Wikipedia talk:New pages patrol/Reviewers/Archive 9

Can someone please explain what causes an article to be indexed, so that it is searchable via Google, before the article has been marked as reviewed? At this point, only an administrator will be able to try to answer this question. The article in question was Sizwe Faith Sithole. Here is the issue. First, User:Haleyquinn688 created the article, which was an unsourced and promotional BLP. Then User:Bluedits nominated it for BLPPROD. I then nominated it for A7. User:JozjuK then removed the deletion tags and added some questionable references. (The only problem with the removal of the speedy tag is that the removing editor was a sockpuppet of the creating editor.) I then did a Google search, and found nothing about the subject except the usual vanity hits AND the Wikipedia article in question. I then nominated the article for deletion, and applied tags to the article. I also reported sockpuppetry. The problem is that, then, when I viewed the logs for the article, it showed that I had reviewed the article. How had the article been indexed by Google before I reviewed it? Administrator User:Mz7 then blocked the socks and deleted the article as G11. Deleting the article is what I was trying to do; one of my primary objectives as a New Page Reviewer is the identification and deletion of crud, and the article was crud, and it doesn’t matter whether crud is deleted as G11 promotional crud, or as no credible claim of significance A7 crud, or non-notable crud after AFD. Why was the article indexed? Robert McClenon (talk) 02:51, 9 June 2017 (UTC)

Since one of the reasons for New Page Review is to prevent the use of Wikipedia for Search Engine Optimization by blatant promoters, this is troublesome as it seems to show that the system was gamed. Robert McClenon (talk) 02:51, 9 June 2017 (UTC)

Kaldari: could these have been past 30 days before you implemented the 90 day NOINDEX? TonyBallioni (talk) 03:02, 9 June 2017 (UTC)

Adding tags to an article using Page Curation automatically makes an article as reviewed, and adding tags/nominating for deletion with Twinkle also marks an article as reviewed, unless the "mark article as patrolled" checkbox is deliberately unchecked. Is it possible that somewhere along the line, between the BLPPROD tagging, A7 tagging, cleanup tagging, and AFD tagging, the article was inadvertently marked as reviewed and therefore indexed? Mz7 (talk) 05:57, 9 June 2017 (UTC)

The only review event in the log shown was mine, and I tagged immediately for cleanup immediately after I tagged it for AFD, but, when I tagged it for AFD, it was already showing in Google. That shouldn't have happened unless some other reviewer marked it as reviewed (e.g., by tagging it) and didn't mark it as unreviewed. Did the BLPPROD event mark it as reviewed, so that when the BLPPROD was deleted by the sock, the article was left in a state ready to be indexed? If so, we have a hole that can be driven through. Robert McClenon (talk) 10:09, 9 June 2017 (UTC)

@TonyBallioni and Robert McClenon: When was Sizwe Faith Sithole created? It's possible either it aged out after 30 days (the previous expiration threshold before it was recently changed to 90 days). It doesn't look like adding the BLPPROD template marked it as reviewed as the first logged event is Robert McClenon reviewing it on June 7, at which point it was already showing up in Google according to Robert. For a thorough explanation of NOINDEXing, see Wikipedia:Controlling search engine indexing#Indexing of articles ("mainspace"). Kaldari (talk) 19:07, 9 June 2017 (UTC)

And FWIW, it usually takes Google at least a day to index an article after the NOINDEX has been removed, but the actual amount of time it takes is pretty random. Kaldari (talk) 19:10, 9 June 2017 (UTC)

It was created at 03:35, 7 June 2017, tagged BLPPROD with Twinkle by User:Blue Edits at 3:36, and tagged A7 with Twinkle by Robert at 3:37. A fourth user removed both deletion tags at 7:33 and, after several more edits, the article was brought to AFD at 21:28. The patrol log shows a revision was marked patrolled at 3:37, when A7 was added; presumably that was enough to make the article indexable once the deletion tags were removed. (It's pretty irritating that patrolling an article like that doesn't show up by default at Special:Log, but doing it through page curation (the 21:32 patrol log shown in Kaldari's link) does; and nearly as irritating that the 21:32 log is worded so as to be easily confusable with the review log, which is something else entirely.) —Cryptic 20:23, 9 June 2017 (UTC)

In that case, the article could have been indexed any time after 7:33 when the A7 tag was removed. Yes, this is a loophole, but it's also a necessary one. If someone legitimately realized that an article was improperly tagged for speedy deletion, we wouldn't want to prevent it from being indexed (until the 90 days had expired). The purpose of the NOINDEX grace period isn't to make sure that no bad articles are indexed (thousands of vandalized articles are indexed by Google every day). It's purpose is to give editors a chance to identify and delete a large percentage of such articles before they are indexed. Typically the system works well, but it isn't foolproof. If we fixed this loophole, it would then be possible for trolls to prevent legitimate new articles from being indexed by temporarily tagging them with a speedy deletion tag even if they had already been reviewed. Kaldari (talk) 22:13, 9 June 2017 (UTC)

(edit conflict)All reasons why Page Curation should be the default GUI for patrolling new pages, and why tagging specifically for deletion should be limited to authorised reviewers (this BLPPROD was a classic example of very poor patrolling - the article was little more than a blatant hoax) . Three accounts that edited that article have been blocked (the creator and their socks ), so this would be a good reason to introduce measures to prevent CSD templates being removed by non confirmed, or even non qualified reviwers. By the same token, it may also be worthwhile considering restricting the removal of PRODs to autoconfirmed users. Kudpung กุดผึ้ง (talk) 22:31, 9 June 2017 (UTC)

Kaldari, one of the main reasons for NO_INDEX is to dissuade spammers, SEO 'experts', and paid editors from creating articles. Trolls should not have access to deletion tagging systems - this can easily be fixed. Kudpung กุดผึ้ง (talk) 22:37, 9 June 2017 (UTC)

Thank you. That answers that. There is a loophole, and a loophole that can be and was and will be deliberately used, in that the removal of the deletion tag apparently made the article available for indexing, without the review of a qualified reviewer. (More precisely, one qualified reviewer had reviewed it, and had tagged it for deletion.) I respectfully disagree that this illustrates that tagging for deletion should be limited to authorized reviewers, because I disagree that the BLPPROD was poor patrolling. It was an entirely proper BLPPROD; the fact that the article was also a candidate for A7 or G11 doesn't make the BLPPROD improper. I agree that it would be appropriate to limit the removal of CSD, PROD, and BLPPROD tags to autoconfirmed editors, or more restrictively to reviewers. The addition of the BLPPROD was not poor patrolling, but the removal of the BLPPROD and of the CSD tag were the abuse, and restricting them would prevent that abuse. Robert McClenon (talk) 00:42, 10 June 2017 (UTC)

In this case, as we have noted, the abuse involved another infraction that is already dealt with, usually quickly and effectively, and that is sockpuppetry. However, the socks shouldn't be able to game the system in the first place before they are laundered (isn't that what one does with socks)? Robert McClenon (talk) 00:47, 10 June 2017 (UTC)

Robert, I looked at the deleted page before I made my comment above. It was indeed an obvious hoax and even practically vandalism. While BLPPROD was technically correct in that the page was about a person and was unsourced, CSD trumps PROD/BLPPROD any time. Skill is often required in order to recognise subtle forms of hoax, advertising, or attack, but in this case it really was blatant - trust me. Kudpung กุดผึ้ง (talk) 00:56, 10 June 2017 (UTC)

I'll take your word that it was a hoax. That wasn't obvious to me, and G3 is meant for obvious hoaxes. It was obviously A7, and I am not sure why I didn't tag it G11 also.

Yes, G3 was better than G11 was better than A7 was better than BLPPROD here, but that's orthogonal to the original question raised - why this page got indexed and how to prevent that in the future - and I submit that that question is the more important. The problem is that the action of patrolling a page is being taken to mean two distinct things: 1) the page doesn't need further review from NPP, and 2) the page is ok to be indexed. The only effective way I can think of to solve that is to train NPP not to mark pages patrolled merely because they're tagged for deletion and make Twinkle not default to doing that. There's really no benefit from marking a page patrolled at that point, since it shows up differently in Special:Newpagesfeed anyway. —Cryptic 01:41, 10 June 2017 (UTC)

I agree that the original question, which is why the page was able to get indexed, is the real question, and that a page should not be considered reviewed simply because it was tagged for deletion and the deletion was then reversed. I would have still taken it to AFD if it hadn't been indexed, but it shouldn't have gotten indexed, and it illustrates a loophole that can be used deliberately. Robert McClenon (talk) 02:38, 10 June 2017 (UTC)

It wasn't considered reviewed because the deletion tag was reverted. It was considered reviewed because Twinkle marked it reviewed when you tagged it an A7. (I understand you can change that at WP:TWPREFS.) Had you tagged the article an A7 without patrolling it, and it deletion tag was removed, it wouldn't have become indexable. —Cryptic 03:08, 10 June 2017 (UTC)

I wasn't using Page Curation to mark the page for deletion. I don't use Page Curation to mark for deletion, which I do directly from Twinkle, but it appears that Twinkle is also marking the page reviewed when I try to have it deleted. Here is another example, this time a more reasonable article. See Lew Gaiter. Its author originally created it as an infobox-only placeholder. I detest infobox-only placeholders, but that is a secondary point. I PROD'd it. The author then removed the PROD and expanded it to an article about a politician. One can argue whether he is notable at this time, but it wasn't a hoax. A Google search turned up the usual vanity hits, and the Wikipedia article, when the only action other than by the author was my PROD. Obviously, tagging an article for deletion and then untagging it causes it to be indexed. Maybe I should always do a Mark as Unreviewed after tagging crud. Robert McClenon (talk) 03:29, 13 June 2017 (UTC)

Robert McClenon, in the Twinkle Preferences you should probably uncheck the various "mark as patrolled" boxes if you're concerned about TW patrolling even when you don't want to. Primefac (talk) 12:14, 13 June 2017 (UTC)

Cryptic, from my understanding page curation marks all pages tagged for deletion as reviewed and there isn't an option to turn it off without clicking unreviewed. TonyBallioni (talk) 03:11, 10 June 2017 (UTC)

This particular article wasn't reviewed by page curation (until much later, at 21:32). And if what you say is true, it renders page curation unfit for the purpose of tagging pages for deletion. —Cryptic 03:14, 10 June 2017 (UTC)

I'll use the most recent page I've tagged for speedy deletion that is still outstanding as an example [1]. I can't find the documentation now, but if I recall correctly, all deletion templates are supposed to add _NOINDEX_ so the marking as reviewed doesn't matter. The loophole is that if the tag is removed, it can be indexed. TonyBallioni (talk) 03:21, 10 June 2017 (UTC)

And that's exactly what this entire section is about. You've just changed that article from one that can only become indexed if 90 days pass or someone with page reviewer rights takes action on it, into one that anybody who's not blocked can make indexed. It's not the article author who hypothetically removes your db-spam tag who's at fault here, nor even (to a lesser extent) the admin who inexplicably declines to delete it. The person who made that page indexable was you. The workaround for the auto-review misfeature in page curation is to do what User:Adam9007 did before you and mark it unreviewed. The workaround for the similar but less-severe misfeature in Twinkle is (I'm told) to disable it in preferences. —Cryptic 03:31, 10 June 2017 (UTC)

From what Kaldari said above, the mark as reviewed feature in page curation seems an intentional inclusion, that the NOINDEX magic word of the deletion tags is supposed to prevent becoming an issue if the tag is placed correctly. In the case linked to, Adam appears to have unreviewed it because another reviewer marked as reviewed a foreign language page without doing anything. Having looked at a fair amount of review logs, from what I can tell, Robert is one of the few reviewers who uses page curation that also marks pages as unreviewed after tagging for issues or for deletion. If that should be the norm, it should be added to the guide at WP:NPP, and if the community wants it changed to not mark as reviewed, then a ticket should be added to phab. TonyBallioni (talk) 03:52, 10 June 2017 (UTC)

That sounds like a bad solution to me. You're basically arguing that an article shouldn't be indexed until all issues with it have been resolved. That seems completely antithetical to the DNA of Wikipedia. Do we really want to turn Wikipedia back into Nupedia (which was a total failure)? There is already a simple solution for this (uncommon) situation, which is to restore the speedy deletion tag. If that isn't adequate, we should build an abuse filter that prevents non-autoconfirmed editors from removing speedy deletion tags from articles. Not marking articles as reviewed is harmful to the encyclopedia. It prevents people from being able to find new articles (that may be important in the case of current events). I don't understand why New Page Patrollers believe that new pages should be held to a higher standard than all the other articles on Wikipedia. It is entirely possible for spammers and vandals to abuse any of the 5 million articles on Wikipedia (and get their abuse indexed by Google). That is the nature of Wikipedia being an open platform. We deal with this by having a huge volunteer pool of editors who eventually fix most of the problems. It will never be a foolproof system, however. If you want a foolproof system that no one can possibly abuse, you're going to end up with a pool of editors the size of Citizendium's. New Page Patrolling was supposed to just be a way to weed out most of the garbage, not a mechanism to put Wikipedia under the control of a small group of gatekeepers. Let's not forget the importance and benefits of keeping Wikipedia open and dynamic (and even a little chaotic). Kaldari (talk) 17:36, 10 June 2017 (UTC)

No, that isn't the argument, but it appears that you, User:Kaldari, are saying that the ability of Wikipedia to be manipulated by spammers and get their abuse indexed by Google is a feature, rather than a misfeature, and that there should be no effort to deal with that. If so, I disagree with Kaldari and agree with what I think User:TonyBallioni and User:Adam9007 are saying. The problem is that a page is being indexed because a reviewer tried to prevent its indexing and to prevent its inclusion in Wikipedia. That is not a feature; it is at best a misfeature, and I think it is a bug. Robert McClenon (talk) 03:16, 13 June 2017 (UTC)

To clarify what I was saying: I don't see many reviewers actually mark it as unreviewed after tagging for deletion or issues. Robert, you are the only one of the few I've noticed do this, and I look at a fair amount of review logs. That's neither a good or bad thing, just an observation. My point was less arguing for either your view or Kaldari's view, but saying that if the community preference is for deleted pages to not be marked as reviewed, that should be added to the instructions at WP:NPP and a ticket placed in phab to try to fix it. TonyBallioni (talk) 03:23, 13 June 2017 (UTC)

Creating an abuse filter to prevent non-autoconfirmed users from removing speedy deletion templates would be a much better solution as it wouldn't make the backlog bigger and it would also insure that speedy deletion tags aren't removed by vandals and spammers (for all articles, not just new ones). Kaldari (talk) 03:43, 13 June 2017 (UTC)

Excellent idea and also a filter that prevents someone removing the tag from articles they have created. If it is forbidden to do it it shouldn't be possible to do it either. Domdeparis (talk) 14:35, 13 June 2017 (UTC)

I'm trying to wrap my head around the issue and seeking some clarification. Technical question first: As far as indexing is concerned, it's perhaps worth noting that articles older than 90 days are automatically indexed and only pages less than 30 days observe the NOINDEX according to the documentation at Template:NOINDEX. So even if I nominate an article for deletion, that only prevents it from being indexed for 30 days. If I were to unreview a 30 day old article, after I marked it for deletion, the magic word would be ignored. Is that everyone else's understanding of indexing works too or did I miss something?

Process related question: I don't understand why we have articles that are nominated for deletion but marked unreviewed. What more reviewing is there to be done? When I nominate an article for deletion with the Page Curation tool, I have done all the reviewing that I need to do; I don't need to then also tag it for all the ways in which it can be improved. I have it on my watchlist, and if the deletion is removed I would notice and can take appropriate action. To those you you who advocate unreviewing new articles that have been nominated for deletion, what purpose does keeping such articles in the queue serve? Mduvekot (talk) 14:51, 13 June 2017 (UTC)

User:Mduvekot - I agree that articles that are marked for AFD should be marked as reviewed. That is all the reviewing they need. They will either be deleted or kept, and, if they are kept, the AFD is a better review than just being okayed by one reviewer. Articles that are PROD'd or tagged for CSD should not be indexed. Anyone can remove a PROD, and that should not mean that the article was reviewed, only that it won't be deleted (although an argument can be made that removing the PROD is an okay to keeping it and indexing it). Anyone or anything can remove a CSD tag, and that doesn't mean that the article isn't crud, only that it isn't speediable crud. Robert McClenon (talk) 15:07, 16 June 2017 (UTC)

A comment was made, with which I agree, that a filter is a good idea to prevent the removal of a CSD or PROD tag by non-auto-confirmed editors, and I think that the suggestion was to prevent the removal of a CSD tag by the author. The removal of a CSD tag by an IP address is usually sockpuppetry, but just preventing its removal by non-auto-confirmed editors is good for that purpose. (For that matter, a filter to prevent the removal of an AFD tag is a good idea. An AFD tag should only be removed by the closing admin, or by the speedy-closing admin in cases of abusive AFD, etc.) Robert McClenon (talk) 15:13, 16 June 2017 (UTC)

I always patrol Unreviewed pages. If someone else tagged it for deletion I don't need to waste my time on it. Legacypac (talk) 08:11, 24 June 2017 (UTC)

I found Mount Gargash originally at Talk:Mount Gargash and moved it. The problem is, 1) it was created by an IP editor, and 2) I have Autopatrolled so apparently moving it "reviewed" the article. Can one of the regular WP:NPP editors please have a look at it, and either unpatrol or tag/CSD appropriately. Thanks! --IJBall (contribs • talk) 01:35, 22 June 2017 (UTC)

  Done TonyBallioni (talk) 01:38, 22 June 2017 (UTC)

@IJBall and TonyBallioni:, Mount Gargash is now typical of the tens of thousands of short articles that might have some potential but which will forever remain perma-stubs. No maintenace tags, or stub tag, or cats, that might attract the attention of someone, and an IP user who doesn't care less. And now (correctly) PRODed by Chris troutman and unless someone comes across a good reason for removing te PROD tag, it will be lost forever, and frankly, although I am not a deletionist per se, if I were cleaning out the expired PROD list like some admins do, it would certainly be lost forever.

It's a shame, because it's the kind of article with which the creator probably genuinely though they were adding something good to the encyclopedia, once more demonstrating what a lousy job the Foundation (except for Jorm) has made all these years of providing help for new uses and is still refusing to do. FYI: Robert McClenon, Atsme, DannyH (WMF). Kudpung กุดผึ้ง (talk) 02:47, 22 June 2017 (UTC)

The sentiment that new editors think "they were adding something good to the encyclopedia" is probably true and speaks to the how un-informed the public is. It's our job to chase off maybe 80% of all new users if we can't educate them to what Wikipedia needs and I wish those useless clods in San Francisco would put more effort into outreach. Someone else could have tag-bombed that article but I'm happier to cut to the chase. My thinking is, every stub article produced by a drive-by editor we allow robs a committed editor of a Four Award. Chris Troutman (talk) 02:57, 22 June 2017 (UTC)

We get a lot of these types of articles near the end of the backlog. Chris definitely cut to the chase, faster than I did obviously, and it was justified. V is a core policy that unfortunately for many similar articles on villages, mountains, and buildings we have no way of meeting. TonyBallioni (talk) 03:02, 22 June 2017 (UTC)

Yuck. By the way, I thought that IP editors couldn't create articles. Did this happen because it was created (in stupid good-faith error) in talk page space? Actually, it probably should have an article, but I don't want to load the encyclopedia with perma-stubs. I am a deletionist, except with respect to things that really exist in the outside world. Of course, geographic places do exist in the outside world, but this doesn't seem to be documented. What a mess. Yes, apparently, doing anything to a crud page marks it as reviewed. Robert McClenon (talk) 03:05, 22 June 2017 (UTC)

Yes, creating articles in talk space is a work around to IP restrictions on creating articles. I've mainly seen it for people trying to sneak spam in. TonyBallioni (talk) 03:07, 22 June 2017 (UTC)

Then the reviewer who moved it to article space is in a role similar to an AFC reviewer moving a draft into article space, and is taking responsibility for the article that has no sources. Hmmm. There is no obligation by a new page reviewer to move a page created by an IP to where the IP couldn't have created it. Oh well. Robert McClenon (talk) 03:46, 22 June 2017 (UTC)

• Good article. Notable mountain. One of Iran's highest, and a planned astronomy site of international significance. --SmokeyJoe (talk) 04:29, 22 June 2017 (UTC)

I don't think deleting permastubs is going to achieve the goal of helping editors focus on improving content. It's going to inflame deletion/inclusion tensions and waste time in WP:PROD and WP:AFD. What's the problem with letting these sleeping dogs lie? ~Kvng (talk) 16:18, 24 June 2017 (UTC)