Wikipedia:Biographies of living persons/RfC to add Pending Changes to all BLP with few or no watchers

From Wikipedia, the free encyclopedia
Jump to: navigation, search
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
This RfC is now closed. For the result go to #May Protection policy be modified to Pending Changes protect BLPs having few or zero watchers?. Armbrust The Homunculus 23:01, 25 December 2013 (UTC)

It has been over a year since Wikipedia:PC2012/RfC 2#May PC be applied to pages to protect against violations of the policy on biographies of living persons (WP:BLP)? This RfC considers a special case of BLPs: unwatched BLPs.


We have a duty to take proactive care of articles on living people. Of all our articles, unhelpful BLP changes have the most potential to create angst. Pending Changes for unwatched and minor BLPs is a simple way to protect article subjects from drive-by vandalism and BLP violations.

Protecting BLPs[edit]

May Protection policy be modified to Pending Changes protect BLPs having few or zero watchers?[edit]

There is a clear consensus against modifying the protection policy to allow the pre-emptive protection for under-watched BLPs. Armbrust The Homunculus 22:50, 25 December 2013 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

The existing rationale is that you shouldn't be protecting something that's not being vandalised (etc), but without watchers how will we ever know if something is being vandalised?

Should Wikipedia:Protection policy#When to apply pending changes protection

be modified to include:

... unless it may be an unnoticed BLP violation.

This will permit Pending Changes protection on BLPs with a justification of the article having few or zero watchers.

Yes (protect)[edit]
  • Support There are far too many articles on Wikipedia that aren't watched by enough people. I looked at a BLP article I wrote a while back and it has two watchers, me and one other user. Widespread application of pending changes to articles where vandalism runs the risk of causing harm to BLP subjects is a moral imperative to protect the project and protect article subjects. —Tom Morris (talk) 13:06, 27 November 2013 (UTC)
    A "moral imperative"? BLP is a policy on an online encyclopedia, not a tenant in some fanatic cult (though I'm starting to wonder). John Reaves 03:51, 28 November 2013 (UTC)
  • Support Another possibility would be to maintain watchlists for these and other unwatched pages, but (particularly on BLPs of non-public figures) I would prefer the option that reduces visibility of unpatrolled edits if that can be done without causing backlogs at PendingChanges similar to those that often occur at WP:Articles for creation. Peter James (talk) 22:53, 27 November 2013 (UTC)
  • Support Implemented thoughtfully this will only add value to the project. Josh Parris 03:26, 28 November 2013 (UTC)
Are nominators/proposers supposed to vote? You initiated this proposal, which makes you the nominator/proposers. Wait, they do probably, but double-voting is disallowed. George Ho (talk) 06:59, 28 November 2013 (UTC)
  • Support, good idea, logical and sound. — Cirt (talk) 18:23, 9 December 2013 (UTC)
No (leave unprotected)[edit]
  • Leave unprotected because it would require too much work over at Wikipedia:Five pillars to rewrite the whole "anyone can edit" thing. John Reaves 17:12, 25 November 2013 (UTC)
    No, it wouldn't. You can edit an article that has pending changes activated. The edits won't appear live immediately, but you can edit the article. In this sense, it is a huge improvement over semi-protection. —Tom Morris (talk) 09:24, 27 November 2013 (UTC)
    You can also full protect every article and require approval for every edit from Jimbo, ArbCom and the Knights of Columbus. That doesn't make it a good, rational idea. The mere existence of procedure (ie pending changes) does not authorize its site wide implementation. John Reaves 03:58, 28 November 2013 (UTC)
  • Bad idea Amongst other things, this would inevitably mean a secret back room decisionmaking process as the basis for the decision (unwatched status of articles) is secret. And we already have too many automated or semi-automated processes that interfere with real editors. Interfering with a real editor should take real people and a public process. North8000 (talk) 15:19, 26 November 2013 (UTC)
    The problem with making these decisions in public is that they would advertise which BLPs were unwatched and thus ripe for vandalism and BLP violations. Whomever makes the decisions would need to do in secrecy, just so that the candidate articles that don't get protected aren't exposed to nefarious risks. Josh Parris 03:25, 28 November 2013 (UTC)
    That is exactly my point and you are reinforcing it. This proposal (i.e. to do it based on the "unwatched" status) would force those decisions to be made in secret. North8000 (talk) 12:55, 2 December 2013 (UTC)
  • This is just going to mean that positive edits hang around invisible to the vast majority of readers for months or years -- if the page isn't watched, who is going to accept the revisions? Espresso Addict (talk) 12:18, 27 November 2013 (UTC)
    The reviewers who patrol Special:PendingChanges. —Tom Morris (talk) 13:05, 27 November 2013 (UTC)
    Who won't stand a chance of correctly assessing most of the changes to articles outside their area of expertise. What is really needed is more subject experts prepared to watchlist biographies of people who are not well known. Espresso Addict (talk) 03:32, 28 November 2013 (UTC)
    This is no different to the current situation, except changes go live immediately and absolutely no-one reviews them. The Wikipedia:Reviewing guidelines are clear: you don't need to be a subject expert, just capable of spotting BLP violations, vandalism, copyvios and legal-disasters-waiting-to-happen. Pending Changes is for protecting against egregious BLP violations, not ensuring each edit is a step towards FA status. Josh Parris 03:40, 28 November 2013 (UTC)
    In my experience with bios of scientists, the main problematic edits are PoV-pushing edits referenced to sources of differing reliability that take quite a lot of work to accept, reject or assimilate in a more appropriate fashion. I've seen little or no obvious vandalism of the sort that's common in celebrity or politician articles. Espresso Addict (talk) 04:16, 28 November 2013 (UTC)
    How is this disaster scenario you speak of prevented from occurring right now? Josh Parris 21:14, 28 November 2013 (UTC)
    By large numbers of editors/IPs patrolling pages within their sphere of expertise. Not restricted to logged-in editors, let alone those who've bothered to apply for bells & whistles; I've seen a lot of sensible reverts by IPs on scientific articles. Espresso Addict (talk) 00:09, 29 November 2013 (UTC)
    If a reviewer misses those subtle problems and accepts the revision, then the IPs will see it (as they do now) and correct it (as they do now). Josh Parris 11:49, 3 December 2013 (UTC)
  • This is to put it bluntly, a proposal to throw the "anyone can edit" principle out of the window. The proposed wording above would allow the indefinite protection of almost any article about a living person, a possibly living person, or an article which might contain material about living people. (The fact that a page is on someone's watchlist doesn't mean that the watcher is checking their watchlist, or that they would be competent enough to recognise and remove a BLP violation if they see one.) There are 640,000 articles in the first category alone. That's a substantial fraction of all the articles on the project. For comparison we have about 800 articles under pending changes protection at the moment, about 4,000 indefinitely semi-protected articles and (I think) about the same number with temporary semi-protection, so the proposal would expand the number of protected articles by a factor of about 100. Yes, pending changes is a form of protection, it's functionally equivalent to semi-protection with a better interface for suggesting edits. Quite apart from the ethical issues, such a drastic expansion of pending changes protection would make it likely that reviewers would not be able to keep up with the flow of edits. This could actually make it harder for people to fix BLP violations. And, of course, closing off so many articles wouldn't help with our problems of editor recruitment. There are a number of things we could do to improve our monitoring of BLPs, but this is an extremely bad idea. Hut 8.5 22:53, 27 November 2013 (UTC)
    There's a difference between editing a page and publishing it automatically as the current version. Peter James (talk) 23:12, 27 November 2013 (UTC)
    And some of that difference is that the enthusiastic newbie who adds their nugget won't get that happy glow of seeing it go live, and won't thereby be encouraged to add their nuggets to other articles. Espresso Addict (talk) 03:36, 28 November 2013 (UTC)
    This is a valid and useful argument. Josh Parris 03:41, 28 November 2013 (UTC)
    Non-autoconfirmed editors can suggest changes to semi-protected pages. They can make some suggestion on the talk page, and if an autoconfirmed editor sees that suggestion and thinks it is good then they can implement it. We have a special system of organising and responding to such suggestions. The fact that we have this system doesn't mean semi-protected pages are considered anything other than protected. What then is the difference between this and pending changes? It's largely technical: pending changes has a better interface for suggesting changes and approving them. If visibility of your edits to the general reader is conditional on someone else's approval then all you can do is suggest changes. The current protection policy recognises this and treats pending changes as an alternative to semi-protection. So yes, this is a proposal to protect an awful lot of pages. Hut 8.5 07:48, 28 November 2013 (UTC)
    [To Hut8.5] With the difference that any autoconfirmed user can implement changes suggested to semi-protected pages, while only a few thousand reviewers & admins can implement those suggested by the pending changes system. Espresso Addict (talk) 00:09, 29 November 2013 (UTC)
    I'm not arguing that pending changes doesn't make it easier to request and process changes. It does, although it isn't suitable in some situations due to unrelated reasons. I'm just trying to counter the view, expressed above, that pending changes doesn't inhibit people's ability to edit pages in a way that semi-protection does. And the number of pages involved is relevant, because it means that protection would "detract from the rest of the encyclopaedia" (as you put it) by tearing up one of the fundamental principles this project is based on. Hut 8.5 08:07, 29 November 2013 (UTC)
    Last night I looked at Category:Wikipedia semi-protected edit requests and at Special:PendingChanges; edit requests had 10 entries, PC had 3. This morning edit requests had 10 entries, PC had zero. Pending Changes is not only a smoother implementation for editors, it's also smoother for reviewers. Yes, the change in wording has the potential to protect a lot of pages; the question is do those pages deserve protection and will the protection add value to those articles; will protection detract from the rest of the encyclopaedia? Josh Parris 21:14, 28 November 2013 (UTC)
    Now Category:Wikipedia semi-protected edit requests is at 22, Special:PendingChanges at 1. Josh Parris 11:49, 3 December 2013 (UTC)
  • Pending changes is flawed in that it is a largely human based operation. Implementing pending changes site-wide (on BLPs) effectively creates a bias/POV filter for everyone with an agenda and a watchlist. Neutral POV goes out the window when a select few get to decide what it is. John Reaves 04:13, 28 November 2013 (UTC)
    Did you already vote? Ah, well, it doesn't matter. Comments like this are new. George Ho (talk) 07:09, 28 November 2013 (UTC)
    not a vote here. John Reaves 00:57, 29 November 2013 (UTC)
    How is this disaster scenario you speak of prevented from occurring right now? Josh Parris 21:14, 28 November 2013 (UTC)
    By the fact that anyone can edit! How many users do you think made their first edit anonymously? How many editors with accounts still edit anonymously at libraries, public computers, etc. when they can't be bothered to log in? How does marginalizing our largest subset of editors help the encyclopedia? John Reaves 00:57, 29 November 2013 (UTC)
    I understand. If "Aaron A Aaronson (born 1983)" is changed to say "he was the first man on the moon" or whatever, an anon will see that and fix it; if it's hidden they won't know it needs fixing. I've seen plenty of quality edits from IPs, including vandalism fixes. Your position is that having things that need fixing be visible to all readers - so that one of them will notice and fix the problem - is superior to hiding edits until they're reviewed by an experienced editor. It's superior because it gives all editors immediacy, right? Josh Parris 07:48, 29 November 2013 (UTC)
  • We can't protect such BLPs from future vandalism. Protection must be based on past violations on any page. As for BLPs, whatever happen to discussion first or one case request? Information about living person is generally sensitive, but we can protect the page that has been edit-warred, disrupted, and/or disputed. We can't protect a page whose sensitive information (and removal of it) has not yet been recorded. --George Ho (talk) 07:09, 28 November 2013 (UTC)
    Yes we can protect pages pre-emptively... by changing policy. That's what the RfC is about. —Tom Morris (talk) 12:29, 28 November 2013 (UTC)
  • If Wikipedia started to protect pages pre-emptively, it would make much more sense to use semi protection. AFAIK, nobody has demonstrated the scalability of pending changes on en-wp on the suggested scales. It very likely would simply end up as another massive backlog and editors burning out. Besides, readers would not understand why the article does not change even though they edited it. Pending change is essentially another cryptic nail to the insular coffin of Wikipedia. Semi protection is simpler: there is simply no edit button. Most of the readers probably do not know what talk pages are or have the interest to figure out how they can request edits, so there would not be a backlog of edit requests either. Keep it simple, folks. (talk) 12:28, 28 November 2013 (UTC)
  • Oppose for two reasons. Firstly it will create an insurmountable backlog resulting in either valid contributions not being added to articles, or if some auto-approve feature is added to prevent this then vandalism will get in anyway. Secondly, this seems like another attempt to bring Flagged Revisions in through the back door - something we saw before with the unending trial phase. The way BLP is always used as a justification for introducing PC/FR is little different in my opinion from governments using the fear of terrorism to introduce draconian security laws and detention without trial. The community rejected sitewide FR, maybe it has some value from a protection standpoint, but it should never be used preemptively. --W. D. Graham
  • Oppose per W.D. Graham. Sven Manguard Wha? 06:23, 29 November 2013 (UTC)
  • Oppose - I've never really understood why BLPs seem to need more "care" than virtually any other kind of article, but I'm against this simply on the grounds that pre-emptive protection of any kind (except for highly-used templates) is a pretty bad idea. Whatever happened to "Wikipedia, the free encyclopedia that anyone can edit"? Pre-emptive protection (for BLPs or otherwise) would discourage constructive edits to articles by non-registered editors, and not everyone would know how to request an edit on the talkpage. Also, the sheer effort needed to implement protections would be difficult, even with bots. Narutolovehinata5 tccsdnew 00:15, 30 November 2013 (UTC)
    The reason BLPs need more "care" is that there are actual people behind them who are affected if we get them wrong. If Wikipedia gets the article about Jupiter or Human rights or Cheddar or Charles Darwin wrong, that hurts our credibility. If we get the article about John Siegenthaler wrong, we not only hurt our own credibility, we also potentially harm and libel a living individual. Which following the age old ethical principle of "don't be a dick" seems like a bad thing to do to most non-sociopaths. —Tom Morris (talk) 10:07, 30 November 2013 (UTC)
    Technically the same could happen to our articles on companies; say, if we get our article on Google wrong, theoretically it could also harm the company, and not only the company itself, but also the (living) people working for the company. Narutolovehinata5 tccsdnew 14:22, 30 November 2013 (UTC)
  • Oppose. We are an encyclopedia here, and encyclopedias shouldn't treat different topics differently. Narutolovehinata5 is precisely correct, and "don't be a dick" applies to anything/everything. We have no good reason to pretend that these articles need more care than any other under-watched article. Nyttend (talk) 20:04, 30 November 2013 (UTC)
  • Oppose. If the articles are unwatched, there's a good chance the edits will sit and rot without approval, especially as I imagine this could lead to a rather large backlog. If the interface for Pending Changes is improved, it might alleviate this problem, but as it stands, it doesn't seem workable to me. Kaldari (talk) 03:51, 1 December 2013 (UTC)
  • Oppose. A "solution" in search of a problem. No evidence has been presented to demonstrate that there is a problem. If there were a problem, this wouldn't be the best way to fix it. Rivertorch (talk) 04:39, 2 December 2013 (UTC)
  • Oppose. "The free encyclopedia that anyone can edit". NOT "the free encyclopedia that anyone can edit, provided they don't do anything that relates to a living person, directly or indirectly." Ypnypn (talk) 19:46, 2 December 2013 (UTC)
  • Oppose The scale of articles that we are talking about here, and the lack of a real need to restrict a vast amount of articles from the free editing and trust that we give to editors, makes this, along with a host of other reasons, a bad idea. Take a look at User:Jimbo Wales's "You can edit this page!" section. Ramaksoud2000 (Talk to me) 19:56, 2 December 2013 (UTC)
  • Oppose per most of the comments above, particularly WD Graham. It's a lousy idea to preemptively protect a whole swath of articles based on nothing more than an abstract statistic rather than a specific, demonstrated need. postdlf (talk) 21:59, 2 December 2013 (UTC)
  • Oppose per pretty much everyone above. Back door, bad for the encyclopedia, and PC is exactly wrong for these types of articles. Hobit (talk) 00:43, 4 December 2013 (UTC)
  • Oppose. If an article has no watchers, then the pending changes potentially would just sit there for ages and ages. Above that, the number of watchers is entirely arbitrary; you could have a whole bunch of blocked sockpuppets watching the article to keep it over the threshold, for starters. And one highly active editor who keeps a close eye on their watchlist is better than a large amount of semi-active editors who don't keep an eye on their watchlist. Lukeno94 (tell Luke off here) 21:59, 7 December 2013 (UTC)
  • Oppose per everyone else, and the anyone can edit pillar. Tazerdadog (talk) 19:47, 8 December 2013 (UTC)
  • Oppose the creeping implementation of flagged revisions (however it's disguised), as always.—S Marshall T/C 08:53, 9 December 2013 (UTC)
  • Oppose agree per above comments. Will require too much work and hamper article improvement. --LT910001 (talk) 03:54, 11 December 2013 (UTC)
  • Oppose. In my experience with pending changes, it should not be used as a substitute for the lack of page watchers. In some of the BLPs under both my watchlist and under pending changes, some vandals periodically try to insert sneaky vandalism. What ends up happening is that either someone on pending changes patrol accepts the changes without realizing it violates the BLP policy, or it stays backlogged for several hours. It then ends up staying that way only until I, or another editor who has it on their watchlist and/or is familiar with the subject, correctly fixes the article. Zzyzx11 (talk) 06:46, 16 December 2013 (UTC)
  • Oppose I reject the argument that this is antithetical to the idea that anyone can edit, PC is a tool like any of the many others we use to protect encyclopedic content from vandals and others whose purpose is something other than providing neutral content. However, I do not believe PC or any other form of page protection should be used preemptively. Beeblebrox (talk) 19:21, 16 December 2013 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Unprotecting BLPs[edit]

Closing section as moot. The idea of protecting under-watched BLPs was rejected in the above section. Armbrust The Homunculus 22:57, 25 December 2013 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

May admin-bots select and Pending Changes protect BLPs having few or zero watchers?[edit]

This presumes that such protection is supported by the community.

BLPs would have to be protected slowly, so that Pending Changes doesn't become flooded. Presumably protecting additional BLPs would occur only when there are few outstanding Pending Changes.

For clarity: only administrators operate admin-bots per Wikipedia:Bot policy#Bots with administrative rights.

Yes (let bots decide and protect)[edit]
  • Support but it needs to be done very carefully so as to not overwhelm our 5,964 users with Reviewer rights. Bots would be able to pick those BLPs watched by the least number of editors, where those editors are the least active, where there have been a history of reverts to the BLP, and those reverts have taken much longer than average - a lot of factors that a human would take a long time to form an opinion on. Josh Parris 03:32, 28 November 2013 (UTC)
  • Support If we impose this suggestion, there will be tons of articles to protect, and requiring humans to protect all of them will waste tons of human time. The idea of "protect page tagged as BLP with X or few watchers" is very very simple for a bot to follow. I don't particularly understand Hut's bit about "non-admins would have the ability to protect pages"; you could do this page-by-page by removing articles from your watchlist, but this would require coordinated action by lots of nonadmins, and nobody could influence protection on pages that aren't already watched. Nyttend (talk) 20:10, 30 November 2013 (UTC)
    • I believe he means you could tag something as a BLP, properly or not. Gigs (talk) 22:32, 3 December 2013 (UTC)
No (this is not a job for bots)[edit]
  • As a general statement, the concept of admin bots is one that unsettles me deeply. This is not a case where I see the task as being so urgent and so trivial that it overrules that unease. Sven Manguard Wha? 06:20, 29 November 2013 (UTC)
  • While practically speaking this would require a bot to implement, using an adminbot like this would mean that non-admins would have the ability to protect pages. Hut 8.5 08:11, 29 November 2013 (UTC)
    Nyttend: what I meant is that anybody could add Category:Living people to an article, and provided it has sufficiently few watchers (most articles will) the bot would come along and protect it. Given the sheer number of articles in that category I doubt there would be much oversight. You would think that deciding whether an article is about a living person would be straightforward, but I have seen a surprising number of people screw it up. Hut 8.5 21:58, 30 November 2013 (UTC)
    Thanks for the response. The ambiguity over whether we consider a page BLP-related is just another reason I'm opposed to this proposal in the first place. However, the sheer number of pages that this proposal considers, together with the extremely small chance of your situation happening (see next sentence), means that we'll still be a lot better off by using a bot than by not using a bot if we end up passing this proposal. In my mind, the situation you describe is vandalism; I suppose that it's likely for someone to do this on one or two pages, but the chances of it happening to a substantial number of pages are extremely small. If it starts happening on lots of pages, we'll discover it, block the vandal, rollback the category, and unprotect the page. Nyttend (talk) 22:27, 30 November 2013 (UTC)
    I admit it is rather unlikely that some vandal would add large numbers of pages to the category in order to protect them, though I doubt it would be detected if they did. There is plenty of scope for good faith abuse of the system though. Possibly a non-admin might add the category to an article with content relating to living people to stop it being vandalised. Simple screwups are likely to lead to some pages being wrongly protected as well - I've seen articles about Ancient Greece and video game characters labelled as "living people". There is a reason we have always given these decisions to admins. Your suggestion below would be helpful though. Hut 8.5 11:37, 1 December 2013 (UTC)
  • The whole idea of or auto or semi-auto protections has clearly been rejected. So "what if we did" questions now have a false premise and should be ended because answers would be meaningless unless they were taken out of context to mean the opposite of the overall result. To avoid this, answers against ANY form of auto and semi-auto protection in the previous question should be considered responses against bot protection, as they clearly are. North8000 (talk) 11:24, 9 December 2013 (UTC)
  • I agree with comment by Sven Manguard, above. Cheers, — Cirt (talk) 18:24, 9 December 2013 (UTC)
  • NO Bots should not ever be used to make these types of decisions. I could see letting them report potential problems at RFPP for a human admin to review, as was done for long time at UAA, but letting them actually decide what to protect is a terrible idea. Beeblebrox (talk) 19:26, 16 December 2013 (UTC)

When should bots be permitted to remove BLPs protection?[edit]

This presumes bots are permitted to protect BLPs, and would only apply to bot-protected articles.

The Mediawiki software thinks that anything with fewer than 30 watchers is so insufficiently watched that it's not going to reveal how many are actually watching it.

Never (only humans should unprotect articles)[edit]
1 watcher[edit]
5 watchers[edit]
30 watchers (per Mediawiki)[edit]
If the Pending Changes backlog is too large[edit]
Other criteria[edit]

In response to the situation Hut describes up above: while it's quite unlikely, imagine that someone's vandalised a thousand articles and wrongly gotten the bot to protect all of them. Unprotection will be a big pain, unless we can get the bot to do it. I'd like to see an instructions page with the following features:

  • Place a list of links to protected articles on the instructions page, and the bot will remove protection from the articles
  • Bot's edit summary should be something like "Removing protection per admin instructions" with the last two words being a link to the instructions page
  • Instructions page is fully protected, preventing vandalism in all but the most unlikely situations
  • In short, this instructions page is patterned after Wikipedia:Categories for discussion/Working

If we find a bunch of articles that have been wrongly protected, copy/pasting the list of articles from Special:Contributions will be quick and easy. Nyttend (talk) 22:37, 30 November 2013 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Other comments[edit]

Any additional comments:

The question to ask here is: how do you know that the contents you are trying to protect isn't already false/defamatory? Why do you presume that future edits are more likely to be worse than past edits to such biographies? Someone not using his real name (talk) 22:13, 4 December 2013 (UTC)

A bot wouldn't; presumably an admin would. In the bot case, further editing to bring it away from a false/defamatory state isn't hindered, but publishing those corrections would be delayed until review. Josh Parris 09:38, 5 December 2013 (UTC)

Questions for someone with a bot or toolserver access[edit]

  • How many poorly watched BLPs are there?
  • How many edits per day do the affected articles draw total?
  • What percentage of the affected articles are in an athletics related category?

I feel like this RfC wasn't fully baked. We need some idea of the scope of the change. I suggest withdrawing the advertising for this RfC until my questions are answered. Gigs (talk) 17:10, 27 November 2013 (UTC)

I'm running the relevant queries now:
Extended content
time sql enwiki_p 'SELECT cl_from, count(*) FROM (SELECT page_namespace, page_title, cl_from FROM categorylinks LEFT JOIN page ON page_id=cl_from WHERE cl_to="Living_people" LIMIT 2000) AS clps LEFT JOIN watchlist ON wl_namespace=page_namespace AND wl_title=page_title GROUP BY cl_from;' > watchers
time sql enwiki_p 'select cl_from, count(*) from (select cl_from from categorylinks where cl_to="Living_people" limit 50000) as clps left join revision on rev_page=cl_from where rev_timestamp>"20130000000000" group by cl_from' > recentrevs;

I should have a sample size of ~50000 entries (the 50000 that were first to be added to the category, so it's a biased sample) later tonight and I'll try to make some pretty graphs.

The data files are/will be available on the toolserver in my home directory /home/valhallasw/watchers and /home/valhallasw/recentrevs; I cannot publish them as this would expose information on which pages are not watched. The question on atheletics related categories is more problematic due to the deep category structure. Valhallasw (talk) 20:46, 28 November 2013 (UTC)

Some statistics[edit]

NB. All my axes will be logarithmic!

There are no pages with zero watchers, but a lot of pages with just a few watchers -- 10% of the sample had a single watcher, another 10% a second, etc:

Living People Watchlist Histogram.svg

A CDF is easier to interpret - it tells you the percentage of pages with less than N watchers:

Living People Watchlist CDF.svg

So roughly 80% of pages have less than 10 watchers, about 50% have less than 5.

In terms of edits, you can do comparable statistics. These are the histogram and CDF for the number of revisions in 2013: Living People Revisions CDF.svgLiving People 2013 Revisions Histogram.svg

Last, but not least, I took a look at a potential relationship between the two. Due to the SQL server returning a different set of pages for my two queries, I only have 113 data points here, but I think it's already clear there's not much of a correlation: Living People Watchers cross revs.svg

I hope this helps a bit. If anyone with toolserver access wants the IPython Notebook to play around with, please send me an email. Valhallasw (talk) 21:28, 28 November 2013 (UTC)

Some back of the envelope stats[edit]

I believe the above data might be significantly biased by being articles that were the first to be added to the category.

The following is a sample of 100 sequential BLPs obtained from "random article".

watchers 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 >20
articles 5 27 24 18 4 6 2 3 1 2 1 1 1 0 0 1 2 0 0 0 0 2*
cumulative % 5 32 56 74 78 84 86 89 90 92 93 94 95 95 95 96 98 98 98 98 98 100

*28 & 81

I didn't record how many were sportspeople, but I'd estimate at least half to two-thirds. The remainder were mainly politicians and actors. Espresso Addict (talk) 02:53, 29 November 2013 (UTC)

In your very small, but more random, sample roughly 85% of pages have less than 6 watchers, about 55% have less than 3. Josh Parris 08:44, 29 November 2013 (UTC)
I've added a cumulative total/%age for clarity. Espresso Addict (talk) 14:40, 29 November 2013 (UTC)
Thanks to both you and Valhallasw for the statistics. So it looks like this RfC would apply to up to 80% of all BLPs. I'm not sure that's a tenable proposition. As Valhallasw noted it is very difficult to get a conclusive count of athletic BLPs but your rough estimate of half to 2/3 doesn't sound very far off. Maybe a larger statistical study could give us some good error bars on athletic percentage without needing to try to navigate the rats nest of athlete categories. Gigs (talk) 22:29, 3 December 2013 (UTC)

The above discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Discussion summary[edit]

There is a consensus against using Pending Changes to pre-emptively protecting BLPs that are under- or un-watched. Some unscientific sampling suggests the vast bulk of our BLPs have five (or perhaps three) or less watchers. The rate at which BLPs get edited and the number of watchers is uncorrelated.

Objections raised included

  • the overwhelming of Special:PendingChanges with hundreds of thousands of articles
  • the necessary secrecy of the selection of articles
  • editors will miss the immediacy of editing an unprotected article, and some new editors will not edit further because of this
  • articles should only be protected once a demonstrated need exists
  • during the delay caused by reviewing, IP editors will not see subtle vandalism that only they can spot and fix
  • the effort involved in reviewing BLP edits is better spent improving other articles
  • there's nothing special about BLPs
  • having Reviewers approve edits is elitist
  • Pending Changes protection is inferior to semi-protection because semi-protection is so difficult only the most determined editor will propose an edit
  • active editors should watch BLPs instead of Pending Changes protecting them
  • doing this to unwatched BLPs is the thin end of the wedge before doing it to everything

Objections that seemed to be based on a misunderstanding of how Pending Changes works included

  • edits to unwatched articles will never be approved
  • IPs should not be blocked from editing

Furthermore, if bots were to PC protect articles, they would have to guard against spoofing.

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.