed to unequivocal. Not everyone making a presentation has easy access to such graphics capabilities. As noted at this site, bad pie charts can be very poor tools, but use of a bar chart for values that naturally some to 100% is not necessarily preferable to a pie chart. I know you included an exception for two values, but United States political articles showing party breakdown are classic examples of data with the values add to 100%, typically have two dominant values but often have a third (or more), and the single most important piece of information is which party is largest and whether it's in excess of 50%. A pie chart with two or three or four values conveys that cleanly and immediately. A bar chart doesn't do the job as well. Imposing a rule that pie charts can only be used if there are exactly two values would lead to the absurdity that when there are no third-party candidates pie charts would be used but if there's a small handful of third-party politicians, then we would switch to a bar chart.

The link does show examples where pie charts can be misused, so again, we are on the same page if we want to avoid their misuse, but my concern is your proposition doesn't sound like advice but in imposition of a rule. S Philbrick(Talk) 16:39, 2 January 2019 (UTC)

Proposal/RfC: Allow ISBN warnings on articles talk pages with ISBN errors

I can do this using my bot. There is a feature on WPCleaner called "Update ISBN warning messages" on WPCleaner. I have been having a discussion about this feature around two weeks ago. The discussion can be found Wikipedia talk:WPCleaner#ISBN warnings. The feature adds and removes ISBN warning messages on the talk pages of the affected articles. It tells the ISBN that is affected and it also tell the reason why the ISBN is affected. This will help Wikipedians fix ISBNs much more easily on English Wikipedia. The template is at User:NicoV/ISBN Warning at the moment. This needs to be moved to Template:ISBN warning. The talk pages with an ISBN warning message will be added into Category:Articles with invalid ISBNs. The WPCleaner script can add the warning message for the affected articles and remove the warning message for the articles that have been fixed automatically. I have tested this on my sandbox and it works. This feature is already available on French Wikipedia and User:NicoV kindly said that he can configure WPCleaner so that this can also happen at the English Wikipedia. The output looks like this:
Your help is requested to fix invalid ISBN numbers (explanation) that have been found in the article:

• • ISBN 0-123-456-78 : The ISBN has only nine digits instead of ten so it is invalid.

Shall we have this brilliant feature which will help the identification of incorrect ISBNs on English Wikipedia? Pkbwcgs (talk) 10:19, 31 December 2018 (UTC)

I thought we already get error messages from apparently faulty ISBNs (and note that this sometimes includes the ISBN as written on the book) - how will this proposal improve things?Nigel Ish (talk) 10:24, 31 December 2018 (UTC)

@Nigel Ish: It will only do this to some ISBNs. At the moment, the ISBN errors are in the form of a list. It will also report incorrect ISBN formatting as well and it doesn't say that directly. Pkbwcgs (talk) 10:29, 31 December 2018 (UTC)

This can also be updated automatically and this makes it easier to locate faulty ISBNs. I am not sure exactly if it will tell the reason why the ISBN is faulty but this can be added manually. It will certainly list the faulty ISBNs. Pkbwcgs (talk) 10:33, 31 December 2018 (UTC)

The main interest I believe is that it's not restricted to ISBN created with templates (which can already be detected as invalid by the template), but it also works for ISBN created with the magic syntax. For example, note 189 on List of shipwrecks in 1946 is using an invalid ISBN but it's not detected because it's done using the magic link. WPCleaner uses several sources for finding invalid ISBN: categories, Check Wiki project, dump analysis... Errors found are listed in Wikipedia:WikiProject Check Wikipedia/ISBN errors but the editor of an article is currently unaware of this. --NicoV ^{(Talk on frwiki)} 10:44, 31 December 2018 (UTC)

It also detects some cases of incorrect syntax which do not produce a magic link (nowiki tags around the ISBN, a colon between ISBN and the number...). --NicoV ^{(Talk on frwiki)} 10:46, 31 December 2018 (UTC)

• I like the idea of a bot that can fix faulty ISBNs. Can I just ask will it do the same for faulty ISSNs (International Standard Serial Numbers)?Vorbee (talk) 14:39, 2 January 2019 (UTC)

  @Vorbee: I think the RfC launched by Pkbwcgs is only for reporting faulty ISBNs with a bot on the article's talk page, not fixing them with a bot. I can probably add some automatic fixing in WPCleaner for ISBN at some point, but it will probably be limited to format problems (wrong dashes, colons, dots...), not problems with the ISBN number itself (wrong length, incorrect checksum...) as it is really difficult to automatically know what is the correct ISBN (or even if the faulty ISBN is real...). Regarding ISSNs, WPCleaner can do the same as with ISBNs, it's just a matter of creating a dedicated warning template and configuring WPCleaner (it's already doing it on frwiki): Pkbwcgs will probably launch another RfC for ISSNs if this one is accepted and it has been put in production. --NicoV ^{(Talk on frwiki)} 16:35, 2 January 2019 (UTC)

  @Vorbee: I am doing a lot of help for ISBNs. I got RegEx ready that is ready for my bot to fix faulty ISBNs using AWB (PkbwcgsBot task 5) and I got approval for updating the list of articles with faulty ISBNs and ISSNs everyday (PkbwcgsBot task 8). The faulty pages are listed at Wikipedia:WikiProject Check Wikipedia/ISBN errors. However, we don't have ISBN warnings on article talk pages which means that users are unaware of some ISBN errors so notices on talk pages that tell which ISBNs are incorrectly formatted or are invalid. If this RfC succeeds, I will wait for User:NicoV to configure this feature for English Wikipedia and then submit a BRFA for my bot to add and remove these ISBN warning messages everyday using WPCleaner. As User:NicoV said, I will submit another RfC for ISSNs if this one succeeds. However, the top priority is for ISBNs because Wikipedia:WikiProject Check Wikipedia/ISBN errors has a huge and long list of faulty ISBNs. Pkbwcgs (talk) 17:11, 2 January 2019 (UTC)

• If you're interested to see the end result, I've configured WPCleaner on my test account to show it: check User talk:NicoV/List of shipwrecks in 1946 for the message that's written. Each part of the message can be configured (either in Checkwiki's configuration or in WPCleaner's configuration). --NicoV ^{(Talk on frwiki)} 18:12, 2 January 2019 (UTC)

  @NicoV: It looks good but that feature is not working on my WPCleaner test version at the moment. It says "You need to define the 'isbn_warning_template' property in WPCleaner's configuration". Pkbwcgs (talk) 18:30, 2 January 2019 (UTC)

  @Pkbwcgs: Yes, it's because I did the configuration only in my test account (NicoVTest), so the configuration is not available to others. And I also modified WPCleaner to allow user's specific configuration for some parameters. If this RfC is accepted, we can put the configuration in the general configuration page. --NicoV ^{(Talk on frwiki)} 20:50, 2 January 2019 (UTC)

  @NicoV: Okay. I have one question. Will this feature automatically report what is wrong with the ISBN (e.g. incorrect checksum, ISBN too long etc.) and tell this information in the ISBN warning message? Pkbwcgs (talk) 20:54, 2 January 2019 (UTC)

  @Pkbwcgs: It will give the same level of details as the example User talk:NicoV/List of shipwrecks in 1946: for example, in the example, the message says that the checksum is incorrect. Messages can be customized i, Checkwiki's configuration, with some variable parts. --NicoV ^{(Talk on frwiki)} 21:25, 2 January 2019 (UTC)

  @NicoV: That is a brilliant feature and I am sure it will benefit English Wikipedia by users being aware of what is wrong with the ISBN so that it can be fixed. Thanks for creating that feature. I am sure it will help. Pkbwcgs (talk) 22:05, 2 January 2019 (UTC)

the first one of these that I check: 0 7209 3995 2 in Lepidosperma squamatum matches the number in the published work at http://museum.wa.gov.au/sites/default/files/1.%20Storr_5.pdf so what do you do about that? Is the number in print authoritative? Graeme Bartlett (talk) 22:45, 2 January 2019 (UTC)

That's a good question Graeme Bartlett, and unfortunately I don't have a definitive answer that works for each case. For your example, the ISBN doesn't match the rules for constructing an ISBN: the computed checksum is "X", but the ISBN shows "2".

As I mostly work on frwiki, I can explain what I do for books in French: the authority for ISBN published in French (ISBN starting with 2, 978-2 or 979-10) is the French National Library. I usually search there for the book and if the notice exists, it will say either

• #1) that's actually the ISBN but it is errorneous (in that case, I use a special parameter "isbn erroné" instead of "isbn" to clearly state that the book has been published with an errorneous ISBN)
• #2) there's an other ISBN (in that case, I user the ISBN on the notice)
• #3) that's actually the ISBN but nothing says that it is errorneous (in that case, I report the problem to the BNF and they usually fix it within a week, so I end up with case #1 or #2)

For books in English, I don't know if there's such an authority... I use Worldcat a lot for them, but I find it useful only when the ISBN in the article is not the same as the one in the notice. In your case, Worldcat has only 2 entries for this book OCLC 24474223 and OCLC 941859713 and none of them gives the ISBN. Same for the National Library of Australia... So I don't know for this one. Maybe remove the ISBN from the article and use the OCLC or NLA link ? Or consider that the ISBN in print is authoritative and have a special parameter for the cite template to acknowledge that ?

I think that before running a bot to put the message on all talk pages, it could be good to do a first pass manually with WPCleaner to fix the easiest ones (Wikipedia:CHECKWIKI/WPC 069 dump: invalid syntax, Wikipedia:CHECKWIKI/WPC 070 dump: the ones that have the 10 or 13 prefix). --NicoV ^{(Talk on frwiki)} 07:53, 3 January 2019 (UTC)

@NicoV: I am also going to configure this feature and test it in my userspace. Pkbwcgs (talk) 11:39, 3 January 2019 (UTC)

@Pkbwcgs: The version of WPCleaner that is available publicly doesn't allow user specific configurations for ISBN warnings (I modified locally my source code to do it with my test account). But I think you can try to configure it on the general configuration page User:NicoV/WikiCleanerConfiguration if you want : in normal use (not bot mode), WPCleaner will only update an existing warning, not create one (that's why I manually created User talk:NicoV/List of shipwrecks in 1946/to do), and only if the modification on the article page has fixed any ISBN related problem. After your test, comment out the configuration to be sure. --NicoV ^{(Talk on frwiki)} 11:50, 3 January 2019 (UTC)

@NicoV: I am trying it on Benfleet railway station (reference 2 has an ISBN error). My configuration is located at User:Pkbwcgs/WikiCleanerConfiguration and the article in my userspace is located at User:Pkbwcgs/Benfleet railway station. The to do page is located at User talk:Pkbwcgs/Benfleet railway station/to do and the talk page is at User talk:Pkbwcgs/Benfleet railway station. However, the test is not working. What am I doing wrong? Pkbwcgs (talk) 12:03, 3 January 2019 (UTC)

@Pkbwcgs: Check my previous message: #1 you have to do the configuration in the main configuration's page User:NicoV/WikiCleanerConfiguration because user's specific configuration is not accepted by WPCleaner for those parameters. And you also have to fix at least something related to ISBN in the page User:Pkbwcgs/Benfleet railway station to trigger update of the ISBN warning. Otherwise, everything is done in bot mode, but it's not possible to select on which page it will work. --NicoV ^{(Talk on frwiki)} 12:41, 3 January 2019 (UTC)

@NicoV: Am I allowed to configure this feature on User:NicoV/WikiCleanerConfiguration and when I am done with testing, shall I revert it? Pkbwcgs (talk) 16:05, 3 January 2019 (UTC)

@Pkbwcgs: Yes, I think you can change User:NicoV/WikiCleanerConfiguration, use User:NicoVTest/WikiCleanerConfiguration to see what is needed. When you're done testing, you can either revert your modifications or just put them in comment (a # at the beginning of a line is a comment). --NicoV ^{(Talk on frwiki)} 16:16, 3 January 2019 (UTC)

@NicoV: For some reason, the feature is not changing anything. It is only saying it is retrieving the contents. Pkbwcgs (talk) 16:35, 3 January 2019 (UTC)

@Pkbwcgs: You have to fix an ISBN problem in your test page with WPCleaner to trigger the update (I did this one for my test). --NicoV ^{(Talk on frwiki)} 16:47, 3 January 2019 (UTC)

@Pkbwcgs: I forgot that there was a menu to add the ISBN warning directly, that's why I was talking about fixing an ISBN in the page. But, I tried to add the warning through the menu and it didn't work: I probably have to modify something in my code to handle the configuration for enwiki (it's different than frwiki). I'll keep you posted. --NicoV ^{(Talk on frwiki)} 16:57, 3 January 2019 (UTC)

@NicoV: It updated the ISBN warning message here. However, I think English Wikipedia would benefit from this if this warning message can be directly added and removed from the talk page of the article by my bot rather than having another "To do" subpage to make this feature work. Pkbwcgs (talk) 18:32, 3 January 2019 (UTC)

@NicoV: Your feature works brilliantly. It was unexpected as I pressed the wrong button but I reverted the edits. However, things like this diff would absolutely benefit Wikipedia. Pkbwcgs (talk) 21:06, 3 January 2019 (UTC)

@Pkbwcgs: WPCleaner can put the warning directly on the talk page or on the subpage depending on the configuration. The parameter general_todo_templates is for adding a template directly on the talk page (that's what happened in this diff): I think it's not correctly configured for enwiki as {{todo}} doesn't accept the list of tasks as a parameter. The parameter general_todo_subpage is for the name of the subpage to use when the warning is put on the subpage. To decide whether or not to put the warning on the subpage or on the talk page: general_todo_subpage_force forces the use of the subpage regardless of the other tests ; if the subpage already exists then the warning is put on the subpage ; if there's already a template on the talk page that provides a link to the subpage (from general_todo_link_templates, which seems to be the correct list for {{todo}}) then the warning is put on the subpage. --NicoV ^{(Talk on frwiki)} 10:16, 4 January 2019 (UTC)

As can be seen in this diff (and in the example in the OP) the rendered template has broken list markup (one asterisk is displayed as such; and a final line has two, not one, bullet points). This needs to be fixed, before any bot run please. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:28, 4 January 2019 (UTC)

@Pigsonthewing: That is a good point. I am actually not sure how to fix that problem. The template is located at User:NicoV/ISBN Warning at the moment if you would like to fix it there. It is transcluded through Template:To do so the problem could also be in that template as well. I have tried to fix it but it didn't work. Pkbwcgs (talk) 12:37, 4 January 2019 (UTC)

@Pkbwcgs: See User:NicoV/ISBN Warning/Sandbox. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:07, 4 January 2019 (UTC)

@Pigsonthewing: That looks good. @NicoV: Can you please amend WPCleaner so that this feature makes the template render as:

{{todo|
* {{ User:NicoV/ISBN Warning | revisionid=875583851 | 978-3-931473-16-4 | Computing the checksum gives 7, not 4 }} -- 21:02, 3 January 2019 (UTC) <!-- This line is updated from time to time by a bot. -->}}

We don't need a "to do" subpage for every talk page; I think it is updating fine so please don't change that. However, we want the line space after {{to do|. Pkbwcgs (talk) 18:57, 4 January 2019 (UTC)

@Pkbwcgs: You can try the last version... --NicoV ^{(Talk on frwiki)} 23:07, 4 January 2019 (UTC)

@NicoV: What do you mean by the "last version"? Pkbwcgs (talk) 12:03, 5 January 2019 (UTC)

@Pkbwcgs: If you're running WPCleaner normally (throught getdown, or with the provided scripts), it should update itself automatically to the release I did yesterday. Version number is the same. --NicoV ^{(Talk on frwiki)} 12:21, 5 January 2019 (UTC)

@NicoV: Shall I move the working ISBN warning template from User:NicoV/ISBN Warning to Template:ISBN warning? Pkbwcgs (talk) 12:25, 5 January 2019 (UTC)

@Pkbwcgs: It will be better if you wish to run the bot. --NicoV ^{(Talk on frwiki)} 13:22, 5 January 2019 (UTC)

@NicoV: Okay. I would still prefer to gain some consensus before I file the BRFA for this to run. Pkbwcgs (talk) 13:31, 5 January 2019 (UTC)

Bot to add Template:reflist-talk

There is a bot proposal to add {{reflist-talk}} to talk pages that would benefit from it.

The bot would do this:

Edit

Do you think this is a good idea?

details inside

How pages look with {{reflist-talk}} Talk:1923 in literature How they look without: Talk:Hayashi Hōkō Talk:Brezovica pri Borovnici Talk:Sa'ad Al-Faqih Talk:Freeport-McMoRan Talk:8 Talk:E Talk:Propaganda in the Soviet Union Talk:Buildings and architecture of Bristol Talk:Detection limit Talk:Plasmodium ovale Problems when {{reflist-talk}} is missing: Citations get mixed together at the bottom of the page with citations from other sections on the page In long talk pages, clicking the sup-script [1] back and forth is burdensome and hover-over doesn't work (Example) The autogenerated list is mixed in with the last section giving the appearance that section has the citations (Example) Will prevent editors trying to fix the problem by adding <references /> in the middle of the page which interferes with citations added below it. When sections get auto-archived it can create unintended effects if other sections in the archive have citations (such as one of the problems above) It is messy and disorganized compared to {{reflist-talk}}

-- GreenC 23:03, 2 January 2019 (UTC)

• I think this is a good idea. I add them a lot, and if a bot is to take over the work that would be great. –Ammarpad (talk) 05:13, 3 January 2019 (UTC)

Do the archive-bots handle this correctly (seeing the the datestamp prior to the template as the date of last message in a thread)? Obviously this is more about the archive-bots an the template in general, but something to double-check before deploying the template more widely... DMacks (talk) 09:28, 3 January 2019 (UTC)

It depends if the archive-bot is scraping dates from signatures, or using APIs in some way to check revisions within a section. However my bot doesn't do section editing it is a whole page save. Really not sure how all those things will play out. But honestly if the choice is to have the template or not, it seems like an archive-bot time counter wouldn't be a major sacrifice. This is something people are doing manually anyway. The template has 26k transclusions which is about 50% of all possible cases where it could be used, has already been done by the community, so it's a good estimation that many of the rest will get updated eventually one way or another (as Ammarpad says "I add them a lot"). -- GreenC 16:56, 3 January 2019 (UTC)

• I personally think it's a great idea, I here and there add reflist-talk to various sections manually when I spot them but It would be easier if a bot could do this, I tend to add it to all sections that use refs (not sure if that's the correct way or not) so I'm not exactly sure if these would need to be added to more than one section on one talkpage .... Dunno but either way support. –Davey2010^Talk 22:12, 5 January 2019 (UTC)
• I love the idea. 142.160.89.97 (talk) 05:22, 6 January 2019 (UTC)

Bot to add {{Section sizes}} to talk pages

There is a bot proposal to add this template to ~6300 article talk pages where the article is longer than 150,000 bytes (per Special:LongPages), like in this edit.. If this is something you would like to have happen, community support for running bots is required. This is the place to voice if you want a bot to do this or not. -- GreenC 15:56, 4 January 2019 (UTC)

150,000 bytes seems too low of a cut-off. Since the point of the template appears to be for oversized articles, something like 250000 would make much more sense as a cut-off because most pages at around 150000 bytes don't really need to be reduced in size (though of course things vary, and certain bare lists that size may need to be cut down or split). Galobtter (pingó mió) 16:24, 4 January 2019 (UTC)

@Pigsonthewing: -- GreenC 17:09, 4 January 2019 (UTC)

I disagree with Galobtter's assessment (most 150000 bytes articles are too long; see WP:AS); but in any case such bike-shedding is irrelevant; this template is to inform editors of the relative size of sections of articles; it does not mandate nor even recommend a split. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:15, 4 January 2019 (UTC)

WP:AS mainly talks about readable prose size, and most 150000 byte articles have a readable prose size much less than what necessitates a split per that guideline (indeed, there are numerous FAs which are more than 150000 bytes long). Every template adding to a talk page adds to the clutter, so there has to be some benefit; since that benefit would mainly be when talking about a split, talking about the cutoff seems important here when automatically adding the template. Galobtter (pingó mió) 15:35, 5 January 2019 (UTC)

A 250,000 limit is about 1000 articles. How was 250,000 arrived at? Asking because it seems to make a big difference how many articles would have the template based on where the cut-off number is. Personally I don't think 6000 articles is very much at all in the scheme of 5700000 articles, with that few I doubt most people would ever see the template in their lifetime at Wikipedia. -- GreenC 16:03, 5 January 2019 (UTC)

Large size articles would very disproportionately be popular ones which already have a very large number of templates on their talk page. (pages like Talk:Cristiano Ronaldo and Talk:Barack Obama would be the ones tagged, not obscure articles) Galobtter (pingó mió) 17:01, 5 January 2019 (UTC)

Like I said, "such bike-shedding is irrelevant". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:57, 5 January 2019 (UTC)

Are we talking 150kb bytes of readable text, or of total page size? 150kb of readable prose is almost certainly too long except in exceptional circumstances when there's a valid argument that all the material has to be kept together, but 150kb save size isn't unusual for pages that include lots of complex template markup (Barack Obama, for instance, uses 340kb of code to produce 79kb of readable prose). I'd quite strongly oppose using a bot to tag everything just based on the size of the code, as that will end up flagging large numbers of completely non-problematic articles as being too long, and consequently annoy a lot of people for little or no obvious benefit. ‑ Iridescent 17:10, 5 January 2019 (UTC)

There seems to be some confusion. The template being discussed does not suggest an article is too long. Jack N. Stock (talk) 17:15, 5 January 2019 (UTC)

Agree with that. And if we aren't talking about articles that are too long, what are we talking about? The template decumentation is (as usual) unclear, as is the proposal above. Johnbod (talk) 17:17, 5 January 2019 (UTC)

We're talking total page size (going off Special:LongPages which uses that), thus my comments. Galobtter (pingó mió) 17:19, 5 January 2019 (UTC)

You mentioned Talk:Barack Obama - why are we tagging that exactly? Johnbod (talk) 17:22, 5 January 2019 (UTC)

Well, the bot would be adding the template to the talk page of any article with more 150kb of wikitext, which would include Barack Obama (which has 340kb of code as Iri mentioned) Galobtter (pingó mió) 17:25, 5 January 2019 (UTC)

Exactly! But why? The original tool request By Andy was very much talking about "very long pages". Johnbod (talk) 17:27, 5 January 2019 (UTC)

It's useful information for whatever purpose. Such as breaking sections up into sub-sections, breaking sections into separate articles, refactoring sections by deleting content. Large article are targeted for the template because they are most likely to have work of this nature it helps users identify which sections they might want to work on. -- GreenC 17:32, 5 January 2019 (UTC)

FWIW Barack Obama is 85,394 as plain-text according to the API [1]. -- GreenC 17:44, 5 January 2019 (UTC)

• In light of the above replies, strongly oppose; this is clearly not an attempt to address an actual problem, but a pretext to use a bot to mass-spam this template under the guise of fixing a non-existent problem. ‑ Iridescent 19:27, 5 January 2019 (UTC)
  • So much for WP:AGF... Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:27, 5 January 2019 (UTC)
• Don't see the point of this. Headbomb {t · c · p · b} 20:58, 5 January 2019 (UTC)
• Oppose - This is a solution looking for a problem. –Davey2010^Talk 22:21, 5 January 2019 (UTC)
• I think this could be useful, but only if it also included other important size information such as readable prose size in bytes and words. Something like the output of this script would be beneficial: User:Dr_pda/prosesize.js. - MrX 🖋 22:24, 5 January 2019 (UTC)
• @GreenC: As the example edit added the template to a talk page that was 65 bytes in size, it appears the opening comment is incorrect and misleading. This is not about the size of the talk page. Suggest a strikethrough and correction to prevent confusion like mine. ―Mandruss ☎ 22:37, 5 January 2019 (UTC)

fixed -- GreenC 23:14, 5 January 2019 (UTC)

Fixed without strikethrough (and underscore), but whatever. FTR, the correct form is "article talk pages that are where the article is longer than 150,000 bytes". ―Mandruss ☎ 23:23, 5 January 2019 (UTC)

• Don't see the useful purpose. Not saying there isn't one, but it has not been explained. Hard to avoid bike-shedding when asked to comment on something where the intended purpose is not specified. · · · Peter Southwood ^(talk): 05:48, 6 January 2019 (UTC)

Talk page health: ranking template enabling editors to rank the health of a talk page.

Background

In July of 2018, Wikimedia made a call for proposals at their grants idea lab. The request called for ideas to improve Wikipedia community health. My first thought was that although some high-tension areas of Wikipedia are well trafficked and watched, there may be other less visible areas where discussion health may be an issue.

Proposal

To make a template which can be optionally placed at the top of a talk page, with radio buttons, check boxes or single-select options ranking a talk page's health.

• Information would then
  • Be used to identify areas requiring feedback
  • Create statistical data about areas requiring attention

Wikimedia proposal page

Some discussion regarding this proposal has taken place at Wikimedia

• the discussion includes:
  • A basic description of the proposal
  • Some example screen shots
  • A call for required experts and collaborators on the project
  • Some ideas about how to proceed on en-wiki.

This proposal has been mentioned at Wikipedia_talk:Community_health_initiative_on_English_Wikipedia

Edaham (talk) 02:52, 4 January 2019 (UTC)

Discussion

What are the ranking criteria? Personal opinion? If so, what useful purpose would this serve? Would there be any actionable consequences? · · · Peter Southwood ^(talk): 06:00, 6 January 2019 (UTC)