Jump to content

Wikipedia:Edit filter/Requested/Archive 3

From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 4Archive 5Archive 10


Filled requests

Portal Vandalism

  • Task: To prevent the blanking and vandalism of portals by new users.
  • Reason: There really is no productive point to blanking, removing sections of a portal, or vandalizing a portal.

- Smallman12q (talk) 21:02, 11 February 2010 (UTC)

Filter 3 already detects this. I believe the reason for keeping it at warn as opposed to disallow is primarily related to WP:BITE, though there isn't actually information in the notes about it. I'm inclined to agree with that. --Shirik (Questions or Comments?) 20:15, 12 February 2010 (UTC)
 Done I missed the "portal" part of this request. I have updated filter 3 to also monitor namespace 100. --Shirik (Questions or Comments?) 20:17, 12 February 2010 (UTC)

Page Blanking

  • Task: The filter would identify edits that remove whole sections of content as "Page Blanking" in the history. This filter would apply to the Tea Party movement and Tea Party protests, 2009 articles.
  • Reason: There's been a lot of what appears to be vandalism recently on these two pages and a lot of time is consumed hunting for unsupported edits, where editors remove a reference, paragraph or entire section without providing a summary or providing a summary that isn't following the WP:FIVE. A simple tag identifying the edit as "Page Blanking" would be a great help.

--Happysomeone (talk) 18:44, 11 February 2010 (UTC)

 Already done by filter 3. --Shirik (Questions or Comments?) 20:16, 12 February 2010 (UTC)


Allcaps "Slut" and "Whore"

  • Task: Triggers a tag when a new user enters text containing either of these words.
  • Reason: Neither of these words, in allcaps, are ever encyclopedic.

- Firsfron of Ronchester 21:40, 24 January 2010 (UTC)

Done Triplestop x3 22:01, 24 January 2010 (UTC)

Thanks! Firsfron of Ronchester 22:33, 24 January 2010 (UTC)

Penis vandalism

  • Task: Disallow "Penis vandalism" by new or unregistered users, specifically the kind where a large block of text containing nothing but the word "Penis" is added.
  • Reason: Does it really need explaining? There is virtually nothing constructive about adding "Penis penis penis penis penis penis penis penis..." to an article.

- The Thing Vandalize me 17:22, 18 January 2010 (UTC)

  • Would is also be possible to include the addition of "P"-related images? An example of this is when a vandal adds an image of a "you-know-what" to an article. I don't know if this already has a filter (or is included in a filter), but this suggestion is closely related to this pal's request! Schfifty3 17:29, 18 January 2010 (UTC)
    The Bad image list usually takes care of that, which prevents images on that list from being displayed inline on pages other than the ones they are specifically allowed on. The Thing Vandalize me 17:42, 18 January 2010 (UTC)
    Created filter 285 which attempts to accomplish this in a slightly more generalized case. Log only for now to see how it does. Note that I'm a little worried about the performance of this filter, so its practicality will depend a bit on how frequent the hit rate is. --Shirik (Questions or Comments?) 07:33, 19 January 2010 (UTC)
    Update: Performance is acceptable on the filter. Leaving it running for just a little longer to verify no false positives are picked up, but so far everything looks good. Zero false positives, 25 hits across 2 days. --Shirik (Questions or Comments?) 05:18, 21 January 2010 (UTC)
     Done May still need some optimization, but the filter is working well and hasn't had a false positive yet. Activated and set to disallow. --Shirik (Questions or Comments?) 23:52, 23 January 2010 (UTC)

New article title in the form of a question

  • Task: Detect if a new user tries to introduce a new article with a title that is written in the form of a question. If such an article creation attempt is detected, warn user that he/she may be attempting to introduce a test page, a personal essay, a how-to guide, or maybe even a piece of original research.
  • Reason: Some new users have apparently been experimenting with new article creation by creating blatantly unnecessary pages that consist solely of questions in the page titles and answers in the article bodies. Furthermore, page titles in the form of questions are possible red flags for personal essays, how-to guides, or pieces of original research.

- SoCalSuperEagle (talk) 00:17, 1 December 2009 (UTC)

Gotcha: Make sure this applies to article namespace only and for users who are not (auto)confirmed. Test should only check for *? (quetion mark at end)   «l| Promethean ™|l»  (talk) 14:59, 6 December 2009 (UTC)
The following is a possible implementation of this filter:
      (article_namespace == 0) & 
      (article_articleid == 0) & 
      ("autoconfirmed" in user_groups) &
      (article_text rlike "\?$")
However I'm not convinced of the need for this. I agree that I've seen things like this myself, but only rarely. If I had to estimate, I'd estimate an average hit rate of 1 per week. I looked at the new pages log (I only went back 2000 pages, as that's as far as it will let me go) and didn't see a single instance it would hit. I'm not sure the execution time can be justified. --Shirik (Questions or Comments?) 02:15, 6 January 2010 (UTC)
In my opinion, the filter should also trigger if the article title starts with an obvious question word such as "how", "when", or "why", even if there's no question mark at the end. While patrolling new pages, I have seen several new article titles that looked like questions, but with the question marks left out. The filter would easily miss such titles if it were to check only for the presence of question marks at the end of them. --SoCalSuperEagle (talk) 18:26, 6 January 2010 (UTC)
Also, in response to the comment that was left by Shirik, when a page is deleted for whatever reason, its respective entry in the new pages log will disappear. --SoCalSuperEagle (talk) 18:37, 6 January 2010 (UTC)
I've filed a toolserver request to get a feel for what kind of hit rate we're talking about here to determine if it's justified. --Shirik (Questions or Comments?) 03:26, 7 February 2010 (UTC)
Data came back, limited to only the article namespace, at approximately 10 deleted pages hitting this condition per day over the last 3 months. Some of these hits seemed to be questions, some were vandalism, and some were "how to" guides. All of this is inappropriate, but I don't think this is something that should be set to disallow, primarily because of WP:BITE. I wouldn't be opposed to setting this up with a warning message that assumes good faith and directs the editor to WP:HOWTO and WP:RD. --Shirik (Questions or Comments?) 03:52, 7 February 2010 (UTC)
That would be fine. --SoCalSuperEagle (talk) 17:59, 7 February 2010 (UTC)
 Done warn-only (with tag on submission) by filter 289 --Shirik (Questions or Comments?) 05:56, 10 February 2010 (UTC)
  • Task: The filter should show a COI warning when a person wikilinks there username in article namespace.
  • Reason: People add themselves to list as with Danilo Garza in this edit. Often times, this creates a COI.

- Smallman12q (talk) 02:11, 30 May 2009 (UTC)

I've implemented a straightforward check for links to any user or user talk page. I can't think of any reason they should appear in article space. One wrinkle though, is that some editors sign internal comments in article space.
The filter is enabled, but takes no action; I'll let it run for awhile that way to test for undesired side effects. —EncMstr (talk) 09:05, 30 May 2009 (UTC)
That's not exactly what I had in mind. I was thinking of something like

(article_namespace == 0) & (lcase(user_name) in lcase(added_lines)) or (article_namespace == 0) & (lcase(added_lines) rlike (("[[" +lcase(user_name) + "]]") | user_name) meaning the username itself is added to the article creating a COI. It probably should also check for removed lines as the user may be trying to vandalize as well.Smallman12q (talk) 16:06, 31 May 2009 (UTC)

I understood that was your intent which is similar to Special:AbuseFilter/148—Users creating autobiographies. I've just made that one a little more general by ignoring capitalization. The new filter (176) has run for most of the weekend and captured 42 edits, most of which should not have been allowed, though it's also triggering on edits near previous insertion of user links. It needs to be tweaked some. Any objection to a more general filter than asked for? —EncMstr (talk) 16:47, 31 May 2009 (UTC)
Nope, no objetion. I'm not an sysop so I can't see 148 so I don't quite know whats in it. One problem I realized is that this filter also catches userfied pages. As in Special:AbuseFilter/examine/log/480969. My question then is,should userfied be pages be linked from article mainspace?Smallman12q (talk) 17:07, 31 May 2009 (UTC)
We could check for length to make it faster. For example:
(article_namespace == 0) 
&((length(added_lines) > length(user_name)) & (((lcase(added_lines) rlike "[[user(talk|):") |   (lcase(user_name) in lcase(added_lines))
| ((length(removed_lines) > length(user_name)) & (lcase(user_name) in lcase(removed_lines)))

The current filter is a inefficient, you could use lcase(added_lines) rlike "[[user(talk|):") Smallman12q (talk) 19:40, 31 May 2009 (UTC)

 Done by filter 176, but this was deleted due to the complexity of exceptions. Contact User:EncMstr if it's desired to complete this filter, or open a new request. --Shirik (Questions or Comments?) 06:14, 10 February 2010 (UTC)

Wikipedia Extended image syntax Vandalism

Edit summary Copyvios

  • Task: Stopping vandals/wallflowers98 socks from adding illegal downloads to media in the form of links in edit summaries, and blocking any that try to, to help decrease the workload.
  • Reason: Do I really need to explain this? It's a horrible violation of copyright, and as far as I am aware, wikipedia has no to zero tolerance for copyright.

- — dαlus Contribs 01:47, 16 January 2010 (UTC)

Bad charts

Maintenance is needed on filter 81. Since it was released, the following charts have been added to WP:BADCHARTS:

  • Australia Country Tracks
  • Chile Singles Chart
  • Chilean Singles Chart
  • Columbia Singles Chart
  • Columbian Singles Chart
  • Latin Areschart
  • Odyssey Albums
  • Hit40uk
  • Mexico Top 100

Kww(talk) 20:38, 23 May 2009 (UTC)

And, in the eight days since I requested maintenance,
  • Top Latino
  • Latvian Singles
  • Latvian Airplay.
  • Lithuanian Top 40

If this isn't the right place to request maintenance of an existing filter, will someone let me know?—Kww(talk) 16:32, 31 May 2009 (UTC)

I moved this request due to its age. As I commented above, if there's a better way to request maintenance of an existing filter, please let me know.—Kww(talk) 13:35, 3 June 2009 (UTC)
Done. This is the right place but there are only a handful of people that work with the more complex filters. For my part I've mostly been preoccupied with the license migration effort. Dragons flight (talk) 01:42, 6 June 2009 (UTC)
Thanks. Sorry for nagging.—Kww(talk) 01:45, 6 June 2009 (UTC)

Abusive usernames

  • Task: A filter that disallows users from creating usernames with abusive words within it or immediately blocks the user once they have created the account. Could cross reference it with the words on Lupin's bad word list.
  • Reason: Abusive usernames are created by people with specific modus operandi.

- DFS454 (talk) 11:10, 22 March 2009 (UTC)

We have a blacklist to handle disallowed usernames. -- zzuuzz (talk) 14:01, 22 March 2009 (UTC)
The filters can do things like character normalization and complex logic that the blacklist can't, but I'd want to see a specific proposal before pursuing those things. When reasonably possible, the blacklist should probably be preferred for blocking bad words in usernames, etc. Dragons flight (talk) 17:19, 22 March 2009 (UTC)
Are blocks, such as ▀, blocked ? I remember having viewed abusive usernames using such blocks. Cenarium (talk) 19:28, 22 March 2009 (UTC)
I suggest you ask at the talk page zzuuzz linked above. There is so much in the blacklist is hard to parse out if you don't regularly work there. Dragons flight (talk) 23:14, 22 March 2009 (UTC)
The block-drawing characters do appear to be blocked. --Carnildo (talk) 01:12, 23 March 2009 (UTC)
The blacklist can't catch block-drawing characters in edit summaries, which is where they are typically used by vandals. Could someone add them to filter #52 (edit summary vandalism) please? NawlinWiki (talk) 02:25, 23 March 2009 (UTC)
I can't find anywhere that the block-drawing characters are blocked, except for one certain sequence of them. Stifle (talk) 23:38, 25 April 2009 (UTC
I think that Lupin's bad word list is too broad to be excluded from user names. I have actually seen a legitimate user name which includes the word "troll" in it, and I think that is nearly as blatant an example of a word for user names that one could come up with (other than specific sockpuppeteers or names intended for outing). We need to be careful about excluding words from user names. עוד מישהו Od Mishehu 08:08, 18 June 2009 (UTC)

 Done by Filter 102 --Shirik (Questions or Comments?) 06:46, 6 January 2010 (UTC)

Creating newpages with #REDIRECTs in the middle of them

  • Task:Suggest to the user that they've probably made an error if they've put a #REDIRECT in the middle of a new page they've just tried to create.
  • Reason: A lot of new users are tempted to put #REDIRECT [[John Smith]] every time they want [[John Smith]]. This clutters articles and can be costly in time of time for humans to both correct and notify the user. Obviously, we're talking about very inexperienced editors here.

I don't know much about the exact abuse filter keywords and efficiencies, but I'll have a stab at a rough process:

  1. Ignore edits which aren't creating newpages from scratch (old_size = 0?)
  2. Ignore edits from experienced users (purely for efficiency purposes) (user_editcount < 10?)
  3. Ignore if the first character is a "#" (perhaps; again, for efficiency)
  4. Now the hard work: does newlines contain #REDIRECT?
  5. If so, give a quick edit warning.

- Jarry1250 (t, c) 16:14, 28 March 2009 (UTC)

Testing as Special:AbuseFilter/113. Dragons flight (talk) 17:49, 28 March 2009 (UTC)
Just seems to have picked up "Test page"-type button pressing efforts so far, but that's not all bad (not limiting it to just newpages was bound to have that effect anyway - and I'm pretty sure we have a stronger filter for that). - Jarry1250 (t, c) 11:45, 29 March 2009 (UTC)
Okay, I think it's time to set this baby to warn mode. If it's possible, this should be lower in preference to the test-edit filter (so people just button spamming don't get the warning). - Jarry1250 (t, c) 21:07, 25 April 2009 (UTC)
 Done by Filter 113 --Shirik (Questions or Comments?) 06:31, 6 January 2010 (UTC)

New users moving another user's user page, user talk page, or subpages

  • Task: To prevent new users from moving another user's userpage, user talk page, or subpages.
  • Reason: I've only ever seen this done by serial pagemove vandals -- there should be no legitimate reason for a new user to move someone else's user page or user talk page. I can't seem to get the coding right for this. Thanks.

- NawlinWiki (talk) 21:21, 30 March 2009 (UTC)

Did you already start a rule, if so which one? Dragons flight (talk) 23:55, 30 March 2009 (UTC)
No, I didn't start one. There's disabled filter 99, but that's meant to do something different. NawlinWiki (talk) 00:21, 31 March 2009 (UTC)
Filter 123 is now running in log-only mode. It initially hit someone moving their own page which has now been excluded. Should work fine and get disallowed entirely in a couple of days. - Mgm|(talk) 11:35, 31 March 2009 (UTC)
Isn't there a "semi protection" for page moves? Why not just automatically semi-protect everything in the user space? Then only the page's owner or an admin can do stuff. A user talk page would still have to be open to edits by anyone, but protected from moves.
I have always felt that a user's main page should be permanently semi-protected by default. ~Amatulić (talk) 00:47, 24 April 2009 (UTC)
Semi protection for page moves is redundant because only autoconfirmed users can move pages anyway. –xeno talk 00:51, 24 April 2009 (UTC)
I have proposed in the past a change to this. Autoconfirmed users should be able to move only non-user pages apart from their own. -- Magioladitis (talk) 04:07, 24 April 2009 (UTC)
 Done by Filter 123 --Shirik (Questions or Comments?) 06:29, 6 January 2010 (UTC)

De-categorizing edits

  • Task: Identify edits that make categorized articles uncategorized. This is usually the result of vandalism but it is occasionally a result of honest mistakes. I suppose it's pretty trivial to simply check for edits that explicitly remove the categories but it's not uncommon for newbies to screw-up the parser (for instance an unclosed <ref> tag, see this test edit on a random page). I'm not sure if that can be verified efficiently but there should at least be a filter in place for the simple removal of the category section of an article.
  • Reason: Pretty self-explanatory. - Pascal.Tesson (talk) 19:29, 4 April 2009 (UTC)
    • I've created a filter that checks for the removal of all categories (Filter 132). It should be possible to create a filter that checks for the "Cite error" message, too. --Conti| 19:55, 4 April 2009 (UTC)
      • Thanks. But now that it's up and running, I can see that the log is going to be gigantic and there will be significant overlap with other existing filters. Maybe integrating this into, say, Filter 3 makes more sense. Pascal.Tesson (talk) 20:10, 4 April 2009 (UTC)
  • That was simple. Special:AbuseFilter/133, for any citation errors. --Conti| 20:21, 4 April 2009 (UTC)
  • The message is good. However, prior to enabling it on the filter, you should have some mechanism to make sure the cite error appeared at the edit in question (e.g. see this log entry). Unfortunately, without old_html the only way I see it as possible is by duplicating the work of the cite.php extension. —Admiral Norton (talk) 10:51, 5 April 2009 (UTC)

 Done by Filter 132 --Shirik (Questions or Comments?) 06:19, 6 January 2010 (UTC)

Anti-specific attacks Abuse

  • Task: To help detect and prevent vandalism edits. It should detect phrases in the edit summary such as:
    • "I am going to saw off your head with a pocket knife. Then I will light your corpse afire and feed it to my dogs. After that, I will murder your entire family."
    • "OXYMORON83’S NAZI ANUS IS STRETCHED AND WIDENED BY GRAWP’S MASSIVE COCK"
  • Reason: While my examples are related to Oxymoron83 history, there are a number of other editors who are vandalised as well. The antiabusebot currently takes care of this, but perhaps the edit filter would be better.Smallman12q (talk) 13:41, 10 April 2009 (UTC)
    The first could be added to our existing filter covering this. The second is already. –xeno (talk) 01:56, 12 April 2009 (UTC)
    But most of us can't see those. So expect a lot of redundant requests like this one. :| Gurch (talk) 17:33, 17 April 2009 (UTC)
    Hopefully people will come to realize if they can't find any recent diffs then it's already been written in... –xeno talk 14:18, 21 April 2009 (UTC)

 Done Handled by many different, specific vandalism filters. If anything additional is needed, open a new request. --Shirik (Questions or Comments?) 06:14, 6 January 2010 (UTC)

Anti-self redirect

  • Task: Firstly, the filter should prevent the creation of pages that redirect to themselves. Secondly, it could possibly check to make sure that the redirect target doesn't redirect back to the host.
  • Reason: This simply bumps up the number of pages without creating any actual content. I can't posit as to why a page would redirect to itself.

- Smallman12q (talk) 00:26, 17 April 2009 (UTC)

Is this actually a problem that occurs? I've never seen someone edit an article to make it a self-redirect. Also, we can't evaluate redirects, so all we could do is test whether the pagename is the same as the redirect target. Dragons flight (talk) 00:33, 17 April 2009 (UTC)
I thought the same thing, but according to the Toolserver, there's currently 5 of these (2 in mainspace), all last edited in the past 4 days. Its either a strange coincidence, or more common than I thought and people just clean them up on a regular basis.
self-redirects
mysql> SELECT page_namespace AS ns,page_title,CAST(rev_timestamp AS DATETIME) AS time FROM page JOIN redirect ON rd_from=page_id JOIN revision ON page_latest=rev_id WHERE rd_namespace=page_namespace AND rd_title=page_title;
+----+------------------------------------+---------------------+
| ns | page_title                         | time                |
+----+------------------------------------+---------------------+
|  0 | Ascaricide                         | 2009-04-17 00:33:19 |
|  0 | Staffan_Kihlbom                    | 2009-04-13 10:05:34 |
|  1 | Prince_Bernhard_of_the_Netherlands | 2009-04-13 17:57:19 |
|  2 | BISHOPN98                          | 2009-04-16 20:42:44 |
|  3 | 128.226.217.63                     | 2009-04-15 05:27:05 |
+----+------------------------------------+---------------------+
5 rows in set (22.01 sec)
-- Mr.Z-man 00:47, 17 April 2009 (UTC)
I'm not sure how frequently it occurs, we do a log-only run to see. And the main problem is testing the pagename to match the redirect target(the other case is very rare. Smallman12q (talk) 00:41, 17 April 2009 (UTC)
Thanks for the info, well at 5 articles, I'm not sure whether its worth implementing a filter, but it would be a good precaution for future spam attacks.Smallman12q (talk) 01:13, 17 April 2009 (UTC)
I would also like to add that the redirect target shouldn't be the page creator or the same name as the page plus user or user talk...no reason for a mainspace redirect to an individual's talk page.Smallman12q (talk) 01:20, 17 April 2009 (UTC)
So what is the status on this one? Can we do a log-only run? I'm interested to see how many come up in a 24 hour period.Smallman12q (talk) 00:47, 20 April 2009 (UTC)
Testing Special:AbuseFilter/151. The first 9 entries in the log are an error (Why does "in" have higher precedence than "+" ?!?). Dragons flight (talk) 00:37, 23 April 2009 (UTC)
It hasn't picked up many, but there are a few such as Hireability.com to User:HireAbility and Michael Muzek to User:Muzekal Mike and Sikku to User:Sikku.Smallman12q (talk) 16:37, 10 May 2009 (UTC)
It seems a bit glitched such as with Clive Anderson's Chatroom to Clive Anderson's Chat Room.Smallman12q (talk) 16:39, 10 May 2009 (UTC)

 Done by Filter 151 --Shirik (Questions or Comments?) 06:04, 6 January 2010 (UTC)

Language Remover

  • Task: To prevent the removal of languages.
  • Reason: Some new users have gone around removing random languages from random articles.

- Smallman12q (talk) 16:58, 10 May 2009 (UTC)

By "language", you mean interwiki links? Gurch (talk) 20:47, 10 May 2009 (UTC)
The interwiki maps are largely constructed by bots, so a lot of interwikilink removals are to remove the random errors. There are also removals of interwiki links to articles deleted on other language Wikipedias. And a lot of apparently new users messing with interwiki links are actually experienced users from other Wikis who came to help clean things up. I'm not suggesting this kind of vandalism doesn't happen, but the number of false positives it would produce would probably preclude having the filter execute any actions. And with those false positives in the log, it would probably be too tedious to be worth looking for the true positives. Regardless, the zillion or so interwiki bots will clean up such vandalism in their regular duties anyway. Someguy1221 (talk) 20:30, 11 May 2009 (UTC)
I didn't see it that way, but you are right. The false positives would probably be overwhelming, but I would still prefer a one dry run to see how high the false positive rate is.Smallman12q (talk) 01:33, 12 May 2009 (UTC)
Possible source code:
(article_namespace == 0)
& !("autoconfirmed" in user_groups)
& (length(removed_lines) > 4)
& (removed_lines rlike ("[[(ar|ca|de|es|fr|hsb|simple|ml:"))
removed_lines

(We should allow rollbackers to undo vandalism).Smallman12q (talk) 01:13, 1 June 2009 (UTC)

 Done by Filter 270 --Shirik (Questions or Comments?) 05:31, 6 January 2010 (UTC)

Warn about using an email address

  • Task: Warn a user if they type in an email address that it is not a good idea. This is for the Wikipedia:Help desk but it could apply to all talk pages (or even articles too). It could just apply to IPs and new users.
  • Reason: New users constantly put their email in questions at the Help desk and other talk pages which exposes them to spam and causes editors to remove the email addresses.

- Commander Keane (talk) 01:20, 22 July 2009 (UTC)

 Done - Special:AbuseFilter/211. Currently in log-only mode. (X! · talk)  · @147  ·  02:32, 22 July 2009 (UTC)
Or you could search for "email" in the list of filters and find Special:AbuseFilter/76 (currently disabled) with quite a long discussion (in the notes). — AlexSm 15:29, 22 July 2009 (UTC)
I can't find the notes, can you provide a link? Thanks. --Commander Keane (talk) 00:34, 23 July 2009 (UTC)
On the page Special:AbuseFilter/76 the "notes" textbox is under the "conditions" textbox. It's only 5 lines in height, so you have to scroll it. Without "Edit Filter manager" right this might be difficult on some browsers because mediazilla:19266 did not seem to fix this issue. — AlexSm 14:09, 23 July 2009 (UTC)
Unresolved

Smallman12q (talk) 23:49, 8 August 2009 (UTC)

I'm not one for advocating hiding logs, but can the filter log be for admin's only? Email addresses can still be seen in the filter hit log (https://secure.wikimedia.org/wikipedia/en/w/index.php?title=Special:AbuseLog&wpSearchFilter=211). As far as I can see, this just makes it easier for bots to farm and spam them.Smallman12q (talk) 21:59, 7 August 2009 (UTC)
Possible optimization (\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,6}) taken from regex cheat sheet=D.Smallman12q (talk) 23:57, 9 August 2009 (UTC)

 Done by Filter 247. Unfortunately the log issue is not addressable. --Shirik (Questions or Comments?) 05:21, 6 January 2010 (UTC)

Edit warring

  • Task: Flag edits that are from new users repeatedly reverting edits of another user. This can be detected by a throttle and looking if the edit summary contains certain keywords (revert, undo)

New users conducting large scale reverts are universally a sign of abuse, usually of sockpuppet accounts. Triplestop x3 00:23, 10 September 2009 (UTC)

Is this sufficient? - Kingpin13 (talk) 16:52, 29 September 2009 (UTC)
 Done Filter 249 --Shirik (Questions or Comments?) 05:03, 6 January 2010 (UTC)

Inhibit "Peter Dais" addition

  • Task: Prevent the text "Peter Dais" from being added to any articles.
  • Reason: For at least four months now, a vandal using newly registered accounts as well as rapidly shifting IPs has been expressing a grudge against Peter Dais by adding his name to various articles. The vandalism is always quite sneaky in nature: [1], [2], [3], [4], [5], [6]. Additionally, the vandalism often looks plausible so it sometimes remains in an article for months before being reverted: 79 days, 74 days, 72 days, 71 days, etc. Since WP:RBI does not appear to be a viable option here, an edit filter looks like the best solution. — Kralizec! (talk) 16:55, 3 October 2009 (UTC)
Comment: The vandal seems to somehow be so on-the-ball that he caught this addition here and reverted it twice.... I guess a checkuser would be useless on such a shifting IP? I gave both the "reverters" of this request only warnings (as it's the only edit they've both made, so we can only assume they're his incarnations).
Peace and Passion   ("I'm listening....") 06:36, 5 October 2009 (UTC)

Denied requests

Boring Filter

  • Task: To tag edits that only add "boring" such as this edit.
  • Reason: Additions of "boring" should only be made when there are reliable sources.

-Smallman12q (talk) 14:12, 13 February 2010 (UTC)

There really is no way to detect the addition of a single word, and surely there is no way for the filter to detect which sources are reliable. --Shirik (Questions or Comments?) 16:59, 13 February 2010 (UTC)
And also, what about if that word was inside a quote? Minimac (talk) 13:18, 18 February 2010 (UTC)
 Not done per concerns and no reply. --Shirik (Questions or Comments?) 19:11, 19 February 2010 (UTC)

Resume Filter

  • Task: To tag edits that put up resumes.
  • Reason: Wikipedia articles are not a place to put up resumes such as in this edit.

- Smallman12q (talk) 14:12, 13 February 2010 (UTC)

Is this an actual recurring problem or just petty vandalism? --Shirik (Questions or Comments?) 17:00, 13 February 2010 (UTC)
 Not done No reply. If this is a recurring problem, please feel free to open a new request.

Short Template

  • Task: If a new template is less than x bytes, this tags the edit as "short new template".
  • Reason: Most really short templates such as my {{short test}} have virtually no use and are either edit tests, or spam.

- Smallman12q (talk) 20:22, 11 February 2010 (UTC)

Suggested code...
      (article_namespace == 10) & 
      (old_size == 0) & 
      (new_size < 50)
It would be something similar to Filter 98.Smallman12q (talk) 20:29, 11 February 2010 (UTC)
Is this really an issue that would warrant a filter? If it's not going to get at least one hit a day, the performance hit probably can't be justified, because we're already really close to the condition limit per edit and we have very few to spare. --Shirik (Questions or Comments?) 20:24, 12 February 2010 (UTC)
 Not done per concerns with performance. --Shirik (Questions or Comments?) 19:11, 19 February 2010 (UTC)


Protection Templates Addition/Removal

  • Task: To prevent non-sysops from adding/removing protection templates
  • Reason: There's no reason for people to be adding/removing these templates unless they are doing administrative work (they're sysops).

- Smallman12q (talk) 20:50, 11 February 2010 (UTC)

What does it hurt if non=-admins add or remove these? It doesn't change the protection of the page, so it's deck chairs on the Titanic, so to speak. Further, I never remove the templates when protection is over, as that allows a non-admin bot or non-admin human to do it. tedder (talk) 22:40, 11 February 2010 (UTC)
 Not done I believe bots do this primarily, and quite a few admins already don't manage these templates. Naturally, the true protection level is handled by the admin tools and is reflected when an attempt is made to edit the page, so the lack (or improper addition) of this template should not be an issue. If this really needs to be tracked, this is more appropriate for a bot task (though as I mentioned, I believe a bot already manages this, though I could be mistaken in its scope). On top of that, there's really no way to verify that an add or remove of the template is proper (there is no way to check the protection level). --Shirik (Questions or Comments?) 20:22, 12 February 2010 (UTC)

Sarey Savy

After dealing with some recent activity around this artist, I'm beginning to wonder if a filter is needed. To date, articles about this artist have been created over the past 6+ months at Sarey, Sarey Savy, Savy Sarey, Sarey (Singer), Sarey Savy (Singer), Me & U (Sarey Savy), Don't Stop Believin' (By: Sarey Savy), and Don't Stop Believin' (Sarey Savy Song). These have been created by several SPA accounts, which are listed at Wikipedia:Sockpuppet investigations/Googleisawesome (see "Comments by other users" section of that page for list of all suspected meat/sockpuppets).

These have resulted in disruption, and multiple AfDs being created: Wikipedia:Articles for deletion/Sarey Savy, Wikipedia:Articles for deletion/Sarey Savy (2nd nomination), and Wikipedia:Articles for deletion/Sarey (Singer) due to the artist failing to currently meet notability criteria under WP:MUSIC.

Due to the persistence and creative name variations being employed, I'm thinking that a filter would be beneficial to prevent further creation - especially by unconfirmed or relatively new user accounts. However, I'm uncertain as to the best way to define the filter. Do others agree that a filter is called for in this case? If not, any other suggestions for a means to be more pro-active in dealing/preventing this disruption? --- Barek (talkcontribs) - 04:29, 13 July 2009 (UTC)

I think this article about this music person has been notable. The article provided he at least one of the notability in music, if this article is made again then someone please remake the article so it makes the article notable due to this is a real singer. (75.92.208.32 (talk) 05:21, 6 August 2009 (UTC))

This might be a subject better suited by WP:RFPP. You can request that an admin apply create-protection to the articles there. Regards, MacMedtalkstalk 15:51, 7 August 2009 (UTC)

I'm going to protect all of these articles. (If anyone has filed an RFPP request, it can be withdrawn.) rʨanaɢ talk/contribs 13:55, 8 August 2009 (UTC)
This comment is coming a bit late, but since Mr Sarey is still out there creating articles about himself as recently as three days ago, I would like to point out that the IP address claiming above that Sarey Savy is notable is in fact Sarey Savy himself, as determined by Wikipedia:Sockpuppet investigations/Googleisawesome. -- Soap Talk/Contributions 02:07, 27 December 2009 (UTC)
Just posting to point out that he's still out there, and his latest attempt to get his page was Sarey Savy (Singer-Songwriter). I do think this is better served by something other than edit filters, if only because edit filters run on every edit and thus incur huge amounts of unnecessary processing, whereas protection incurs none and a title blacklist entry (unlikely to be granted, but theoretically possible) only runs on new page creations. -- Soap Talk/Contributions 01:54, 16 January 2010 (UTC)
I'm inclined to agree with User:Soap here. Is this frequent enough to warrant a filter? It's trivial to exempt most edits in a single condition with (article_articleid == 0), but if we're talking about less than one page a week here, then it's probably not warranted here. Even 1 a day doesn't really add too much work to admins. --Shirik (Questions or Comments?) 07:19, 10 February 2010 (UTC)
 Not done On second thought, this really isn't reasonable for an edit filter. The only real condition that could handle it would be looking for Sarey Savy in the title, and this is something more appropriate for the blacklist than it is for an edit filter. If there is a more specific condition to be handled by a filter, feel free to open a new request. --Shirik (Questions or Comments?) 07:27, 10 February 2010 (UTC)

Warn on creating article ending in forward slash

  • Task: Warn editors if they're creating a page that ends in a forward slash (/)
  • Reason: Rarely if ever needed - can cause naming problems if some folks don't read instructions for forms (see WP:MEDCAB's form, for example; if someone doesn't fill it out, it will be categorized as just a date)

- Xavexgoem (talk) 07:55, 25 January 2010 (UTC)

Does this really happen enough to be problematic? Additionally, it's not really abuse for which the edit filter is intended. --Shirik (Questions or Comments?) 16:58, 26 January 2010 (UTC)
 Not done no response. --Shirik (Questions or Comments?) 17:59, 3 February 2010 (UTC)

Best & Less

  • Task: Disallow the adding of the text "Best & Less opened its first store at Parramatta, NSW on 27 May 1965. In April 1998 it was acquired by Pepkor Limited - a leading listed retail group in South Africa. The acquisition by Pepkor has enabled Best & Less to grow to the current 185 stores, with an anticipated 200 stores by the end of 2010. The current Best & Less Head Office is located in Leichhardt, New South Wales." - to prevent IP's owned by the company to constantly readd this text. They have constantly done this for a few months now to Best & Less.
  • Reason: To prevent Best & Less from readding their preferred version of a page. Please see page history and talk page for details.

- Sk8er5000 (talk) 00:31, 19 January 2010 (UTC)

I have that page watchlisted; I've only seen edits appear off and on every few weeks at most (with sometimes gaps of many months). Is it really worth the overhead of running a filter on every edit just to catch this one? -- Soap Talk/Contributions 00:53, 19 January 2010 (UTC)
Well, the options that are open are semi-protecting the article, which would deny legitimate IP's the ability to correct the article; blocking the IP's (which I don't thinks allowed per policy a indef IP block); this; or send a abuse complaint to Best & Less (I have a draft in my email), but that would fall on deaf ears if this is part of their stragigy. Simply put, this would be the best option to protect the page (yes, I understand this could slow the Wiki up, but I think this is the best option if we still want to have legitimate IP's still contribute to the article). Thanks -- Sk8er5000 (talk) 04:10, 19 January 2010 (UTC)
There's two sides to this. Soap raises legitimate concerns about running a filter for a single page. On the other hand, short-circuit logic could be exploited to restrict the typical consumed conditions to 2 and run extremely quickly for everything except that one page, in which case the "meat" of the filter begins its work. However, I worry about setting a precedent, too. If it's done for this page, what's to stop the edit filter from becoming the "new" page protection, with requests coming here instead of WP:RPP? I'm not sure the edit rate on that article is high enough to argue that the use of {{editsemiprotected}} is too much overhead to keep the article up-to-speed. Myself, I'm torn in both directions, but leaning towards not creating the filter for those concerns. But that's not to say my mind can't be changed should there be discussion to encourage the opposite move. --Shirik (Questions or Comments?) 07:53, 19 January 2010 (UTC)
True. I also don't believe this is WP:RPP, I came here thinking that it would be the best idea considering we should be welcoming newbies to help catch typos and help with the three maintance tags that are currently on the page. It would have also forced B&L to have to find some other, NPOV way of expressing that infomation. I'll leave it to someone else to decide if the filter should be added. -- Sk8er5000 (talk) 09:28, 19 January 2010 (UTC)
 Not done Insufficient activity on the page to warrant the overhead of a filter. If activity escalates, feel free to request again or go to WP:RPP. --Shirik (Questions or Comments?) 18:03, 3 February 2010 (UTC)

New user change reference name

  • Task: Tag the edit when a new user change a reference name and does not change all the refs with the same old name.
  • Reason: Anomiebot is often able to correct the errors that result from deleting the parent reference, but not the ones that result from changing a reference name. You may ask why not simply browse Category:Pages with broken reference names? But correcting this error requires going through the history to find out who caused the error and how, and to sort out which errors are likely to be corrected by bots and which are not.

- Sole Soul (talk) 09:02, 16 January 2010 (UTC)

Honestly, I'm not even sure if this is technically feasible, but even if it was, I fear it would be extremely complex and too demanding on resources. How about a feature request to Anomiebot? --Shirik (Questions or Comments?) 03:45, 17 January 2010 (UTC)
I assumed that detecting a change in <ref name="name" /> is not that complex, and not any different from other filters that detect similar changes. I thought about the Anomiebot request but fixing this error requires usually reverting the user which is unlikely to be acceptable if done by a bot. Sole Soul (talk) 06:44, 17 January 2010 (UTC)
The problem isn't really the detected change in a particular reference. It's the second half: Determining if there are other reference tags that were not switched over. Because of the (admittedly sane) limitations of what the edit filter can do, there isn't really any capability of feed-forward information that I'm aware off, and even if there was, this would be performance-intensive because it would require reviewing the entire page on every edit, instead of just the part that changed (believe it or not, this can make a huge difference). --Shirik (Questions or Comments?) 10:10, 17 January 2010 (UTC)
I see. What about: new user + change in ref name only as a test. Sole Soul (talk) 10:31, 17 January 2010 (UTC)
Sorry for the delayed response. I've been trying to think about this and I still can't get around this issue: detecting the change actually is more difficult than I first considered. A change is essentially a removal followed by an addition. We can detect that this removal and addition occurred, but have no way of detecting that it's not something else on the same line (because the entire line is detected as removed). So if you change the line on which a reference is located, but don't change the name of the reference, it would still be detected. There's no capability to store regex match groups, so building a regex off that doesn't seem possible. Unfortunately, it's looking more and more like this filter can't be implemented. --Shirik (Questions or Comments?) 23:34, 22 January 2010 (UTC)
OK. Thank you, I appreciate it. Sole Soul (talk) 23:37, 22 January 2010 (UTC)
 Not done per the above concerns --Shirik (Questions or Comments?) 19:41, 23 January 2010 (UTC)

New user removing images

  • Task: Tag edits by non-autoconfirmed editors which remove, change, or add an image to an article
  • Reason: MediaWiki:Bad image list has significantly cut down vandalism from inappropriate images, but that depends on a particular image already having been identified as problematic. Removal of an image generally by new or anonymous users is likely to be blanking vandalism or misguided censorship of a controversial topic, and the existing blanking tags would not catch this unless the user deletes a larger portion of the text along with the image. If the image is a validly used non-free image, it can get automatically tagged for deletion as an orphan if no one catches its removal in time, which at minimum causes wasted labor.

Shirik (Questions or Comments?) 03:41, 6 January 2010 (UTC) on behalf of User:Postdlf

I think this is a reasonable request, and it shouldn't cause too many performance problems. However, is detecting an added image really necessary? Or could it just be removed and changed images? The reason I ask is because checking for an added image would nearly double the execution time (as that is the most expensive part, and "changed" can be caught at the same time as "removed", as a change is fundamentally a removal then an addition). I'm not entirely sure the addition of an image indicates as much of a problem as removing or changing an image does, but feel free to convince me otherwise. --Shirik (Questions or Comments?) 03:41, 6 January 2010 (UTC)
No, I'm more concerned with removing or changing images too. postdlf (talk) 14:21, 6 January 2010 (UTC)
Filter 280 is in unit testing. --Shirik (Questions or Comments?) 20:16, 6 January 2010 (UTC)
Running, log only, but unfortunately I think the false positive rate may be a bit too high, with no obvious way to correct the issue. Performance is concerning as well. --Shirik (Questions or Comments?) 10:37, 7 January 2010 (UTC)
Filter 280 disabled for now. There seem to be too many hits for this to be useful as-is. If the use case could be clarified perhaps it could be investigated. See the logs for my concerns. Perhaps we could consider only cases where images are removed, and not changed (though this can be performance intensive). --Shirik (Questions or Comments?) 19:49, 10 January 2010 (UTC)
 Not done Simply to complex for an edit filter to be both useful and sufficiently condensed for the site. I will keep this in the back of my head, though, should any bright ideas arise. --Shirik (Questions or Comments?) 20:26, 31 January 2010 (UTC)

Digimon & Resident Evil

- 上村七美 (Nanami-chan) | talkback | contribs 23:52, 22 November 2009 (UTC)

Vandalism on the involved Digimon articles is still continuing intermittently, even after this request has been posted. - 上村七美 (Nanami-chan) | talkback | contribs 14:07, 27 November 2009 (UTC)
I'd like to personally devise a suitable filter, but don't know how. I've made a request for some basic guidance should someone want to quickly show me the ropes. Cheers. Nja247 15:32, 3 December 2009 (UTC)
Note: Filter 274 has been created for this purpose, but it is currently disabled. --Shirik (Questions or Comments?) 08:17, 7 January 2010 (UTC)
done and not done; created, but disabled. Won't reenable under the assumption that it can be re-enabled on demand if necessary. --Shirik (Questions or Comments?) 05:58, 10 February 2010 (UTC)

Japanhero spam

  • Task: Block the text "japanhero.webs.com" from being placed anywhere.
  • Reason: Sockpuppeteer Japanhero (talk · contribs) owns the Japanhero.webs.com fansite. There is never going to be a use for his links on Wikipedia. However, due to the nature of webs.com, there are multiple domain names that he uses (see [7], [8], and [9]). As it is highly unlikely that he will ever change his domain name, this is the most effective way to prevent him from editing as well as track any new sockpuppets that may show up (the nature of the links means that blacklisting the links will be ineffective).

- —Ryūlóng (竜龙) 06:30, 1 November 2009 (UTC)

"Rush is right" sockpuppeteer

  • Task: Detect and disallow edits made by the "Rush is right" sockpuppeteer. Technically, that would mean disallowing edits where the summary contains either "Rush" or "Limbaugh" and where a lead sentence is inserted containing either "Barack" or "Obama," and "communist" (case-insensitive). A bot (e.g. Mr.Z-bot (talk · contribs)) would immediately file a report at WP:AIV.
  • Reason: We're dealing here with a tireless vandal who claims to have access to thousands of IPs and has used dozens for this purpose already. See the modus operandi here. The addition is made to totally unrelated articles.

-  Blanchardb -MeMyEarsMyMouth- timed 16:01, 19 October 2009 (UTC)

Tag additions referring to Roman Polanski

Task: Tag all edits adding the word Polanski to any article in main space.
Reason: There are currently a lot of suspect edits either WP:BLP contraventions or adding WP:UNDUE weight to articles with reference to the current Roman Polanski news. This filter would allow these edits to be checked easier. Martin451 (talk) 06:02, 2 October 2009 (UTC)

I was just coming here to request the same. I agree, it should tag "Roman Polanski" or "Polanski" in any articles except his own. They are happening on many other articles. tedder (talk) 09:41, 3 October 2009 (UTC)
 Not done I've been monitoring this situation and it seems a bit stale. If it's still necessary, feel free to add a new request. --Shirik (Questions or Comments?) 06:01, 10 February 2010 (UTC)

Automated page creation

  • Task: Warn about automated page creation w/o approval.
  • Reason: A recent discussion decided that all automated/semi-automated creation of new articles/stubs should go through the BRFA process, even if they don't technically meet the specification of a BOT. Naturally, most editors don't know about this any automated creation of stubs has continued. It is doubtful that a significant % of the community will hear about this change, regardless of how well it is advertised. As such, I am requesting a filter to inform people of this change when they hit say 5 new pages created within 5 minutes, using the throttle function.

I am well aware that this won't catch much compared to its processing cost, but feel it is worth a try. Due to the nature of the task, it should match for all users except bots. -- ThaddeusB (talk) 16:09, 29 September 2009 (UTC)

Prevent creation of new talk pages when there is no corresponding article page

  • Task: Prevent the creation of talk pages without corresponding article pages. This is mainly done by IP/anon editors (though not always), who presumably find they are unable to create an article in the main article space. Sometimes it is plain test edits; sometimes it is vandalism; sometimes it is a means to ask questions; sometimes it is draft articles; sometimes a similar article exists but the editor has misspelled some part of the title. Such attempts should instead direct editors to the sandbox, the help/reference desks, the create account page, or requested articles. I haven't checked, but perhaps other talk namespaces (except User talk) have a similar problem but on a smaller scale.
  • Reason: The problem is that while talk page vandalism is usually quickly reverted, these orphaned talk pages don't seem to get any attention except from myself and a couple of other editors. For example, recent examples are:Talk:Maharoli, Talk:National treasures of china, Talk:Sonoclot. Many get speedily deleted, but some fall through the net such as Talk:Sign-in Seal, Talk:Romano Amodeo, Talk:Eric durnin and Talk:Augusto Cicaré

- Astronaut (talk) 12:59, 17 August 2009 (UTC)

  • I'm not sure if the edit filter can check whether the corresponding article exists. Also, archive subpages and other talk subpages would be inappropriately caught by this filter. Stifle (talk) 11:50, 3 September 2009 (UTC)
Should we get over the first problem (detecting whether the corresponding article exists), the second would be simple - check if the page name has a "/" character in it. עוד מישהו Od Mishehu 12:13, 6 September 2009 (UTC)
 Not done Not really possible for a filter. Should be fairly trivial for a bot, though. That's the best way to go. --Shirik (Questions or Comments?) 06:04, 10 February 2010 (UTC)

Disallow creating articles with multiple single quotes

Inexperienced users sometimes create articles at a title containing runs of multiple single quotes (e.g., '''foo. Although such titles are legal, they are nearly always an error - an attempt to include wikimarkup in a page title, to bluelink a broken wikilink that contains wikimarkup, or to compose a double quote mark from two single quotes. In most cases, page titles with this error begin with a string of such single quotes, so detecting this should be efficient. Gavia immer (talk) 05:24, 13 August 2009 (UTC)

See bug 20213. MER-C 09:15, 13 August 2009 (UTC)
Poke, poke - any update on implementing this? Gavia immer (talk) 15:58, 1 November 2009 (UTC)
Is this really a big enough issue that it deserves a filter? I'm just taking a guess here, but I wouldn't think it would be very common, and can be addressed fairly easily with a move and delete, etc. --Shirik (Questions or Comments?) 10:16, 17 January 2010 (UTC)
 Not done No response --Shirik (Questions or Comments?) 06:07, 10 February 2010 (UTC)

File talk: clarification

  • Task: If a user is an IP or has less than 20 edits and attempts to create a page in the File talk namespace, they get a warning similar to MediaWiki:Editnotice-7 reminding them that their edit is probably better served at WP:MCD. They would still be able to commit the edit, just an additional step to make sure they understand the odds are against anyone responding to them.
  • Reason: I've deleted over 6,000 pages in the file talk namespace, nearly 1/7th the entire namespace and people continue adding off-topic, nonsensical, or other pages to it. MBisanz talk 08:11, 22 April 2009 (UTC)

- MBisanz talk 08:11, 22 April 2009 (UTC)

I'm not sure this is a good use for a filter. Would it be better to modify MediaWiki:Newarticletext to provide a bolder / scarier message when creating a new page in file talk? One could also provide links to other likely venues like MCD. That message already has namespace dependent behavior, so it wouldn't be much of a stretch. Dragons flight (talk) 00:18, 23 April 2009 (UTC)
It was scarier to begin with, but I was told it was overloading the parser or something, so it had to be normalized to a plain box. MBisanz talk 00:25, 23 April 2009 (UTC)
Huh? Could you point me at that discussion? Seems unlikely. Dragons flight (talk) 00:45, 23 April 2009 (UTC)
Here was the edit [10] I was thinking of. MBisanz talk 00:57, 23 April 2009 (UTC)
Do you have the wrong link? I'm not sure how that is relevant. Dragons flight (talk) 01:13, 23 April 2009 (UTC)
This would be the other edit [11], I took that to mean that namespace-wide edit notices should only use a basic box structure. MBisanz talk 01:21, 23 April 2009 (UTC)
I see you've already modified Editnotice-7, to me that feels good enough. Dragons flight (talk) 18:05, 2 May 2009 (UTC)
 Not done per User:Dragons flight
  • Task: Checks added lines for any links containing. youtube. Displays a warning about linking to pages where the content may be unauthorised copyrighted material or copyright status is in serious doubt (PER COPYLINK). Also logs the incident so experienced users can keep tabs on torrent additions and check wether they are needed (and legal). Would run on all namespaces.
  • Reason: Given the recent Pirate Bay case decision we've gone through and dicovered many links to illegal torrents that we have now removed. There are very few cases were torrents are used legitimatly and as such I think a notice reminding users about not posting links to illegal torrents is required.

  «l| Ψrometheăn ™|l»  (talk) 13:32, 17 April 2009 (UTC)

I believe that wikipedia's policy is to not censor. It'll be very difficult to determine which torrent is legal, and which is not. Likewise, I would propose that youtube should also get censored for its copyright infringement. Smallman12q (talk) 22:07, 17 April 2009 (UTC)
Not Really Smallman12q, Wikipedia's policy is to remove blatant copyright infingement. Links such as this are pretty obvious copyvios and given our policy i dont think we should turn a blind eye to this sort of stuff. The question is wether or not we allow people to LINK to illegal torrents or not. An experienced user (such as myself) can spot an illegal torrent a mile a way, as a simple google result for what the torrent is will reveal wether it is a "free download" or a copyrighted product or artistic work. YouTube isn't a problem because they remove copyrighte dmaterial as its reported and most of youtubes material isnt even illegal. Whereas torrent sites such as the pirate bay primarily deal in illegal downloads and most do not remove illegal downloads when reported.   «l| Ψrometheăn ™|l»  (talk) 04:36, 19 April 2009 (UTC)
Perhaps we should seek a policy addendum or a policy for torrents in particular. Youtube has quite a bit of controversial copyrighted content in the form of copyrighted songs. I'm not against banning torrents, its just that I believe there should be a specific policy/guideline written to deal with them.Smallman12q (talk) 00:45, 20 April 2009 (UTC)
I would support a check for torrents and also YouTube links, with a warning about linking to copyrighted content. Stifle (talk) 10:44, 20 April 2009 (UTC)
Can User:XLinkBot be of help here. --Dirk Beetstra T C 10:55, 20 April 2009 (UTC)
Not really, Im after a filter to display a warning message to ALL users who add torrents advising wether or not they should be added and logging when they are. XLinkbot seems to be automatically reverting additions of some links (Both legitimate or otherwise) by un-registered or new users (I'm struggling to understand why something like that was even approved when the collateral damage ratio must be enormous). Wikipedia has got 2 concurring policies regarding linking to copyrighted material (WP:ELNEVER WP:COPYLINK).   «l| Ψrometheăn ™|l»  (talk) 14:02, 21 April 2009 (UTC)
Some torrents are legal (and are not copyvios), and some are not. Same thing with YouTube; some copyright violations, and not. --Kanonkas :  Talk  14:11, 21 April 2009 (UTC)
Exactly, That is why XLinkbot would fail at the task. The filter would makes users think about what they are linking to on torrent sites and youtube before they do it by displaying a message and would log addditions so that they can be checked for context and/or copyright status. We have thousands of links to Youtube (Most legit but a sizeable preportion not.) [12][13][14][15]. Wikipedias policies are quite clear on this matter (That unauthorised copyrighted material should never be linked to) and thus we should start doing something about it and that a warning message when youtube and torrent links are added letting users know not to link to unauthorised copyrighted material or were copyright status in serious doubt (and that if it isn't or is unlikely to be unauthorised then procede with the save) is a good start. Policy asside, From a legal perspective Intellectual Reserve v. Utah Lighthouse Ministry established the principal that if a website (such as wikipedia) deep links to unauthorised copyrighted material on another site knowingly that it is a contributory copyright infringement.   «l| Ψrometheăn ™|l»  (talk) 14:20, 21 April 2009 (UTC)
That case was struck out. Are you thinking of a different one? Stifle (talk) 22:59, 25 April 2009 (UTC)
I'm writing something for this. Stifle (talk) 23:11, 25 April 2009 (UTC)
Special:Abusefilter/155 written. I haven't enabled it because I'm afraid I'll crash the wiki. Can someone who's better at abusefilters check that nothing's obviously wrong and enable it for logging? Assuming we don't get mad amounts of false positives, the long-term aim would be to set it to warn using MediaWiki:Abusefilter-warning-copyvio. Stifle (talk) 23:31, 25 April 2009 (UTC)
I'd have a look, but you seem to have made it private. Gurch (talk) 13:50, 26 April 2009 (UTC)
How in the world are you not a sysop? :\ I suppose we could make you an AF editor. Stifle (talk) 17:41, 26 April 2009 (UTC)
Or people could just not make filters private when there is nothing to hide. Gurch (talk) 18:58, 26 April 2009 (UTC)
Meanwhile Dragons flight has checked and improved this and it's now enabled on logging. It's getting a reasonable level of hits. I'll leave it a few more days and set it to warn. Stifle (talk) 17:42, 26 April 2009 (UTC)
I'd like to see a larger discussion, in a more general forum, per enabling a warning on all YouTube, PirateBay, and torrent links. Dragons flight (talk) 21:56, 26 April 2009 (UTC)
I would have liked to have seen the filter regexp myself, but alas, I guess I can't. As I pointed out on the AN thread and during my lengthy discussion with Promethean, the majority of "torrent" links are not named "*.torrent", they can be named anything and hosted anywhere. It also turns out the Children of Hiroshima movie torrent that was removed and initially turned into a posterchild for "bad" torrent links in the AN thread is in fact in the public domain. The only reason I've not added the link back myself is I am still working on a design for a standardized template for linking to these types of media that will include the copyright/licensing information. I intend to follow up on AN once I've finished compiling information, but for the curious, most Japanese films from 1953 and earlier are currently in the pubic domain. Tothwolf (talk) 21:09, 26 April 2009 (UTC)

Can we please remove YouTube from this? Abuse Filter cannot supercede WP:EL, and there's never been a consensus there to block YouTube. rootology (C)(T) 01:47, 27 April 2009 (UTC)

I disabled 155 temporarily--I don't know the specifics of the syntax, so I just turned it off pending review. But for blocking PirateBay--and especially YouTube--we need a much wider discussion than just this page. rootology (C)(T) 01:49, 27 April 2009 (UTC)
Nothing was being blocked; the filter was in log-only mode. Mr.Z-man 01:52, 27 April 2009 (UTC)
This is why I only hit the "enable" button. :) I flipped it back on to log. Still, it shouldn't even be probably set to warn, on this, since the nature of the content and its validity is still periodically under dispute. I'd rather we block 'illegal' torrents, like say for Windows XP, or films, but that's going to be the trick. YouTube is generally fine to link to, and even warnings today would be inappropriate, when even the White House posts to YouTube. rootology (C)(T) 02:02, 27 April 2009 (UTC)
The handful of links that went to bootleg torrent files linked from The Pirate Bay had already been removed. Right now, accounting for the link that was in the AN thread, which has been archived (since AN gets transcluded) we should have right about 160 legitimate links to *.thepiratebay.org. This number had been stable for better than a week. As of a moment ago that was down to 153. One I can account for in this edit, which was a link removal of a film that is in the public domain. Tothwolf (talk) 02:18, 27 April 2009 (UTC)
The filter looks for any URL with a youtube video link, or ending with .torrent, or containing piratebay. I strongly disagree that "youtube is generally fine to link to". The warning (MediaWiki:Abusefilter-warning-copyvio) that I would propose to use would acknowledge that there is some valid content that could be added, but may be enough to discourage people from casually linking to music videos and TV interviews that are up there illegally. Stifle (talk) 08:42, 27 April 2009 (UTC)
That warning is way too harsh. Threatening users to be blocked for adding a youtube link? Seriously? The message should be more of a friendly notice, and not a warning, stating that the addition of a link to a torrent file/youtube video is probably against our policies. --Conti| 11:51, 27 April 2009 (UTC)
Please feel free to edit it accordingly. Stifle (talk) 15:41, 27 April 2009 (UTC)
Get consensus (not here, but on WP:EL) that we can do this for YouTube, then this filter can get added for YouTube. We simply can't do this in such a small niche environment for this, for something that is patently not "abuse". Conti's suggestion above is spot on for torrents, but not for YouTube (or Vimeo, or Hulu, or whomever), when you have multiple official government and corporate content providers now using these sources. You did know that YouTube even just signed a massive deal to stream copyrighted content like Hulu, right? Legally? And that the US President and multiple people in Congress and Senate post updates there? rootology (C)(T) 13:18, 27 April 2009 (UTC)
I'm aware. However, most of the links I'm seeing in the filter log aren't to any of these. I will undertake not to set this on warn on my own motion, given the opposition here. Stifle (talk) 15:41, 27 April 2009 (UTC)
I don't think a warning template of that nature is appropriate at all for any situation where a site may have legitimate content. It would no doubt result in the removal of links for freely licensed and public domain material due to overly cautious editors. After the discussion on AN, I'm still not even sure why a filter for "*.torrent" and "*piratebay*" was attempted. The majority of the torrent metadata file links currently in use on Wikipedia are not deep links to "*.torrent" files (most sites forbid deep linking to the torrent metadata files to prevent url-rewriting abuse) and are not links to thepiratebay.org as The Pirate Bay is only one of 100s (or actually 1000s) of popular bittorrent sites. In theory it would be possible to create a regexp hostname check for the majority of the larger sites, but such a large filter would no doubt result in a noticeable performance hit. As I've previously mentioned, torrent metadata files can be named anything and hosted anywhere. The current checks will simply not catch most links and will not be able to differentiate between links to metadata files for freely licensed content and copyright-infringing content. Considering the small number of links that were found and removed in the immediate aftermath of the Pirate Bay scare, and considering external links seem to already be regularly checked by editors, I personally consider this filter implementation a waste of server resources. Tothwolf (talk) 18:41, 27 April 2009 (UTC)
The issue here has been slightly over exaggerated and very generalised. I agree that the warning template should not be a warning at all, rathor a friendly notice. Secondly (mainly aimed at Rootology), We must not forget that the questionable XLinkbot (approved by the BAG, in a similar fashion to this) reverts additions of certain links by anonymous and new users without even telling them seoncds after addition, wether they were usefull or not. I'm hoping that this filter will remove youtube from the list because as you said youself, youtube is hosting more and more usefull content. Surely displaying a simple notice informing users about our policy regarding unauthorised copyright material and that if you know its unauthorised dont add the link is far better than stealthily reverting thier additions? I also find it ammusing that Rootlogy disabled a filter that he had very little idea what it did (as demonstrated by his misguided statements above) and sysops wonder why a large number of users want the abuse filter right seperated from sysop entirely. I'll say it one more time, we are not blocking youtube or PB, we are notifying users of a relevent policy that is (imo) all too frequently violated, and in effect, if this filter does what im hopping it will I can probably ask for youtube to be removed from the XLinkbot revert list, effectivly UN-blocking youtube for anons and new users...   «l| Ψrometheăn ™|l»  (talk) 01:04, 29 April 2009 (UTC)
I would be all for a warning prior to adding certain sites by new and IP editors. The warning should be extremely friendly, remarking them about the caveats with these links (we tried this earlier, there was or is an earlier rule which did the same!): do NOT link if the site is a copyvio, be sure to be on-topic, etc. etc. When that is done (i.e. warning), I would suggest to remove the rules from XLinkBot, as otherwise double warning may occur, which is certainly bitey.
By the way, where was established that XLinkBot was questionable? I have seen no (recent) thread saying that (and I believe that all concerns and objections were cleared in the BRFA). I did however a recent quick look at some XLinkBot-youtube-reverts (did not do full statistics and evaluation as I have done once for myspace), and did find quite a number which were correctly reverted as they were copyvio's, and quite some which failed WP:EL on other reasons (nothing to do with subject, indirectly linked, etc.). However, it did contain also one or two reasonable additions.
Seeing that quite some additions of youtube by 'new' editors are questionable (to say the least) against WP:EL, a properly written, good, friendly warning (or better: remark) may cut down significantly on the problem additions and on the mistaken XLinkBot reverts, which undoubtedly will be there as well. --Dirk Beetstra T C 11:40, 29 April 2009 (UTC)
Questionable is my own personal assesment based on conversations with newer users on IRC and I personally feel it is questionable that we utilise a bot that would have a hit and miss rate far higher than most, you can't teach a bot what a copy vio is, but you can teach a user. The page blanking filter is an excellent example of what we are trying to do here, the stats say that once a user recieved an automated notice about blanking pages, the vast majority did not carry out the edit and only a small number did.   «l| Ψrometheăn ™|l»  (talk) 11:46, 29 April 2009 (UTC)
OK, I'd like to discuss with those newer editors myself as well to get a better feel, I generally see that the bot is doing quite well, though a bit quick, indeed. And copyvio is really not the only (or even, the biggest) problem with youtube, though still a quite occuring one! But I agree that using a filter would be a way better than the bot on youtube (and also on blogspot, myspace, etc.!). --Dirk Beetstra T C 11:53, 29 April 2009 (UTC)
I'm not even sure why "The Pirate Bay" is still being discussed here. If Youtube links need to be checked then this should be a filter strictly for those. I've already explained this to Promethean at least once before as well as multiple times in all these discussion threads: A check for "*piratebay*" and "*.torrent" it worthless, see my comments above. I've also yet to find either of these checks triggered in the logs for filter 155. See the past discussions on Administrators' noticeboard here and a current discussion here. Tothwolf (talk) 16:11, 29 April 2009 (UTC) (Edit: Updated AN archive link. Tothwolf (talk) 21:14, 3 May 2009 (UTC))
According to a grep -Eil "(torrent|piratebay)" run against the current 520 logs, this filter has only logged three changes that potentially matched anything other than Youtube:
  • 249056 – edits to The Pirate Bay (diff); vandalism, although ironically also a Youtube video link addition (matched "youtube\.[a-z]{2,}/(watch|\?v=)")
  • 267678 – edits to Square (geometry) (diff); vandalism and inclusion of browser-incompatible tracker links (matched "\.torrent" and "piratebay")
  • 268772 – edits to Imaginationland Episode III (diff); website spamming, appears to be a subscription based private tracker site (matched "\.torrent")
Log entries 267678 and 268772 also show that the regexp intended to match "*.torrent" files is flawed because it matched "\.torrent" as a partial domain name.
Molehill =/= Mountain
--Tothwolf (talk) 19:32, 29 April 2009 (UTC)
OK, I'm convinced enough that I would drop piratebay and torrent from this as they're not really necessary. I agree with Promethean's suggestions above. Stifle (talk) 21:16, 29 April 2009 (UTC)
Seconded, it seems the focus of this filter has changed somewhat since the inital request.   «l| Ψrometheăn ™|l»  (talk) 08:14, 30 April 2009 (UTC)

Arbitrary break - YouTube, MySpace, blogs, Facebook, Image sharing sites etc.

I would, per my remarks above, really suggest to shift this to Youtube and other typical sites like MySpace, the blog sites like BlogSpot, Facebook, Image sharing sites (there may be more). Those are the XLinkBot rules which (even if I believe the mistake rate is pretty low) give the most errors, and I really believe that we can, like with the page blanking, get down the number of 'problematic' edits with a good, friendly 'warning' (prefer 'remark' here), and those edits that still go through are 'logged' so easy to check by the RC patrollers (e.g. when tagged). --Dirk Beetstra T C 08:37, 30 April 2009 (UTC)

P.S. I tried this earlier (days/hours after the abusefilter was enabled) with rule 48, but at that time the rule had serious problems with the performance of the wiki, very long runtime. --Dirk Beetstra T C 08:39, 30 April 2009 (UTC)

I think for the sake of taking baby steps, lets try this with youtube for 2 weeks and then revist wether the idea is working or not? Then we can add other sites as required   «l| Ψrometheăn ™|l»  (talk) 08:18, 1 May 2009 (UTC)
Oh, sure! I was already looking at the regex we would need for more sites, that certainly needs a look. Lets try it first with one or two quite frequently used sites and see. Please comment out the rules in User:XLinkBot/RevertList when (before) enabling warning on these rules. --Dirk Beetstra T C 08:51, 1 May 2009 (UTC)
@ Stifle: If it hasn't been done already can you change the "warning" to a "notice" style template, remove the block threat and just put some general info on what not to link to (and that if the link is ok click save again) Im also thinking we'll just set this to tag and warn for !autoconfirmed users at this time because they are the main group that we need to get a feel for :)   «l| Ψrometheăn ™|l»  (talk) 14:08, 1 May 2009 (UTC)
Why would we go to !autoconfirmed users? There are plenty of users who have been here for a long time and have no clue about WP:LINKVIO and other policies. Stifle (talk) 15:24, 1 May 2009 (UTC)
I've amended the warning, see this diff. Stifle (talk) 15:26, 1 May 2009 (UTC)
We're distinguishing here between sites that carry a lot of copyvios and sites that are just terrible sources. I would not be holding the latter in scope of this filter, but feel free to propose another. Stifle (talk) 15:27, 1 May 2009 (UTC)
Point taken, leave it for all then :)   «l| Ψrometheăn ™|l»  (talk) 16:07, 1 May 2009 (UTC)

Whatever we do, can we please not bundle all the popular web 2.0 sites into any one singular filter? E.g., Vimeo, YouTube, Facebook, MySpace, Twitter, etc. if they get rules here (but volume wise aside from YouTube at best, I can't see the point) please put them all in separate rules if we're going to test to see how often they pop up with logging-only.

Since blogs can be acceptable sources, why would we even consider an abuse filter in general for blogspot, typespace, and the like? For example, there was a discussion here where I laid out how you can demonstrate (I was pretty much in "overkill" mode, however, with how much evidence I presented) how an "apparently" anonymous blog is a fine source. There is no policy-supported reason to abuse filter blogs. rootology (C)(T) 14:28, 1 May 2009 (UTC)

The same is true for youtube links, Rootology. The official youtube video's from official TV channels can be great sources. --Dirk Beetstra T C 14:34, 1 May 2009 (UTC)
Oh, certainly. I just get twitchy when my "web 2.0 sense" tingles. ;) rootology (C)(T) 14:43, 1 May 2009 (UTC)
Yeah we wil put all the links to commonly abused web 1.0 stuff in the same filter after this little "test" run.   «l| Ψrometheăn ™|l»  (talk) 16:08, 1 May 2009 (UTC)

Note: Per the above discussion, I've removed the .torrent / piratebay components, so currently the log is just youtube links. Dragons flight (talk) 18:58, 2 May 2009 (UTC)

Why don't we just create a filter that uses XLinkBot's regex and then decomission it?--Ipatrol (talk) 02:05, 3 May 2009 (UTC)

Whilst the same regex can and most liekly will be used. The actions of xlinkbot and abusefilter are quite different, hence a test period is warrented. I just went over the logs given the recent regex change and didnt spot anything abnormal. Given this and that the notice has been changed to be more friendly shall we enable tag and warn? If someone does this please also remember to comment out the youtube line on XLinkbot per Dirk Beetstra   «l| Ψrometheăn ™|l»  (talk) 12:28, 3 May 2009 (UTC)

Well, no, I am suggesting to completly supplant XLinkBot with the filters.--Ipatrol (talk) 02:14, 5 May 2009 (UTC)

We can look at that after we do a trial with youtube to see if a filter is meerly enough to stop the consistent innapropriate linkage, you agree?   «l| Ψrometheăn ™|l»  (talk) 06:45, 5 May 2009 (UTC)
semi-but-not-really-done. Was done by 155 but never completed (only log only and then deleted). Consensus here seems to be pretty against a sweeping ban on this content (which is in contradiction with WP:EL) but that something should be done. I believe a bot is already tracking these types of edits, and if not, a bot should probably be devised to do such a thing to avoid the condition hit of a filter. --Shirik (Questions or Comments?) 06:34, 10 February 2010 (UTC)

simple image transclusion

  • Task:Prevent the placement of images in templates using the <includeonly> tag. Apply to the template namespace and probably ip/new editors.
  • Reason: this is more of a question than a request, but would a filter that could catch include only additions of an image in a template edit be possible? I imagine that we have much more complex filters for other image transclusion problems, so I was surprised to see this one make it to mainspace. Would a brute force restriction on edits like that catch too many good edits (how many ips need to edit a template to only include an image on transcluded pages?)?

- Protonk (talk) 19:28, 12 August 2009 (UTC)

 Not done Primarily because there seems to be too many appropriate uses for this. I'm not really sure how putting it in <inludeonly> is any worse than anywhere else in a template, except for its "hidden" nature. Related changes checks should be able to identify such problems. --Shirik (Questions or Comments?) 06:28, 10 February 2010 (UTC)

Edit breaks formatting

  • Task: Spot changes in the balance of left & right braces and possibly square brackets.
  • Reason: In general, a "correct" article should have an equal number of left and right braces, i.e. everything that is opened is closed. A change in the balance of braces will often indicate that a GF editor has borked a table or template (or fixed something when the balance changes to zero), or a vandal has removed part of the infobox at the top of the article. Comparing the before-after brace-balance would be tolerant of existing imperfections, but would flag new problems. In fact, this should probably be a function of the Preview button, but a filter would be handy to add a note to the edit summary. Obviously, the braces could still be incorrectly placed, this would only catch a subset of these problems. Franamax (talk) 20:02, 20 August 2009 (UTC)
I have tested a bit with (count("(",new_wikitext)!=count(")",new_wikitext))&(count("(",old_wikitext)==count(")",old_wikitext)), which returns a few. However, the filter is not that practical for this and it may be server-heavy, you could ask at check Wikipedia if they can run this. Cenarium (talk) 19:08, 15 November 2009 (UTC)
 Not done Not really an abuse problem and would cause significant server load. User:Cenarium's recommendation seems appropriate. --Shirik (Questions or Comments?) 06:26, 10 February 2010 (UTC)

Entire article blank

  • Task: Disallow any user except admins from blanking articles in the article space. If the article size = 0, disallow/warn.
  • Reason: A common vandal action where the user simply selects all the text and deletes it. Usually in an attempt of censorship/dislike of the article. Special:AbuseFilter/3 picks it up but only warns and is open to false positives, while this wouldn't be.

- Otterathome (talk) 13:12, 21 August 2009 (UTC)

Signing in Article Namespace

  • Task: To warn users and tag articles when they sign their name in article namespace.
  • Reason: There is no reason for people to be signing their name in article namespace.

- Smallman12q (talk) 13:05, 15 January 2010 (UTC)

 Not done This was already attempted in Filter 253 and it was determined the filter would be way too complex to work properly. Additionally, I fear this might be too bitey, as the users that are most likely to do this would be newcomers. If it's a major issue, I think this could be fairly easily cleaned up with a bot. --Shirik (Questions or Comments?) 03:43, 17 January 2010 (UTC)

Icon Group International

  • Task: Warn anyone trying to add "Icon Group International" or "Webster's Quotations, Facts and Phrases" to an article.
  • Reason: Icon Group International publishes a large number of books automatically generated from Wikipedia articles, which show up in Google Books. Many people don't notice and use them as sources here, accidentally creating circular references. It's only a few additions per week and I used to clean up these regularly, but I have been neglecting it recently. I don't really like the edit filters expanding too much, but this would be useful. Is the overhead too much?

- Apoc2400 (talk) 21:14, 21 November 2009 (UTC)

 Not done No response to queries posted on Apoc2400's talk page inquiring whether or not this is still needed, so assuming a stale request. Re-request if necessary. --Shirik (Questions or Comments?) 09:01, 18 January 2010 (UTC)

Islamic honorifics

- Otterathome (talk) 14:22, 1 November 2009 (UTC)

Not abuse, and may be biting or prevent good-faith editing. Cenarium (talk) 19:08, 15 November 2009 (UTC)
Doesn't seem like a job for the edit filter. Edit and inform others instead. --Apoc2400 (talk) 20:58, 21 November 2009 (UTC)
 Not done per the top of this page, style issues really shouldn't be covered by the edit filter. --Shirik (Questions or Comments?) 07:16, 9 January 2010 (UTC)

Testosterone vs diabetes vandal

ChemNerd (talk) 12:09, 27 December 2009 (UTC)

Of the 23 suspected socks, only 11 of them have ever used the phrase "CANCER VIRUSES IN VACCINE" or the name "Hilleman". I fear that, were this implemented, it would only persuade the editor to change the edit summary, and thus not be productive to justify the execution time. --Shirik (Questions or Comments?) 01:39, 6 January 2010 (UTC)
OK. Thanks for taking the time to look into this. ChemNerd (talk) 12:48, 7 January 2010 (UTC)
 Not done per the above concerns. --Shirik (Questions or Comments?) 23:19, 7 January 2010 (UTC)

Anonymous reverts

  • Filters check every single edit, but cannot make edits, only avoid them. Since reversions of an IP are pretty hard to discover automatically, I suggest you go to WP:ANI instead. - Mgm|(talk) 13:03, 2 December 2009 (UTC)
  • That is incorrect, well, to the extent that your making an assumption that the filter would need to make an edit to do this (when in fact it wont). A filter can be set to block certain actions by certain IPs on certain pages where the action=revert or undo...That would stop this happening.   «l| Promethean ™|l»  (talk) 14:50, 6 December 2009 (UTC)
So it is possible? Geoff B (talk) 21:53, 6 December 2009 (UTC)
The reverts are not similar enough to block them with a single filter. The edit summary will soon change if you test for that. It would be possible to completely block all BT users from editing the pages, but that might be seen as overkill. Testing for ten different article titles and several IP ranges is also likely to have a not insignificant performance hit. So much so, that I'd be disinclined to start writing it. -- zzuuzz (talk) 22:07, 6 December 2009 (UTC)
Sigh, please read the abuse filter documentation. The reverts dont have to be similar, we can block edits where the ACTION is Undo, where the IP ADDRESS is one of those above and the PAGE TITLE matches one of those above. Its apparent that you are not familiar with action based filters. If the filter was written in a very narrowative manner than I imagine the performance hit could be minimised. If a filter is not written then there will be no option but to semi protect 20+ pages, which would be a disappointing outcome.   «l| Promethean ™|l»  (talk) 01:16, 7 December 2009 (UTC)
I'd be interested to see the documentation for action=undo or action=revert, since I was under the impression they only existed as action=edit. After 9 months of editing filters here it's not something I've seen in any other filter. -- zzuuzz (talk) 10:47, 7 December 2009 (UTC)
There is no "undo" action. Imho the only option is a filter that prevents editing from this IP range in this subset of articles (which is defined as a list or by common category if possible). — AlexSm 15:31, 7 December 2009 (UTC)
Whilst it seems rollback has its own action, undo does not which is what this vandal is using. However, something needs to be done and perhaps its worthwhile just preventing the entire range editing those articles. Furthermore, I would recommend filing an abuse report so that we get in touch with the ISP as well.   «l| Promethean ™|l»  (talk) 01:49, 8 December 2009 (UTC)
I think you'll find rollback is also an edit action. An ISP, especially BT, is unlikely to want to get involved in content disputes. What you will need is a list of the titles and a list of the ranges you want to prevent editing those titles, though see my comment above. -- zzuuzz (talk) 01:53, 8 December 2009 (UTC)
Rollback is its own action, for I double checked once the error in my above idea was pointed out. The reason undo isnt is because no additional permission is required to do it. BT will act if told one of they're users is IP hopping to circumvent on-site blocking measures and that failing to do so will see most thier customers affected (for an extended period of time) due to one specific user, my involvment at the IP abuse page has taught me this much. Again, a filter would be the best and easiest resolution
Shall I file an abuse report in the meantime then? Geoff B (talk) 23:51, 11 December 2009 (UTC)
 Not done Per above, this isn't really suitable as an edit filter. --Shirik (Questions or Comments?) 07:27, 6 January 2010 (UTC)

Uncy 2

  • Task: Disallow/warn/tag edits to articles where 'uncyclopedia' and 'Oscar Wilde' are newly introduced. Should apply to all users except admins.
  • Reason: A popular vandal action seems to be replacing articles with articles from Uncyclopedia. This has been observed many times on Special:AbuseFilter/103. If you didn't know, a common trait between most uncyclopedia articles are quotes from Oscar Wilde, see examples of vandalism which it would have been blocked: [16][17][18] (big pages).--

- Otterathome (talk) 14:28, 6 August 2009 (UTC)

Posted proposed code to the talk page of the project. MacMedtalkstalk 05:56, 7 August 2009 (UTC)
Hmm seems that the admins have modified Special:AbuseFilter/103 to do this whilst reducing how effective it is, see Wikipedia_talk:Edit_filter#Special:AbuseFilter.2F103_not_working.--Otterathome (talk) 13:03, 21 August 2009 (UTC)
Stale
 – If this is still needed, post a new request

--Shirik (Questions or Comments?) 06:51, 6 January 2010 (UTC)

Uncy 3

- Otterathome (talk) 16:38, 8 August 2009 (UTC)

Stale
 – If this is still needed, please file a new request.

--Shirik (Questions or Comments?) 06:51, 6 January 2010 (UTC)


Uncy warning

  • Task: Warn users who add the word 'uncyclopedia' anywhere in the article to pages that don't already contain the word. Should also detect any links that contain the word such as uncyclopedia.org and uncyclopedia.wikia.com etc. Should only apply to article, article talk, file, and file talk space.
  • Reason: Commonly linked to petty vandalism, see uncyclopedia article. Also WP:BLP/WP:SPAM problems, see full essay.

- Otterathome (talk) 22:16, 22 March 2009 (UTC)

It is probably in your essay, but would you mind separating out some examples of users who have done this? Dragons flight (talk) 23:45, 22 March 2009 (UTC)
[20][21] are some recent ones. 99.9% of all edits that involve the word are unwanted or are vandalism. And yes, there are plenty in my essay but that is a long essay.--Otterathome (talk) 23:50, 22 March 2009 (UTC)
Why article talk and file talk? Dragons flight (talk) 01:28, 23 March 2009 (UTC)
Because of WP:BLP issues with the site, user_talk space should also be included. My big essay goes in to detail about why it shouldn't be used. Also interwiki spamming of the link cannot be tracked because of Help:Interwiki_linking#Interwiki_links_versus_external_links.--Otterathome (talk) 11:04, 23 March 2009 (UTC)
Created Special:AbuseFilter/103 in log mode. Waiting to see what it hits. Dragons flight (talk) 03:43, 26 March 2009 (UTC)
It appears it only detects users who are not autoconfirmed, could we get this changed to detect everyone?--Otterathome (talk) 17:49, 28 March 2009 (UTC)

Why do we need this filter? I changed the filter to apply to articles only, in addition to the non-autoconfirmed users only restriction. If you want to ban linking to uncyclopedia from user pages, talk pages, user talk pages, etc, you need more consensus than that essay. I would argue that it isn't a common enough vandalism MO to bother with any filter at all. Assuming no further developments, I intend to delete the filter. Prodego talk 15:59, 29 March 2009 (UTC)

It is needed because interwiki links cannot be searched like external links, so there is no other way to keep track of the interwiki's to uncy. So what happens is that it can take over 1 and a half years to remove the spam[22][23]. It is not an outright ban, simply a way to monitor it. It also needs to monitor outside of article space per the reasons in my essay.--Otterathome (talk) 16:50, 29 March 2009 (UTC)
A filter to monitor only would not be awful (although it could be a waste of resources, which examining the log of hits will tell us) but you suggested the filter give a warning. It isn't really acceptable to warn someone just for adding an uncyclopedia link. Prodego talk 17:45, 30 March 2009 (UTC)
So far all article space edits this filter has picked up have been vandalism, can we change this to give a warning now?--Otterathome (talk) 14:45, 1 April 2009 (UTC)
Filed bug 18683: Remove Uncyclopedia from interwiki list. There's no reason for it to have an interwiki prefix; an external link is quite sufficient in the few cases it needs to be linked to. Gurch (talk) 21:17, 4 May 2009 (UTC)
  • Comment I see nothing useful in the creation of this particular filter. Much of what's on Wikia isn't worth linking to from an encyclopædia - many wikis there either are dead/abandoned or serve as little more than fan clubs for individual TV and movie franchises. A link to a wiki about Muppets from any article other than Muppet or a very few related pages makes no more sense than a link to Uncyclopedia from anything other than a page on a humour-related topic (for instance, backronym or Russian reversal). Singling out one wiki out of thousands on that site seems therefore arbitrary and capricious. Perhaps a filter to warn if any more links to [[User:Otterathome/Uncy]] aka [[User:Otterathome/Uncyclopedia]] would be warranted, however, as special:whatlinkshere indicates that particular link spammed to nearly twenty pages - rather many if Wikipedia is an encyclopædia and WP:NOT one user's personal soapbox? --66.102.80.212 (talk) 11:38, 10 September 2009 (UTC)

 Not done The interwiki link for this was removed, so this filter is no longer needed. --Shirik (Questions or Comments?) 06:35, 6 January 2010 (UTC)


Noinclude tags when tagging templates with deletion templates

  • Task: Whenever someone adds a {{db-xxx}} (like {{db-t2}} etc.) or {{tfd}} tag to a page in Template: namespace, the filter should check if <noinclude>-tags are present. If they are, it should output a warning like "Please make sure the tag you placed is inside the <noinclude>-tags". If there are none, it should disallow adding the tag with a message like "Please make sure the tag you placed is enclosed by <noinclude>-tags". It should apply to all users and groups, because we all make mistakes (well, maybe we can expect sysops to know it)
  • Reason: Adding such tags without <noinclude>-tags leads to all pages with the affected template to be listed in CAT:CSD, a situation which will take days of job queue or dozens or hundreds of null-edits to resolve, creating unnecessary workload and incorrect category listings.

- Regards SoWhy 06:27, 1 April 2009 (UTC)

Stale
 – Was done by Filter 135 but that filter has since been deleted.

--Shirik (Questions or Comments?) 06:28, 6 January 2010 (UTC)

Tireless sock creating film pages of non-existing films

  • Task: Detect articles likely created by well-known banned user (Lyle123). This should be easy because I don't care if the filter creates lots of false positives. The pattern is pretty simple. He creates hoax movie articles, mostly about animated or stop-motion movies, cartoons and other kid movies. These are systematically created with a fresh new account. So a filter that checks for new articles by new users and containing, say, the words "movie" or "film" as well as any of "animated", "stop-motion", "Disney" and the like should do it. Like I said, false positives are not a big problem: the pattern is really easy to recognize, in part because the usernames also tend to follow standard patterns (though they change over time). I've stopped tagging every blocked account as a sock (per BEANS and DENY) but for examples of the pattern, check my log of blocks and search for either Lyle123 or StealBoy. There might be better filter ideas.
  • Reason: Tireless vandal. I've stopped counting the number of accounts I and other admins have blocked. Early detection allows rapid checkuser to identify other socks that might have flown under the radar.

- Pascal.Tesson (talk) 15:49, 3 April 2009 (UTC)

I'll try with Special:Abusefilter/129, checking for new pages with some of the infoboxes and templates the user has used recently. I'll keep an eye on it. --Dirk Beetstra T C 16:30, 3 April 2009 (UTC)
Great, thanks. Pascal.Tesson (talk) 18:07, 3 April 2009 (UTC)
  • Suggestion: This user's naming conventions are pretty obvious. He CamelCases a movie studio name along with a non-existent film and follows it up with a string of numbers, usually the "year of release" of that non-existent film. --PMDrive1061 (talk) 23:10, 28 April 2009 (UTC)
Stale
 – Old request that is probably no longer needed. If it is still needed, please make a new request.

--Shirik (Questions or Comments?) 06:20, 6 January 2010 (UTC)

Addition of !s and ?s

  • Task: Warn new users trying to add a phrase in the article namespace containing three or more succeeding exclamation marks (!!!) or question marks (???) if the article does not already contain any.
  • Reason: There is a high probability of the edit constituting vandalism, except in some articles such as this one.

- Mxipp (talk) 20:37, 5 April 2009 (UTC)

Just a comment. You might get false positives when someone is adding a quote. I suppose making an exception for !!!" avoids the problem. Pascal.Tesson (talk) 20:49, 5 April 2009 (UTC)
Take a look at 135. It looks for anything repeated seven or more times. Good enough? —Wknight94 (talk) 16:43, 6 April 2009 (UTC)
Stale
 – No response; Filter 135 seems to handle this sufficiently.

--Shirik (Questions or Comments?) 06:18, 6 January 2010 (UTC)

Newspages with talkheader

  • Task: (This is my first request here and so maybe it is not the best place to request this) Prevent new talk page creating only with {{talkheader}} or {{talkpage}}.
  • Reason: This template is not intended for every talk page and for those that a discussion has already started and many editors are participating causing edit warring, forum-like discussion, etc. -- Magioladitis (talk) 13:56, 8 April 2009 (UTC)
Support this if it also stops people creating talk pages with only a "wikiproject" template on them (I hate those things). Gurch (talk) 17:33, 17 April 2009 (UTC)
Sometimes wikiproject can be usefull. The project knows, for example, with how many articles is dealing with. -- Magioladitis (talk) 16:52, 27 April 2009 (UTC)
I personally strongly oppose this as a talkheader serves as a reminder to those who wish to start a conversation to be civil. The wikiprojects list related projects so interested parties viewing the talk page may find a related wikiproject. In the spirit of remaining neutral, I will provide a possible source code.

Prevent talk header

(article_namespace == 1)
& (article_recent_contributors == "")
& (length(added_lines) < 15)
& ("{{talkheader}}" in lcase(added_lines))

Prevent wikiproject and talkheader

(article_namespace == 1)
& (old_size == 0)
& (length(added_lines) < 110)
& (lcase(added_lines) rlike "{{(talkheader|wikiproject)}}"

Enjoy. Smallman12q (talk) 01:41, 1 June 2009 (UTC)

I won't be neutral. Loads of wikiprojects tag "their" articles with these tags and dozens of editors would have a fit if we enabled this filter. Can't be done.--chaser (talk) 02:39, 2 June 2009 (UTC)

We can enable only the talkheader part. -- Magioladitis (talk) 15:45, 1 September 2009 (UTC)

 Not done Per the above issues. Also, not an abuse issue. --Shirik (Questions or Comments?) 06:17, 6 January 2010 (UTC)

Anti-referral

  • Task: Detect any external links that contain ref= refid= Log mode only, applies only to article space.
  • Reason: Blocks referral marketing links, and alerts users to new links that may be blacklisted.

- Otterathome (talk) 21:18, 11 April 2009 (UTC)

There are some sites that use refid as a type of login token, and as such it may break some links. Will have to see log.Smallman12q (talk) 01:54, 12 April 2009 (UTC)
If such a filter is created, please set it to log-only. Gurch (talk) 17:32, 17 April 2009 (UTC)
Testing Special:AbuseFilter/152. Dragons flight (talk) 01:10, 23 April 2009 (UTC)
Looking at the hits it seems "ref" is quite a common parameter in URL query strings, things like http://www.nytimes.com/2009/05/02/business/02gordon.html?ref=obituaries, that have nothing to do with referral marketing. While the "ref" parameter can be removed from the link and it still works, that does not make inclusion of it malicious. Gurch (talk) 20:08, 2 May 2009 (UTC)
Stale
 – Filter was disabled, but is handled by Filter 152

--Shirik (Questions or Comments?) 06:13, 6 January 2010 (UTC)

CSD tag removal by author

  • Task: Detect any CSD tags being removed by the creator of an article
  • Reason:The creator of an article is only supposed to add a "hangon" tag, never remove the CSD tag itself, because of the creator's inherent bias.—Kww(talk) 03:10, 12 April 2009 (UTC)
That may encourage the users to create sock puppet accounts to remove them. In fact, a user can simply log out and remove the tag to circumvent the filter. I would say that there should be a policy change so only administrators can remove the tag. Anybody, including the author can contest the deletion by using the {{hangon}} tag. -- IRP 22:11, 13 April 2009 (UTC)
While I disagree regards who should be able to removes SD tags, it would certainly be useful to silently log removals - possibly feed them into some sort of IRC feed as well - so they could be fixed. - Jarry1250 (t, c) 09:09, 14 April 2009 (UTC)
First, the instructions in the CSD banner specifically instruct the creator that he is not permitted to remove the tag. Second, the problem really is the edit-warring: when the creator removes the tag to protect the article, usually the nominator slaps it back on, and it cycles. Usually, the creator winds up blocked after an AIV report. Lots of fuss and commotion, and frequently the creator just really doesn't understand what's going on, and the block serves only to piss him off. Preventing the removal in the first place keeps the unpleasant situation from developing in the first place.—Kww(talk) 13:31, 14 April 2009 (UTC)
Only administrators? Uhm... IRP, you are aware that there are more types of contributor to this project than just "admins" and "vandals", right? ...right? :| -- Gurch (talk) 12:43, 14 April 2009 (UTC)
That's why we need a "trusted" user group. -- IRP 21:42, 14 April 2009 (UTC)
No, it's why we need to stop treating all new users as suspicious and harassing them with abuse filters. Gurch (talk) 17:35, 17 April 2009 (UTC)
I can't see any way to implement this without throwing a significant number of false positives. One could create a log for people to sort through manually if that is sufficient. Maybe even a warning if the false positives aren't too common, but I don't think the abuse filter code is adequate for disallowing this with sufficient accuracy. Dragons flight (talk) 17:24, 14 April 2009 (UTC)

Anyone remember Special:AbuseFilter/29? —Admiral Norton (talk) 17:47, 14 April 2009 (UTC)

That doesn't target creators, and doesn't disallow.—Kww(talk) 18:31, 14 April 2009 (UTC)
Yes, because consensus was against disallowing anons, new users, and creators removing CSD tags. –xeno (talk) 21:55, 14 April 2009 (UTC)
I saw new users and anons ... please indicate where there was a consensus to allow creators to remove the tags.—Kww(talk) 04:07, 16 April 2009 (UTC)
There have been times when it is actually justified--when such tags are added for vandalism or harassment. I've seen this. I agree its a problem, but a log or warning is sufficient. DGG (talk) 06:23, 7 June 2009 (UTC)
Stale
 – Closest match is Filter 29, which seems sufficient.

--Shirik (Questions or Comments?) 06:12, 6 January 2010 (UTC)

Page blanking by creator

  • Task: Any time the creator of a page (outside userspace) blanks a page with less than 25 revisions, it should be automatically tagged with {{db-author}}
  • Reason: This way we can maybe stop those overeager rollbackers who revert any blanking on sight, despite blanking being a good-faith way to request G7 deletion and it saves some patroller the time to tag it with the corresponding G7 tag.

- Regards SoWhy 08:37, 16 April 2009 (UTC)

Not possible. The filter can't edit page content, and hence can't add a db tag. It also has no way to count the number of revisions the page has. Dragons flight (talk) 00:35, 17 April 2009 (UTC)
How about using a message that says "to request deletion of a page you created, insert "{{db-author}}" at the top of the page"? -- IRP 22:45, 17 April 2009 (UTC)
IRP's suggestion is a good alternative I think. Could that be done? Regards SoWhy 10:29, 19 April 2009 (UTC)
There isn't a reliable test for page creator. One can test for whether someone is a page's only editor. Is that good enough to be useful? Dragons flight (talk) 00:40, 23 April 2009 (UTC)
Unfortunately, the usual scenario is the page creator blanking the page after it gets tagged for deletion by someone else. While it would still be useful to have such a message when the only editor blanks a page, it would not be the most usual scenario. I'd be happy if you wrote a filter for that situation as well, if it's not too time consuming. After all, it would help at least some new users and that's never a bad thing. Regards SoWhy 09:01, 24 April 2009 (UTC)
Stale
 – I did not search the deleted filters, but did not see it in non-deleted ones. If this is still needed, please make a new request.

--Shirik (Questions or Comments?) 06:09, 6 January 2010 (UTC)

Shmede/Brown Town vandals

  • Task: Here's an easy one. A recurrent group of abusive vandals keeps creating pages about a couple of guys nicknamed "Shmede" and "Brown Town" respectively. Both of these strings are implausible additions to an article so the filter will get very few hits. No action needed, the filter log will be sufficient.
  • Reason: As noted above, recurrent problem. Past accounts have been abusive to admins deleting the stuff and one of them even hijacked an account. I can provide details if anyone is interested. Pascal.Tesson (talk) 16:19, 28 April 2009 (UTC)
Stale
 – No indication that this is still needed. If so, open a new request.

--Shirik (Questions or Comments?) 05:54, 6 January 2010 (UTC)

Heart and Swastika Blocker

  • Task: To block the random insertion of ♥ ♥ ♥ ♥ and 卐卐卐into music/song/art/porn star articles
  • Reason: Occassional fans decide to spam an article with hearts. I can't see why several hearts in a row would be needed in an article.

- Smallman12q (talk) 23:31, 30 April 2009 (UTC)

Hearts (suit)? Xclamation point 01:30, 1 May 2009 (UTC)
Examples? Dragons flight (talk) 18:06, 2 May 2009 (UTC)
Can we do a 12 hour log only run to see how many we get?Smallman12q (talk) 20:08, 4 May 2009 (UTC)
I've added 卐 because it is appearing more frequently.Smallman12q (talk) 01:27, 12 May 2009 (UTC)
Stale
 – The filter may have been created before, but looks like it was deleted because I can't find it (I didn't search the deleted filters). If it's still needed, open a new request.

--Shirik (Questions or Comments?) 05:52, 6 January 2010 (UTC)

Arabian Gulf vandalism

  • Task: Block vandalism related to the Persian Gulf naming dispute. I figure that blocking any edit adding the words "fake", "false", "true", "incorrect", "correct", "wrong", or "does not exist" to Arabian Gulf, or any edit adding those words or the word "Persian" to Arabs Gulf should stop most of it.
  • Reason: This vandalism makes up >95% of the editing history of Arabian Gulf, despite extended periods of protection.

--Carnildo (talk) 00:25, 3 May 2009 (UTC)

  • That seems like it would generate a heck of a lot of false positives and increase edit time on the {{NUMOFARTICLES}}-1 other articles in the encyclopedia. Protection is probably the lesser of two evils here. Stifle (talk) 11:58, 6 May 2009 (UTC)
    • Would it really generate that many false positives? A filter that blindly denied all edits to those two articles would have a false-positive rate of less than 5%. By filtering on words that are unlikely to occur in non-vandalism edits to those article, it should reduce the false-positive rate. I'd rather not permanently full-protect those articles: in the four years I've been watching Arabian Gulf, there have been productive edits -- just not many. --Carnildo (talk) 22:52, 6 May 2009 (UTC)
      • I read up on performance issues on the abuse filter and this might not be as bad as I thought it would. I can't write regexen for toffee though. If you can, then add it (just make sure that the check for the article title is the first clause). Stifle (talk) 14:01, 7 May 2009 (UTC)
Stale
 – After a look at the current state of the article in question, this filter does not seem needed anymore. If it's still needed, make another request.

--Shirik (Questions or Comments?) 05:38, 6 January 2010 (UTC)

Pete the Flying Cat

  • Task: Prevent insertion of "Pete the Flying Cat"
  • Reason: Multi-IP vandalism over the past several months ([24]), placing this fake game title in lists. I was alerted to this on WP:AIV ([25]). Subsequently I removed several of these fake entries that had remained for a month or more.

- Evil saltine (talk) 18:57, 15 May 2009 (UTC)

Stale
 – Unlikely that this still needs attention. If it does, open a new request.

--Shirik (Questions or Comments?) 05:28, 6 January 2010 (UTC)

Blank image description pages

Users should be prevented from uploading an image with an entirely blank description page, or perhaps one that lacks any template calls whatsoever. These images are invariably tagged {{subst:nsd}} and deleted. Even {{Don't know}} should be accepted though. Stifle (talk) 11:50, 25 May 2009 (UTC)

Alas, but it seems that uploads are simply ignored by AbuseFilter. — AlexSm 21:59, 26 May 2009 (UTC)
 Not done Not possible, per above. --Shirik (Questions or Comments?) 05:26, 6 January 2010 (UTC)

What exactly is this filter trying to accomplish? It matched me listing a template for deletion. In the way that you're supposed to. Can it be fixed to do something useful or deleted please? Gurch (talk) 17:28, 3 June 2009 (UTC)

[26] I think it's looking for tfd tags that aren't noincluded. –xenotalk 17:52, 5 June 2009 (UTC)
WP:AF/FP is second door down the hall on the right. Stifle (talk) 17:22, 6 June 2009 (UTC)

 Not done Not a filter request. --Shirik (Questions or Comments?) 05:22, 6 January 2010 (UTC)

Disallow creating subpages of User:Example

User:Example is an account set up to be usable as an example (it's a real registered account that doesn't do anything). Because it's guaranteed not to be a real user, various templates and instructional text use User:Example as a stand-in where the name of another user would normally appear. This has the side effect that new users attempting to start a sandbox article in their own userspace sometimes end up creating it as a subpage of User:Example instead. In many cases, the creating user then can't even find the page they created. There are legitimate subpages of User:Example, such as User:Example/sub page, but most pages created here are errors. Gavia immer (talk) 05:24, 13 August 2009 (UTC)

Are there that many subpages created to warrant a filter? Or perhaps a bot could do a once a week cleanup? Often times, new users are creating these pages, and a filter may confuse them even more.Smallman12q (talk) 12:09, 13 August 2009 (UTC)
It varies. The worst-case failure involves some new user repeatedly creating the same article under variant titles on multiple subpages, because they can't keep track of it, can't understand how to find it, and don't get why "it keeps disappearing". However, this is the least important of my three requests; if anyone can point me towards a bot that could manage these, that would be fine. Gavia immer (talk) 22:56, 13 August 2009 (UTC)
Put in a bot request. It's really a simple task for a bot. (Check the article's listed in userspace against a whitelist->delete one's that don't match->notify page creators.)Smallman12q (talk) 14:25, 14 August 2009 (UTC)
 Not done due to the concerns above. The overhead isn't warranted. --Shirik (Questions or Comments?) 05:18, 6 January 2010 (UTC)
This could be done via the title blacklist. -- Soap Talk/Contributions 02:02, 16 January 2010 (UTC)

Warn when creating or moving pages to a title that starts with a slash

When moving a draft page out of userspace, a very common error is to simply delete the username from the destination, leaving the / character in place, and then move the page to an incorrect title beginning with a slash. Less often, new pages get created at titles beginning with a / or \. This is generally an error, but not universally so - /dev/null and /b/, among others, are perfectly legitimate uses. Since it's generally an unintentional error, a warning like "Do you really want to create this title?" ought to be sufficient to discourage such creations without affecting false positives at all. Gavia immer (talk) 05:24, 13 August 2009 (UTC)

Seem like a good idea, perhaps we should have a filter to catch odd titles such as %lol@.
/*Catches all moved/created article that begin with / or \ */
(action == "create")
| (action == "move")
& (article_text rlike "^(/|\\)")

Smallman12q (talk) 13:01, 13 August 2009 (UTC)

Poke, poke - any update on implementing this? Gavia immer (talk) 15:58, 1 November 2009 (UTC)
Users notice when they make the mistake; they are visible enough and deleted quickly, [27], [28]. So it doesn't seem necessary. Cenarium (talk) 19:08, 15 November 2009 (UTC)
There's no particular reason for you to know this, but those listings are clean because I clean them out; I believe that I'm the only one doing it. I'm trying to find a solution that doesn't depend on fragile old me. Gavia immer (talk) 18:38, 18 November 2009 (UTC)
 Not done Stale request, but I share the same concerns as User:Cenarium -- it is extremely rare and should be visible and easily corrected. The performance hit isn't warranted. --Shirik (Questions or Comments?) 05:11, 6 January 2010 (UTC)

Second person or you

  • Task: This filter would indicate the addition of the word you (your, yours, you're, you'll...) on an article page. These edits would be allowed to be saved, perhaps after being asked to visit WP:YOU, but the summary would notify other users that the second person has been introduced. (I recommend looking for you followed by a wildcard to catch other forms, and preceded by a space to exempt words like layout.)
  • Reason: Except for quotes, use of the second person is frowned upon and such edits will need to be reworded, either by the saving editor or another supervisor. Moreover, I have found that many anons will insert good faith material phrased with you, where this wording is only one of many issues. In both cases, the attention of an experienced user is required, and this tag would give heads-up that an edit may need attention.

- HereToHelp (talk to me) 20:24, 22 December 2009 (UTC)

 Not done This isn't really an abuse issue, it is a style issue, and thus is not a suitable candidate for a filter. --Shirik (Questions or Comments?) 04:42, 6 January 2010 (UTC)

Addition/removal of featured article / featured list templates

  • Task: Flags non-bot edits in which {{featured article}} or {{featured list}} is added to a page.
  • Reason: GimmeBot (talk · contribs) adds and removes featured article/list stars to pages, so there is no reason for individual editors to do so. Every so often, new editors will add one of these templates accidentally because they copy-pasted a featured article/list into their draft of a page, or they will add one because they do not know that there is a process by which articles/lists become featured. This filter would help with monitoring misuse or inappropriate remoals of those templates. Please ping me if there are problems and if this filter is created. Thanks, Dabomb87 (talk) 01:36, 13 September 2009 (UTC)
 Not done - I'm not sure it's large enough of a problem to warrant a filter. Each filter takes time to process, even for edits which don't fall under its realm. Basically, if there's not at least a few hits per day, it's not worth a filter. King of 03:06, 13 September 2009 (UTC)
Alright, thanks anyway. Dabomb87 (talk) 03:08, 13 September 2009 (UTC)

Warning users about signing posts

Not practical, we can't be sure an addition is indeed a new comment, and too bitey. Cenarium (talk) 15:39, 1 September 2009 (UTC)
Also, sine-bot does a good job of catching these. Skier Dude (talk) 04:46, 3 September 2009 (UTC)

Change to a referenced number without a reference change

  • Task: Flag edits to referenced numerals by new/IP users without a corresponding change to the reference. In other words, when a user changes a number in a sentence that is supported by a reference tag, without changing the corresponding reference tag.
  • Reason: Probable introduction of a deliberate factual errors. If the filter is overambitious, could be applied only to edits made without an edit summary. I can help develop a reg ex for this pattern if necessary (I'm unfamiliar with the abuse filter syntax, but I can work with POSIX regexs)

- Shadowjams (talk) 06:42, 5 June 2009 (UTC)

I'd like to see a suggested syntax. I have doubts over whether this can be practically accomplished with the abuse filter. Dragons flight (talk) 01:49, 6 June 2009 (UTC)

Dead Image detector

  • Task: To prevent the insertion of nonexistent images by ips.
  • Reason: Vandals often do this. I'm not sure if a function exsists that will allow for this filter, but at least I tried.=D

- Smallman12q (talk) 01:29, 2 June 2009 (UTC)

I don't think this is a good idea to prevent the edit, as some users might be trying to insert a valid image, but accidentally mistype a valid file name, with a invalid one (I sometimes accidently type a invalid one myself). Perhaps we can display an information message such as

Warning! The file you have attempted to add does not seem to exist.

  • Make sure you have typed the name in the brackets correctly.
  • Check the deletion log for the file, to make sure the file hasn't been deleted.
  • If you do not know the name of the you wish to add, use the search function.
  • If you still have problems feel free to ask any Wikipedian.

I think the above (I'm no artist) would be less bitey, and more useful then simply preventing the edit. Vivio TestarossaTalk 23:17, 10 June 2009 (UTC)

I think it can be done using new_html.Smallman12q (talk) 02:00, 10 August 2009 (UTC)

Recently reverted user

  • Task: Check the edit summaries of recent contributers for the user name and the word "reverted"
  • Reason: Edits made by new users who have been recently reverted by other experienced users in the same article are more likely to be troublesome. I don't know if the edit filter could check the edit summaries of recent contributers.

- Sole Soul (talk) 01:03, 27 November 2009 (UTC)

No thanks, Too broad and would trip on allot of edits for little or no reason. I also note that the idea that new users are more likely to be trouble makers is a sweeping judgement.   «l| Promethean ™|l»  (talk) 14:56, 6 December 2009 (UTC)
I didn't say that new users are more likely to be trouble makers, I said "New users who have been reverted by other experienced users are more likely to be trouble makers", this is purely about probabilities and implies that some experienced editors are also trouble makers, although I could've phrase it better. I agree with you that it would trip on a lot of edits, but that is not necessary a bad thing if most of them are true positive. This idea can be refined more narrowly by combining it with other criterias if necessary. That is more smart than dismissing it altogether. Sole Soul (talk) 19:20, 6 December 2009 (UTC)
To be more clear about how it would work:
  • New User A make an edit.
  • User B with rollback right revert the edit with an edit summary that contain User A name and the word "Reverted".
  • User A repeat the edit.
  • The Edit Filter detect the edit by checking the edit summaries of recent contributers, in this case the edit summary of User B. Sole Soul (talk) 22:28, 6 December 2009 (UTC)
In any case, this is not possible with Edit Filter. Could be done with a bot. — AlexSm 04:36, 7 December 2009 (UTC)
Hello, this is currently not technically possible to do with the abuse filter. Triplestop x3 19:59, 13 December 2009 (UTC)
Thank you. I suspected that. Sole Soul (talk) 20:05, 13 December 2009 (UTC)
 Not done Infeasible request. --Shirik (Questions or Comments?) 04:18, 6 January 2010 (UTC)

Preppy vandal filter

  • Task: Warn and log when stings such as "hi!," ":)," "(:," "=]," "♥," "&hearts," "is awsome," "i love," "<3," and any similar strings that come to mind are added to the main article space.
  • Reason: I've been seeing a lot of people write these string in articles related to names, schools, school/youth activities (especially articles related to cheerleading), work places, and sometimes just random articles. For the most part, I find no legitimate reason for someone to write hearts, smiles, or I love... in an article. It actually surprises me that we didn't already have a filter to reduce the occurance of this type of test edit.

- PCHS-NJROTC (Messages) 19:30, 15 October 2009 (UTC)

  • Is this a widespread problem? The edit filter adds a fair amount of overhead to each time a page is saved, so if it's just a couple times a week I don't see it as constructive. Stifle (talk) 15:44, 20 October 2009 (UTC)
    • Yes, it is indeed a wide spread problem; there are articles all over Wikipedia that see this "cheerleader-like vandalism" as I call it. Those are just ones that see it more than others, but you see this in all kinds of articles. I should have mentioned school subjects as well; it's not uncommon to see this articles related to science, mathematics, history, literature, politics, etc. It seems to be something that little kids and preppy girls like to do not to single anybody out or anything. Something else you might consider is making the filter check for things like "i like pie" as there seems to be a lot of that going around as well. PCHS-NJROTC (Messages) 18:07, 20 October 2009 (UTC)
And cake. Don't forget cake. We already have a "poop" filter, don't we? These could probably go in there. Wknight94 talk 18:19, 20 October 2009 (UTC)
I think I heard that the cake and pie stuff is related to 4chan. PCHS-NJROTC (Messages) 18:41, 20 October 2009 (UTC)
Anyone still reading this? PCHS-NJROTC (Messages) 23:41, 1 November 2009 (UTC)
False positives would be the biggest concern here. We already have a dozen or so exceptions in the "poop" vandalism filter alone because of them. And we're constantly battling the performance limits of this tool. So adding 6 or 7 more silly words with a dozen exceptions per silly word is probably not worth the trouble. Wknight94 talk 03:18, 2 November 2009 (UTC)
This may be true with "is awesome," "i love" and the hearts, but how could a filter on :), =], and HI! trigger a lot of false positives? There's not too many articles that I can think of where a heart would be legitimate either, but I'd be cautious about a filter on "<3" as that could catch people adding "greater than three." PCHS-NJROTC (Messages) 20:20, 2 November 2009 (UTC)
"SHANGHI!"? Various programming languages might use :) and =]. It's just not worth the aggravation. I wouldn't have thought "poop" would cause so much concern, but some people get irate at false positives. Wknight94 talk 21:31, 2 November 2009 (UTC)
Lots of politics (sigh). Probably better just to keep reverting, blocking, and sending abuse reports; false positives a menace. PCHS-NJROTC (Messages) 01:20, 5 November 2009 (UTC)

 Not done As discussed, this is pretty easily detected with current anti-vandalism tools and use of a filter hasn't been able to be justified. --Shirik (Questions or Comments?) 07:54, 6 January 2010 (UTC)