Wikipedia:Edit filter/Requested

From Wikipedia, the free encyclopedia
    Requested edit filters

    This page can be used to request edit filters, or changes to existing filters. Edit filters are primarily used to address common patterns of harmful editing.

    Private filters should not be discussed in detail. If you wish to discuss creating an LTA filter, or changing an existing one, please instead email details to wikipedia-en-editfilters@lists.wikimedia.org.

    Otherwise, please add a new section at the bottom using the following format:

    == Brief description of filter ==
    *'''Task''': What is the filter supposed to do? To what pages and editors does it apply?
    *'''Reason''': Why is the filter needed?
    *'''Diffs''': Diffs of sample edits/cases. If the diffs are revdelled, consider emailing their contents to the mailing list.
    ~~~~
    

    Please note the following:

    • Edit filters are used primarily to prevent abuse. Contributors are not expected to have read all 200+ policies, guidelines and style pages before editing. Trivial formatting mistakes and edits that at first glance look fine but go against some obscure style guideline or arbitration ruling are not suitable candidates for an edit filter.
    • Filters are applied to all edits. Problematic changes that apply to a single page are likely not suitable for an edit filter. Page protection may be more appropriate in such cases.
    • Non-essential tasks or those that require access to complex criteria, especially information that the filter does not have access to, may be more appropriate for a bot task or external software.
    • To prevent the creation of pages with certain names, the title blacklist is usually a better way to handle the problem - see MediaWiki talk:Titleblacklist for details.
    • To prevent the addition of problematic external links, please make your request at the spam blacklist.
    • To prevent the registration of accounts with certain names, please make your request at the global title blacklist.
    • To prevent the registration of accounts with certain email addresses, please make your request at the email blacklist.



    Vandalism to meme pages[edit]

    bad_desc := "(cringe|worst|best)"; any_meme := "(skibd|skidibi|skibid|rizz|bozo|\(meme\))"; meme_cat := "(meme\}\}|fads\]\]|trends\]\]|slang\]\])"; !"confirmed" in user_groups & page_title irlike any_meme & ( rcount(meme, added_lines) / 3 > rcount("\.", added_lines) + 0.5 | /* prevent excessive use of the meme */ rcount(bad_desc,added_lines) > 3 | /* prevent defamation of the meme */ (removed_lines irlike meme_cat & !old_wikitext irlike meme_cat) /* prevent decategorization from meme categories */ )

    • Actions: Disallow

    Faster than Thunder (talk | contributions) 01:06, 6 May 2024 (UTC)[reply]

    Looks like the third filter log entry should be disallowed by filter 1233 (hist · log) but wasn't caught, the second log entry looks like your everyday run-of-the-mill disruption, and the first log entry is likely low-effort disruption that may want to be prevented by some filters. Maybe we could set 1163 (hist · log) to warn+tag or disallow.
    By the way,  !( (removed_lines + page_title) irlike abuseStr) basically means that Skibidi Toilet additions are excluded from said article describing this meme itself. Codename Noreste 🤔 La Suma 02:13, 6 May 2024 (UTC)[reply]
    Also, please note that before disallowing, we always test filters on log or tag before to minimize the possibility of a huge amount of false positives. If this is made into a seperate filter then, I highly doubt it will be set to disallow immediately. – PharyngealImplosive7 (talk) 13:45, 6 May 2024 (UTC)[reply]
    I hereby retract saying to set 1163 to disallow after seeing your comment, but couldn’t we at least set this to warn with the tag? Codename Noreste 🤔 La Suma 14:58, 6 May 2024 (UTC)[reply]
    Yes, but I believe that usually, filters are first set to log or tag just to see if they work well or not, as even warning could be problematic if the filter has too many FPs. – PharyngealImplosive7 (talk) 15:01, 6 May 2024 (UTC)[reply]
    Can someone start this filter with no actions enabled first please? Faster than Thunder (talk | contributions) 20:43, 8 May 2024 (UTC)[reply]
    Can you break down what each part of that filter is trying to do? It doesn't make sense to me. Suffusion of Yellow (talk) 21:06, 8 May 2024 (UTC)[reply]
    Done. Faster than Thunder (talk | contributions) 22:45, 8 May 2024 (UTC)[reply]

    Done. Faster than Thunder (talk | contributions) 21:22, 8 May 2024 (UTC)

    Alright. So we have:
    meme := "(?i)(" + str_replace(page_title," ","|") + ")";
    and
    length(meme) * 2 < rcount(meme,added_lines) | // prevent excessive use of the meme
    First, you're generating meme by splitting apart the title. That's clever, but what about a title like "Bozo the Clown"? One of your words is going to be "the". Second, rcount() counts the total number of matches, not the total length of the matches put together. If you want to prevent excessive use of a word, say something more like:
    rcount(meme, added_lines) - rcount(meme, removed_lines) > 2
    But I don't that's a good idea. It's natural for the title of the article to be repeated many times throughout the page.
    Now we have:
    get_matches(bad_desc,added_lines) > 3 | // prevent defamation of the meme
    But get_matches() returns a fixed-size array. I'm not sure what the "3" is supposed to mean.
    And finally:
    (removed_lines irlike meme_cat & !old_wikitext irlike meme_cat) // prevent decategorization from meme categories
    This won't match anything, but could be fixed by using added_lines instead of old_wikitext. But we already have 132 (hist · log) for category removal.
    Thanks for this, but I think it's just inevitable that "meme pages" are going to end up semi-protected, at least temporarily. There are just too many creative ways to vandalize. Suffusion of Yellow (talk) 21:24, 9 May 2024 (UTC)[reply]
    Maybe the 3 in get_matches(bad_desc,added_lines) > 3 | // prevent defamation of the meme is supposed to be compared to the array length so maybe @Faster than Thunder really just meant length(get_matches(bad_desc,added_lines)) > 3. I also do sadly agree that vandalism to meme pages is bound to happen, and we'll probably need to protect them at some point. – PharyngealImplosive7 (talk) 22:34, 9 May 2024 (UTC)[reply]
    @Faster than Thunder: Also, if we have a bad_desc variable to prevent defamation, wouldn't another issue be to say that the meme is the "best"? So would it also be a good idea to create a separate variable to prevent additions like that? – PharyngealImplosive7 (talk) 02:57, 10 May 2024 (UTC)[reply]
    I implemented your suggestions. Faster than Thunder (talk | contributions) 17:40, 14 May 2024 (UTC)[reply]
    Could we also block "skibidi toilet", "skibidi", and such as per the thread below? I don't know how the filters work. Myrealnamm (💬talk · ✏️contribs) at 19:17, 14 May 2024 (UTC)[reply]
    They are already added because of "skibid." Faster than Thunder (talk | contributions) 19:34, 16 May 2024 (UTC)[reply]
    See this, which an IP vandalized using "skibidi toilet" as the edit summary. This should be added @Faster than Thunder. Myrealnamm (💬talk · ✏️contribs) at 23:37, 23 May 2024 (UTC)[reply]
    That is the wrong filter to request it in, because that was in another article, but this could be added to a new filter idea or something like 614 (hist · log). – PharyngealImplosive7 (talk) 00:32, 24 May 2024 (UTC)[reply]
    Very funny. XD Faster than Thunder (talk | contributions) 20:22, 24 May 2024 (UTC)[reply]
    Now that some improvements have been made to the filter idea, what new changes need to be made to the filter before it can be created? Faster than Thunder (talk | contributions) 17:39, 22 May 2024 (UTC)[reply]
    I share the same concerns as SoY. Vandals on meme pages are going to come up with new ways faster than a filter can catch them, and it's far more efficient to just protect the small number of "meme-type" pages than to try and craft a filter that has every single variation and type of petty vandalism out there. It's possible for general vandalism filters, because the terms in those are spread throughout the encyclopedia, but for specific pages, it's going to just end up with vandals getting around the filter on purpose. EggRoll97 (talk) 22:29, 26 May 2024 (UTC)[reply]
    @EggRoll97 That's very true. However, these vandals seem to vandalize with "meme words" on all the pages, so hmmmm. Myrealnamm (💬talk · ✏️contribs) at 00:22, 27 May 2024 (UTC)[reply]
    Yeah. I think the best thing to do is to just semi-protect the meme pages, instead of creating and constantly changing a filter that won't catch all the vandalism sadly. – PharyngealImplosive7 (talk) 04:45, 27 May 2024 (UTC)[reply]

    New Filter Request - "Skibidi Toilet"[edit]

    Myrealnamm (💬talk · ✏️contribs) at 20:47, 9 May 2024 (UTC)[reply]

    Oh, the previous request is similar. Myrealnamm (💬talk · ✏️contribs) at 20:47, 9 May 2024 (UTC)[reply]
    That could be accomplished by expanding 614 (hist · log) to article talk pages. Here's a quick test from last month. Suffusion of Yellow (talk) 21:05, 9 May 2024 (UTC)[reply]
    @Suffusion of Yellow Example: https://cuddlysnowman.com/w/index.php?title=Stuart_Gibbs&curid=54184329&diff=1223521681&oldid=1223051823 Myrealnamm (💬talk · ✏️contribs) at 18:13, 12 May 2024 (UTC)[reply]
    They didn't use "skibidi", they used "skbidi". Maybe all forms of "skibidi", including typos such as "skbidi" and "skibid" should be added to the filter as well. Myrealnamm (💬talk · ✏️contribs) at 18:14, 12 May 2024 (UTC)[reply]

    Disallow changing result parameter on Infobox military conflict by IPs/new users[edit]

    Per this discussion (pinging @GreenC):

    • Task – In the |result= parameter of {{Infobox military conflict}}, disallow edits between sides of "X victory", in addition to edits away from or between "X victory", "Inconclusive", and "See (article section)" by IP addresses or very new users.
    • Reason – Widespread tendentious editing by those unfamiliar with site guidelines, at a bare minimum with MOS:MILHIST. After parameter is in accordance with said guideline, it almost never needs to be changed.
    • Diffs:

    Remsense 01:17, 11 May 2024 (UTC)[reply]

    This is a hard one, because there could be so many false positives, like if someone corrects a typo in the result parameter and gets a disallow message. I would suggest something like tag or warn at most unless someone can find a non FP-prone way of filtering these types of edits, but this should definitely be a log-only filter at first. The regex should also probably be similar to something like 391 (hist · log). – PharyngealImplosive7 (talk) 02:15, 11 May 2024 (UTC)[reply]
    Agreed that initial caution is required, but unfortunately I don't see a warning saying "changes require reliable sources" being effective in the end? Remsense 02:21, 11 May 2024 (UTC)[reply]
    Totally agreed. But first we should make the filter ready to be disallowed by minimizing the amount of FPs as much as possible. – PharyngealImplosive7 (talk) 03:15, 11 May 2024 (UTC)[reply]
    @Remsense: As to the prospect of disallow, I'm going to say  Not done. The top of this page even states, Edit filters are used primarily to prevent abuse. Contributors are not expected to have read all 200+ policies, guidelines and style pages before editing, and I don't think it's a far stretch to assume that edits are all in bad faith. Even in the diffs provided, the edit to Fourth Crusade seems misguided and wrong, but not necessarily in bad faith. Any filter that catches this would end up with a non-zero amount of false positives. I'm not against a log or maybe a tag filter, though. I'll see if I can work one up, but if anyone wants to have a shot at trying the code in the meantime feel free. EggRoll97 (talk) 04:04, 24 May 2024 (UTC)[reply]

    Creating mainspace articles which begin with your username[edit]

    • Task: (This is my first post here, so please let me know how I borked it up.)

      I think a filter which logs (and eventually warns?) people who attempt to create an article which begins with your username would be beneficial. I have seen multiple people who create (e.g.) HouseBlaster/sandbox as opposed to User:HouseBlaster/sandbox (and I have personally done something similar).

      It also might catch people who try to write autobiographies and people whose usernames violate WP:CORPNAME, both of which seem like positive side-effects.

    • Reason: Self-explanatory
    • Diffs: They are all deleted fairly quickly as WP:G6 (if it is a benign mistake), and I don't have any evidence that the autobiography/CORPNAME thing is a problem (I just think that it is a something else which this filter would happen to catch).

    Thanks, HouseBlaster (talk · he/him) — Preceding undated comment added 13:49, 17 May 2024 (UTC)<diff>[reply]

    Something like the following, maybe? '''[[User:CanonNi]]''' (talkcontribs) 13:54, 17 May 2024 (UTC)[reply]
    Don't we already have Special:AbuseFilter/148 or something similar? Codename Noreste 🤔 La Suma 00:51, 18 May 2024 (UTC)[reply]
    My proposal is slightly different, in that it would catch people with more than 100 edits who make a mistake rather than a deliberate attempt to create an autobiography. HouseBlaster (talk · he/him) 01:09, 18 May 2024 (UTC)[reply]
    page_id == 0 &
    (
     page_namespace == 0 &
     (
      page_title rlike user_name | user_name in page_title
     )
    )
    
    Such a filter might make life interesting for the likes of User:F, but generally there seem to be few false positives. Certes (talk) 17:43, 18 May 2024 (UTC)[reply]
    Would it make sense to additionally check that the title/username is longer than x? Not sure which is more efficient. HouseBlaster (talk · he/him) 17:46, 18 May 2024 (UTC)[reply]
    Possibly. We might also need to convert spaces to underscores in user_name before matching to page_title. This query may be of interest. Certes (talk) 18:13, 18 May 2024 (UTC)[reply]

    "Skibidi" username filter[edit]

    I've noticed that new usernames which contain "Skibidi" in them often are used only for disruption/vandalism/trolling. Is there any way we could add a filter which blocks all usernames with "Skibidi" and/or sends them to UAA? If you reply here, please ping me. Thanks — thetechie@enwiki: ~/talk/ $ 02:33, 29 May 2024 (UTC)[reply]

    I'm not sure if creating a filter that prevents Skibidi (toilet) usernames is necessary (after all, it compares every account creation when set to action == "createaccount"); there is User:AmandaNP/UAA/Blacklist in which you can propose adding s+k+[i1bdt]{4,}y*\b on the talk page. Codename Noreste 🤔 La Suma 03:15, 29 May 2024 (UTC)[reply]
    Also note that the regex above would need to be continuosly updated as the filter changes. – PharyngealImplosive7 (talk) 19:47, 29 May 2024 (UTC)[reply]
    Not really; there's no need to catch them all. I don't like disallowing usernames which scream "I am NOTHERE" but aren't so offensive as to require a revdel; those usernames just make the vandalism easier to spot. (Plus the first word to disallow should be "Truth".) And reporting to UAA on account creation isn't really helpful unless the username is block-on-sight. They might wait hours or days to edit, or never edit at all. Now, we could have filter which reports to UAA on the first edit, at which point it's usually clear what the user is up to. But as CN points out, DeltaQuadBot already does that, so why not just add to DQB's list? Suffusion of Yellow (talk) 19:56, 29 May 2024 (UTC)[reply]