Community Wishlist Survey 2015/Moderation and admin tools/arz

Voting has CLOSED! Thanks for your votes!

Better history pages

History pages are a key tool for article maintenance! Some possible improvements:

• More reliable layout; clearer divide between rows, and/or better wraparound behaviour
• Better separation of data from actions. "Data" includes revision links, timestamp, user info, edit summary, tags, et cetera. Actions include rollback, undo, thank, et cetera.
• Fewer mildly-cryptic things that might be confusing to newbies. For example: "cur" and "prev" links aren't self-explanatory as "diff with current revision" and "diff with previous revision", respectively.
• Visual representations of data. For example, graphical links between net-null revisions (usually between a revert edit and the revision to which it reverted). The key idea here is "information on history pages that doesn't require reading words or numbers", so that it's easier to understand a page's history at a glance.

I've played around with some of this already using JavaScript; interested parties can paste importScript("User:Nihiltres/nothingthree.js"); and then nothingthree.customRevs.testRun(); into the console of an English Wikipedia history page to see how my experiments ended up. For example, I make the byte-difference more self-explanatory by changing "(65,176 bytes) (+1,234)" into "+1,234 → 65,176 bytes".

I stopped messing around a) because it was tiring and b) because this is something that should be implemented in MediaWiki proper, rather than reimplementing the whole damn history page in JavaScript. Maybe something the Community Tech team would like to take a run at?

{{Nihiltres|talk|edits}} 20:25, 17 July 2015 (UTC)

Earlier discussion and endorsements
I've wanted for a long time some visual representation of the history inspired by the lines on the left of e.g. git histories. That would be something very interesting to have. Helder 12:23, 10 November 2015 (UTC)  Endorsed for voting. This, that and the other (talk) 08:28, 17 November 2015 (UTC)

الاصوات

1.  Support Ckoerner (talk) 17:07, 1 December 2015 (UTC)
2.  Support --Usien6 (talk) 19:59, 1 December 2015 (UTC)
3.  Support although this should run as a beta before full rollout, as I'm sure there would be lots of feedback and differing opinions on various aspects. It's possible that if there is too much consternation over the changes, that this may work best as a gadget. We'll see. Stevie is the man! ^{Talk • Work} 22:49, 1 December 2015 (UTC)
4.  Support Casliber (talk) 13:32, 2 December 2015 (UTC)
5.  Support This is a really fundamental feature of Wikipedia that needs some people to think about what they can do to make it better. But be real sure you have community support before actually implementing it. Wnt (talk) 23:32, 2 December 2015 (UTC)
6.  Support Pengo (talk) 01:58, 3 December 2015 (UTC)
7.   I support most of this proposal, but I strongly oppose replacing "cur" with "diff with current revision". Hovertext would be OK, but please don't use up real estate with needless repetition. YBG (talk) 06:26, 3 December 2015 (UTC)
8.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
9.  Support Kvardek du (talk) 10:34, 5 December 2015 (UTC)
10.  Support all of the above are great suggestions. --Waldir (talk) 15:12, 8 December 2015 (UTC)
11.  Support Matěj Suchánek (talk) 21:04, 9 December 2015 (UTC)
12.  Support Alkamid (talk) 22:26, 13 December 2015 (UTC)

Enhanced per-user, per-article protection / blocking

Tracked in Phabricator:
Task T2674

There are currently two mechanisms to stop a user editing an article on Wikipedia - protect the article so nobody can edit it, or block the user so they can't edit anything. That's it.

I really feel quite strongly that we need something in between those two extremes - the ability to protect or block a specific user from a specific article. It would allow topic bans to be enforced technically, and prevent editors from going towards a specific area of disruption while still accommodating them as much as possible elsewhere. We need editors, and sometimes we just have to take what editors we're given, and manage the disruptive elements.

Take a look at ANI on the English Wikipedia or read any discussions that involve a block and see how much drama that is - anything that can reduce that is welcome in my view. I've hacked a local installation of MediaWiki I maintain to allow "automatic G7" ie: any user (not just admins) can delete any page that they created, so I'm sure it's technically doable.

Ritchie333 (talk) 13:02, 29 May 2015 (UTC)

Earlier discussion and endorsements
Occasionally one both needs to both protect the article and than continually block new socks as they are created :-) It is to easy to get around this. You just keep creating socks. Getting them auto confirmed and editing around the semi protection. We have a few people who do this. Not sure what the solution is. Doc James (talk · contribs · email) 13:20, 29 May 2015 (UTC)  Endorsed +1. We are currently using AbuseFilter for personal restrictions, but native solution would be better --AS (talk) 13:50, 9 June 2015 (UTC)  Endorsed This would be a useful tool in the administrator's arsenal, and be fairer to involved participants who want to edit other pages and uninvolved ones who want to edit the page in dispute. Fluffernutter (talk) 17:15, 11 November 2015 (UTC)  Endorsed This would give a much less bitey way to deal with edit warrers WereSpielChequers (talk) 21:20, 11 November 2015 (UTC)  Endorsed--Shizhao (talk) 07:55, 12 November 2015 (UTC)  Endorsed - This should definitely be coded. Whether it is turned on for a given Wiki should be a local decision. Davidwr/talk 05:56, 16 November 2015 (UTC)  Endorsed Yes, this is long overdue... IJBall (talk) 03:36, 17 November 2015 (UTC)  Endorsed 4nn1l2 (talk) 23:51, 18 November 2015 (UTC)  Endorsed on itWiki abusefilter are in my experience not particularly efficent. This tool for these situations is definitely better.--Alexmar983 (talk) 18:28, 24 November 2015 (UTC)

Votes

1.  Support 4nn1l2 (talk) 03:16, 30 November 2015 (UTC)
2.  Support MER-C (talk) 09:45, 30 November 2015 (UTC)
3.  Support Samwalton9 (talk) 10:32, 30 November 2015 (UTC)
4.  Support IJBall (talk) 13:54, 30 November 2015 (UTC)
5.  Support NE Ent (talk) 14:05, 30 November 2015 (UTC)
6.  Support
   ⋙–Berean–Hunter—► ((⊕)) 15:38, 30 November 2015 (UTC)
7.  Support - And also a system to block a user except a small, pre-defined list of pages. This would certainly be useful in many cases on English Wikipedia. עוד מישהו Od Mishehu 16:18, 30 November 2015 (UTC)
8.  Support Tryptofish (talk) 18:11, 30 November 2015 (UTC)
9.  Support. --Stryn (talk) 19:11, 30 November 2015 (UTC)
10.  Support - All-or-nothing blocks are too binary. In my experience on other websites, granular blocking is a much better way to police behaviour.Jo-Jo Eumerus (talk) 20:14, 30 November 2015 (UTC)
11.  Support Matiia (talk) 20:28, 30 November 2015 (UTC)
12.  Support Orlodrim (talk) 20:30, 30 November 2015 (UTC)
13.  Support --Grind24 (talk) 20:49, 30 November 2015 (UTC)
14.  Support This is currently possible, but only with the abuse filter, and it would be good if we could do this without requiring the abuse filter to run on every edit by everyone. Nyttend (talk) 21:52, 30 November 2015 (UTC)
15.  Oppose This will only encourage more socking. I do believe we need more tools in this area, but I'd prefer to see tools which focus on the content rather than the editor, or auto-blocks based on previous content by the editor (using ORES). John Vandenberg (talk) 02:09, 1 December 2015 (UTC)
16.  Support. I think this would give much needed flexibility in dealing with problems, and might even decrease the amount of work at an/i and arbcom. DGG (talk) 02:21, 1 December 2015 (UTC)
17.  Oppose Per-topic bans already solve this problem and they are more effective.--Isacdaavid (talk) 02:27, 1 December 2015 (UTC)
18.  Support--Shizhao (talk) 09:40, 1 December 2015 (UTC)
19.  Support--Kimdime (talk) 12:18, 1 December 2015 (UTC)
20.  Support--Максим Підліснюк (talk) 14:48, 1 December 2015 (UTC)
21.  Support tufor (talk) 15:12, 1 December 2015 (UTC)
22.  Support - Whaledad (talk) 15:15, 1 December 2015 (UTC)
23.  Support --AS (talk) 15:33, 1 December 2015 (UTC)
24.  Support Goombiis (talk) 16:43, 1 December 2015 (UTC)
25.  Support If this system will be easy for learning, I'll support it. Maybe there should be an option that allow sysops to block user edit all articles in category and it's subcategories and an option for block user edit in all articles in one namespace or allow only one namespace (for example user can edit talk pages but he cannot edit articles). --Urbanecm (talk) 17:38, 1 December 2015 (UTC)
26.  Support--Alexmar983 (talk) 18:30, 1 December 2015 (UTC)
27.  Strongly oppose --Usien6 (talk) 20:02, 1 December 2015 (UTC) // The en:Wikipedia:Edit filter already does the job.

    We want something that does this with a lower performance impact, so that we can use more of them and free up the scarce edit filter resources for something else. MER-C (talk) 20:25, 1 December 2015 (UTC)

28.  Support Great idea! --I am One of Many (talk) 20:28, 1 December 2015 (UTC)
29.  Support Good idea, but I worry whether socks will make this useless. StevenJ81 (talk) 22:31, 1 December 2015 (UTC)
30.  Support for the flexibility this gives to administering Wikimedia projects. I understand the sock issue, but this proposal doesn't seem to be trying to solve that. At the same time, if a user is blocked only from editing particular articles or topic areas, they could just as easily decide to work on other parts of the wiki than socking to return to articles they were blocked from editing. Also, I like that this reduces pressure to implement full blocks and demonstrates we're a kinder site that doesn't levy ultimate punishments for narrow crimes. I can also see this being used for 3RR blocks. Stevie is the man! ^{Talk • Work} 23:02, 1 December 2015 (UTC)
31.  Support - Yes, please! RonnieV (talk) 11:23, 2 December 2015 (UTC)
32.  Support --β₁₆ - (talk) 11:56, 2 December 2015 (UTC)
33.  Support Technical enforcement of topic bans would be useful and would lighten the load of those editors who help make sure those bans are enforced.  DiscantX 12:31, 2 December 2015 (UTC)
34. Comment This needs to be expanded for all pages and not just articles. Please see Jimbo's response here.
    ⋙–Berean–Hunter—► ((⊕)) 12:49, 2 December 2015 (UTC)
35.  Support Casliber (talk) 13:33, 2 December 2015 (UTC)
36.  Oppose, Special:AbuseFilter already does this — NickK (talk) 15:18, 2 December 2015 (UTC)
37.  Support Fluffernutter (talk) 15:51, 2 December 2015 (UTC)
38.  Support, though nervously. If admins/arbitrators could be persuaded to actually tailor their topic bans to a pre-agreed enumerated list of articles, then people who are topic-banned could actually avoid being sucked down the drain by detractors waiting for a gotcha. OTOH I fear this may be prone to be used like a "less than lethal weapon", i.e. on anything that moves because hey, it's not banning everything. Still, it is interesting enough that it should be researched. Note ultimately its viability as a tool depends on the level of self-control that administrators can manage, rather than the people blocked. Wnt (talk) 23:30, 2 December 2015 (UTC)
39.  Support although slightly skeptical of when this would be used. JQTriple7 (talk) 23:33, 2 December 2015 (UTC)
40.  Support Rzuwig^► 09:31, 3 December 2015 (UTC)
41.  Support on condition that it would also apply to all "confirmed sockpuppets" as I could see users creating a new account to go around the protect/block. Krett12 (talk) 16:12, 3 December 2015 (UTC)
42.  Support Though I think this could be done with an improved edit filter, eg edit filters attached to individual users, or pages so as to seriously cut the performance impact of running for all. Graeme Bartlett (talk) 04:35, 4 December 2015 (UTC)
43.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
44.  Support GY Fan★ 11:28, 4 December 2015 (UTC)
45.  Support Graham87 (talk) 11:33, 4 December 2015 (UTC)
46.  Support Bináris ^{tell me} 18:34, 4 December 2015 (UTC)
47.  Support --H4stings (talk) 14:27, 7 December 2015 (UTC)
48.  Support --Waldir (talk) 15:13, 8 December 2015 (UTC)
49.  Support If properly implemented, this could really cut down on both AN/I drama and the ArbCom's caseload. Daniel Case (talk) 19:23, 8 December 2015 (UTC)
50.  Support - Bcharles (talk) 00:37, 9 December 2015 (UTC)
51.  Support Lsanabria (talk) 15:23, 10 December 2015 (UTC)
52.  Support AlbinoFerret 18:27, 10 December 2015 (UTC)
53.  Support as nominator. In all seriousness, I thought this proposal was a perennial one that would be tossed into the bit bucket! Regarding some of the opposes - I appreciate this won't necessary solve socking, but I don't think any tool can solve every problem ever. It is indeed technically possible to use the edit filter to do some user / page level blocking, but IMHO the user interface (or rather the absence of it!) isn't enough to make it safe for general-purpose use. Admins do not need to be programmers. Ritchie333 (talk) 13:35, 11 December 2015 (UTC)
54.  Support --Aervanath (talk) 20:23, 12 December 2015 (UTC)
55.  Support --Sphilbrick (talk) 15:35, 13 December 2015 (UTC)
56.  Support Alkamid (talk) 22:27, 13 December 2015 (UTC)
57.  Support --MisterSanderson (talk) 04:36, 14 December 2015 (UTC)
58. Neutral This proposed tool is a blade, metaphorically: it is a two-handed-vorpal-flaming-broadsword when a rouge admin hands out dozens of rapid-fire page-blocks to "calm" things down (cf: backfire), but when used as a scalpel, by surgically precise admins, it is less bitey to page-block for 31 hours than to enWiki-block for 31 hours, because the admin can block edit-warriors solely from the page that was the locus of the disruption, and not from enWiki entirely. Upside: proportionate, and therefore preventative-not-punitive. Upside: edit-war participants can go edit some other page. Downside#1: page-block will halt edit-warriors on THAT page, but if the dispute is hot enough, the edit-warriors will simply migrate the dispute to another page, such as usertalk (discussion is fine but hurling invective is not). Downside#2: unintended consequence, will encourage more socking. (Ritchie333: the worry is not that page-block fails to SOLVE socking, the worry is that page-block will tend to generate MORE socking than the regular enWiki-block does... BECAUSE it is a surgical block, it is hypothesized to be more likely to encourage socking "to get around that biased admin" ... i.e. the blockee will see an page-block as "that admin tried to keep me from putting the truth into enWiki" whereas the blockee will *more often* see enWiki-block as "that admin does not want me to edit-war anywhere.) Downside#3: again with the unintended consequences, blockees will perceive corruption where admins are page-blocking in order to control content (in the eyes of the blockee at least), rather than halting edit-warring. My even-more-extensive comments are hidden here. Some implementation-suggestions, *if* this is decided to be a wise feature to implement, are hidden here. In any case, although I can see value in this proposal, there are a lot of risks here. Please be exceedingly careful, if you do implement this stuff. 75.108.94.227 12:15, 14 December 2015 (UTC)

Improve date range searches on Special:Contributions

Tracked in Phabricator:
Task T120733

There is no good way to search a window of time in Special:Contributions. Currently, you can set a date but then you get from that date AND ALL earlier contribs which can be too many to sift through. It would be good to refine the date searches from Point A in time to Point B in time. As an example, we should be able to search the last three months and ONLY the last three months. Every editor and admin that hunts socks, spammers, paid editors, long term abusers, etc. would benefit from this as it allows them to refine their searches for relevance. At the present, very few editors are manipulating the search strings in the URLs to force time windows but it is very cumbersome. You add a string that matches this pattern: " ?ucstart=yyyymmddhhmmss&ucend=yyyymmddhhmmss " Please see this example of URL string manipulation and the tail end of this thread to see that folks have been looking for this. Please modify the queries and front end of this interface to have start and end dates so that we may search time windows.

⋙–Berean–Hunter—► ((⊕)) 23:26, 9 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed This would be nice. This, that and the other (talk) 01:06, 10 November 2015 (UTC)  Endorsed This would be very useful at RFA and might even focus people's attention on the recent edits of the candidate rather than dragging in ancient stuff that no longer applies. WereSpielChequers (talk) 19:38, 10 November 2015 (UTC)  Endorsed. This is an annoyance with core software that can be easily addressed. MER-C (talk) 21:52, 10 November 2015 (UTC)  Endorsed Fluffernutter (talk) 17:24, 11 November 2015 (UTC)  Endorsed Heck, yeah! In addition, they need to change the search functionality on Special:Contributions so you can at least search BY DAY (if not by hour!) and not "by month" (which is ridiculous!) IJBall (talk) 03:40, 17 November 2015 (UTC) I agree that it is good to have this improved granularity in time range control. I should point out that anyone evaluating IP ranges for blocks would likely be interested in searching through only the last three months to understand the potential for collateral damage. These changes would allow for filtering out all of those edits from previous years. ⋙–Berean–Hunter—► ((⊕)) 16:34, 17 November 2015 (UTC) Comment These improvements are scalable to all page histories (articles, talk pages, etc.) as they are currently constrained in the same way (example). I hadn't thought of that when I first posted but adding time window searching to those would also be an improvement and anywhere that form of query may be found. ⋙–Berean–Hunter—► ((⊕)) 17:26, 17 November 2015 (UTC)  Endorsed And similar improvements to History and Special:Log pages as well if feasible. the wub "?!" 01:02, 25 November 2015 (UTC)

Votes

1.  Support Jenks24 (talk) 10:40, 30 November 2015 (UTC)
2.  Support Should take less than a day, too. MER-C (talk) 11:02, 30 November 2015 (UTC)
3.  Support NE Ent (talk) 14:05, 30 November 2015 (UTC)
4.  Support, but as per my Endorsement comments the search "range" function should also be improved from a "per month" value to a "per day" (at least!) value (i.e. improved "granularity"). IJBall (talk) 14:12, 30 November 2015 (UTC)
5.  Support as proposer.
   ⋙–Berean–Hunter—► ((⊕)) 15:37, 30 November 2015 (UTC)
6.  Support. --Stryn (talk) 19:11, 30 November 2015 (UTC)
7.  Support — Bilorv (talk) 19:57, 30 November 2015 (UTC)
8.  Support Orlodrim (talk) 20:30, 30 November 2015 (UTC)
9.  Support Dalba 20:44, 30 November 2015 (UTC)
10.  Support --Grind24 (talk) 20:49, 30 November 2015 (UTC)
11.  Support An easy win. John Vandenberg (talk) 02:14, 1 December 2015 (UTC)
12.  Support Mahdy Saffar (talk) 06:52, 1 December 2015 (UTC)
13.  Support--Alexmar983 (talk) 09:53, 1 December 2015 (UTC)
14.  Support --Steinsplitter (talk) 11:27, 1 December 2015 (UTC)
15.  Support. Mathonius (talk) 14:57, 1 December 2015 (UTC)
16.  Support · · · Peter (Southwood) ^(talk): 15:01, 1 December 2015 (UTC)
17.  Support Sadads (talk) 15:55, 1 December 2015 (UTC)
18.  Support Goombiis (talk) 16:44, 1 December 2015 (UTC)
19.  Support Ckoerner (talk) 17:09, 1 December 2015 (UTC)
20.  Support--Calak (talk) 18:30, 1 December 2015 (UTC)
21.  Support.--Nahum (talk) 19:35, 1 December 2015 (UTC)
22.  Support. Jules78120 (talk) 19:54, 1 December 2015 (UTC)
23.  Support --Usien6 (talk) 20:05, 1 December 2015 (UTC) // Great value, given it's pretty easy to implement.
24.  Support  Trizek ^{from FR} 22:12, 1 December 2015 (UTC)
25.  Support - Kertraon (talk) 22:31, 1 December 2015 (UTC)
26.  Support Stevie is the man! ^{Talk • Work} 23:06, 1 December 2015 (UTC)
27.  Support Spencer (talk) 01:09, 2 December 2015 (UTC)
28.  Support Risker (talk) 04:21, 2 December 2015 (UTC)
29.  Support Graham87 (talk) 10:29, 2 December 2015 (UTC)
30.  Support Quick and useful fix.  DiscantX 12:28, 2 December 2015 (UTC)
31.  Support, quick, easy and useful fix — NickK (talk) 15:19, 2 December 2015 (UTC)
32.  Support Fluffernutter (talk) 15:52, 2 December 2015 (UTC)
33.  Support Mule hollandaise (talk) 18:41, 2 December 2015 (UTC)
34.  Support Wnt (talk) 23:43, 2 December 2015 (UTC)
35.  Support --AS (talk) 09:53, 3 December 2015 (UTC)
36.  Support — Arkanosis ^✉ 14:07, 3 December 2015 (UTC)
37.  Support as long as you could put "earliest" or "current" into the boxes. Krett12 (talk) 16:13, 3 December 2015 (UTC)
38.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
39.  Support GY Fan★ 11:29, 4 December 2015 (UTC)
40. Very strong  Support Bináris ^{tell me} 18:36, 4 December 2015 (UTC)
41.  Support Avgr8 (talk) 21:02, 4 December 2015 (UTC)
42.  Support --Emptywords (talk) 01:45, 5 December 2015 (UTC)
43.  Support — Rhododendrites ^talk \\ 00:41, 7 December 2015 (UTC)
44.  Support — seems easy enough to implement. Wbm1058 (talk) 18:58, 7 December 2015 (UTC)
45.  Support --Waldir (talk) 15:14, 8 December 2015 (UTC)
46.  Support Wugapodes (talk) 18:36, 11 December 2015 (UTC)
47.  Support Ekkt0r (talk) 06:38, 13 December 2015 (UTC)
48.  Strong support as an easy-to-implement project, which already is supported in the core codebase, and just needs a few GUI-bits added to let folks take advantage of it. Prefer a datepicker, but anything is an improvement over hand-modifying the querystring. If you want a more difficult task, implement an improved contrib-search feature, broadly construed. 75.108.94.227 21:12, 13 December 2015 (UTC)
49.  Support --MisterSanderson (talk) 04:37, 14 December 2015 (UTC)

Improve MediaWiki's blocking tools

Our blocking tools suck. It is a trivial matter to defeat blocks, and any vandal worth his salt can do it in his sleep. (User:Philippe). Anyone who works a sockpuppet investigation page or is a victim of repeated on-wiki harassment can attest to this. Block evasion doesn't have to be deliberate -- in some parts of the world, users can be assigned a different IP address by their ISP for every edit they make. We effectively have no measures against these users.

The aim here is threefold -- to make it more difficult to evade blocks, make it easier for us to identify and deal with potential sockpuppets as soon as possible-- ideally before they start editing -- and keep collateral damage at a minimum. This proposal has multiple facets:

• Look at other forum/blog/wiki software (e.g. Wordpress, vBulletin) and determine which blocking tools can be feasibly integrated into MediaWiki (not as an extension) subject to the constraints in our privacy policy. Implement them.
• Implement the existing tickets:

• Sets a blocked user cookie

• Added 27 November: investigate the use of Evercookie-like and other obnoxious tracking techniques to make the cookie harder to remove, while remaining within our privacy policy.

• Block by user agent only, don't block the underlying IP
• Throttle account creation by user agent
• When the account creation throttle is hit, flag all accounts that were created by that user (limit to Checkuser if necessary)
• Recursive checkuser
• Checkuser watchlist feature

• Block by device ID (needs check with WMF Legal first)
• Added 27 November: if an account with a registered email address is blocked with account creation blocked, prevent creation of any new accounts with that email address. (Optional, because we don't require email on registration.)
• Any other suggestions from the community that will help tackle this problem.

Improved blocking tools may be assigned to the Checkuser group initially to get a feel for how much collateral damage they cause. This will ideally reduce the burden of cleaning up after spammers, vandals and long term abusers, reduce the amount of on-wiki harassment of the type referred to in the diff above and will benefit all good faith editors of any MediaWiki installation.

MER-C (talk) 12:06, 8 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed I've always wondered why blocking didn't set a cookie. ONUnicorn (talk) 14:02, 8 November 2015 (UTC)  Endorsed Oh yeah...expedite and roll out to remaining admins after brief trial with Checkusers. ⋙–Berean–Hunter—► ((⊕)) 23:51, 9 November 2015 (UTC)  Endorsed Recommend to eliminate need for check user. If 2 sock puppet or more is block within same week from a same IP, the IP is block for 1 week without the need to checkuser. Genuine contributor can always register from the same IP address. Yosri (talk) 02:26, 10 November 2015 (UTC)  Endorsed --Yann (talk) 16:06, 10 November 2015 (UTC)  Endorsed Amir (talk) 10:06, 11 November 2015 (UTC)  Endorsed Finer-grained block tools are desperately needed on many projects. UA blocking is something I've been wishing especially hard for. Fluffernutter (talk) 17:21, 11 November 2015 (UTC)  Oppose. fewer tools; more leadership. it is unclear that more tools will decrease vandalism / socking and no study of unintended consequences. Slowking4 (talk) 00:11, 13 November 2015 (UTC)  Endorsed This won't stop the expert evaders but it will make it easier to stop the majority, freeing up people to do other things. Davidwr/talk 05:24, 16 November 2015 (UTC)  Endorsed Better tools to improve the anonymous grief folks can impose on positive volunteers is a must. Ckoerner (talk) 20:53, 16 November 2015 (UTC)  Endorsed Swpb (talk) 19:54, 17 November 2015 (UTC)  Endorsed This discussion [[1]] suggests this is a serious problem and a malicious individual with fairly basic hacking skills could seriously disrupt Wikipedia unless there are better tools for automatic detection and blocking of IP-hopping vandals. Msubotin (talk) 20:40, 18 November 2015 (UTC)  Endorsed As a former admin on fawiki who has had to deal with many trolls, I think improvement in MediaWiki's blocking system is vitally important to reduce the amount of work needed to deal with some vandals and with harassment in general. Dalba 08:28, 19 November 2015 (UTC)

Votes

1.  Support 4nn1l2 (talk) 03:14, 30 November 2015 (UTC)
2.  Support as proposer. MER-C (talk) 09:45, 30 November 2015 (UTC)
3.  Support Samwalton9 (talk) 10:31, 30 November 2015 (UTC)
4.  Support Jenks24 (talk) 10:38, 30 November 2015 (UTC)
5.  Support Lugnuts (talk) 12:07, 30 November 2015 (UTC)
6.  Support
   ⋙–Berean–Hunter—► ((⊕)) 15:28, 30 November 2015 (UTC)
7.  Support Yes please- right now, the only solution for IP-hopping vandals is literally to wait until they get bored. PresN (talk) 18:32, 30 November 2015 (UTC)
8.  Support. --Stryn (talk) 19:11, 30 November 2015 (UTC)
9.  Support Matiia (talk) 20:22, 30 November 2015 (UTC)
10.  Support Dalba 20:46, 30 November 2015 (UTC)
11.  Support --Grind24 (talk) 20:50, 30 November 2015 (UTC)
12.  Support Mike V • Talk 02:43, 1 December 2015 (UTC)
13.  Support Doc James (talk · contribs · email) 09:32, 1 December 2015 (UTC)
14.  Support --Steinsplitter (talk) 11:27, 1 December 2015 (UTC)
15.  Support Goombiis (talk) 16:44, 1 December 2015 (UTC)
16.  Support wow--Temp3600 (talk) 16:45, 1 December 2015 (UTC)
17.  Support --Snaevar (talk) 16:48, 1 December 2015 (UTC)
18.  Support One big way to help folks feel more welcome is to improve the tools to keep the jerks out. Ckoerner (talk) 17:09, 1 December 2015 (UTC)
19.  Support--Calak (talk) 18:33, 1 December 2015 (UTC)
20.  Support --Usien6 (talk) 20:07, 1 December 2015 (UTC) // Neutral on the other proposals.
21.  Support StevenJ81 (talk) 22:35, 1 December 2015 (UTC)
22.  Support generally - whatever can be proven to work and not cause new issues. Stevie is the man! ^{Talk • Work} 23:12, 1 December 2015 (UTC)
23.  Support In veritas (talk) 04:30, 2 December 2015 (UTC)
24.  Support ... I especially like checking the machine code and setting a cookie. Binksternet (talk) 06:14, 2 December 2015 (UTC)
25.  Support Graham87 (talk) 10:33, 2 December 2015 (UTC)
26.  Support--Syum90 (talk) 12:49, 2 December 2015 (UTC)
27.  Support any reasonable solution. It is very easy to delete a blocking cookie, however, but we should think of possible improvements — NickK (talk) 15:21, 2 December 2015 (UTC)
28.  Support Matěj Suchánek (talk) 15:52, 2 December 2015 (UTC)
29.  Support If I had to pick only one improvement to vote for, it would be this. While mediawiki's software has gotten slicker and more accessible, the tools that underlay much of the work people do with it have lagged behind. Fluffernutter (talk) 15:55, 2 December 2015 (UTC)
30.  Support -- Dave Braunschweig (talk) 22:12, 2 December 2015 (UTC)
31.  Oppose Blocking the worst problem users is a pipedream. Helping out the NSA and its ilk by doing mass surveillance of regular users and occasional visitors by storing troves of user-agent data or sending people out on the web bleeding fancy hacked Evercookie type data polluting their browser ... that's the reality. Just don't. Wikipedia would be better off if it never blocked anyone, and relied solely on positive certification of good users, but at least don't dig the hole any deeper. Wnt (talk) 23:35, 2 December 2015 (UTC)
32.  Oppose if you don't like playing whack a mole, stop. problem is not lack of tools, but lack of leadership. Slowking4 (talk) 02:46, 3 December 2015 (UTC)
33.  Support I think you should put all of these in, but beware even though evercookies are in theory harder to remove, in practice they are quite easily circumvented. Krett12 (talk) 16:16, 3 December 2015 (UTC)
34.  Support GY Fan★ 11:32, 4 December 2015 (UTC)
35.  Support It won't stop everyone, but every reduction in abuse is obvious win. Alsee (talk) 16:24, 5 December 2015 (UTC)
36.  Support Courcelles 08:34, 8 December 2015 (UTC)
37.  Support I wouldn't go so far as to say our blocking tools suck—IME there are far more of the users easily deterred by the extra effort involved to circumvent them than the proposer, or Phillippe in his linked edit, seem to realize—but I won't argue that they could be better. Daniel Case (talk) 19:30, 8 December 2015 (UTC)
38.  Oppose - I don't see how clearing a cookie is more difficult than changing an IP. There are bigger concerns with a persistent cookie strategy, as mentioned by User:Wnt at point 31 above. Bcharles (talk) 00:48, 9 December 2015 (UTC)

List of contributors

Problem: CC-BY-SA and GFDL require to add a list of contributors whenever content is copied from one project to another or outside a wiki (WP to WB, en-WB to de-WB, but also from this forum to de-WB or from de-WB to any external PDF). In all these cases it's necessary to add such a list. In the past, there was a tool by Duesentrieb at Toolserver which is deactivated before July, 2014. Next, there were Xtools which don't work since June, 2015 -- see GitHub issue. There's no way to get one list for all pages (incl. subpages) of one book in Wikibooks.

Users: All users who copy content, mainly admins.

At the moment, one has to export a page as pdf or book via Tools and then copy the created list inside the exported work.

Proposed solution: Such a list may be created by analyzing the history of a page. It should be possible to include this feature into the MediaWiki software and make available via ToolsPrint/Export. The feature should offer some options: users with account yes/no, IPs yes/no, including subpages yes/no, including pages by prefix yes/no. (The last request concerns something like b:de:Mathe für Nicht-Freaks where subpages are marked by colon instead of slash.)

-- Juetho (talk) 14:48, 20 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed Etamni (talk) 01:33, 22 November 2015 (UTC)  Endorsed -- Axum (talk) 09:48, 22 November 2015 (UTC) Comment (German) Die Liste wird auch benötigt, wenn zwei Seiten (z.B. WP-Artikel) zusammengefasst werden: Nachdem Inhalt per Cut&Paste übertragen wurde, kann die überflüssige Seite gelöscht werden. Dadurch verschwindet die Versionsgeschichte; stattdessen müssen die bisherigen Autoren genannt werden. -- Axum (talk) (try to translate in English) This list is necessary in cases where two sites (WP articles, e.g.) are merged: After content is transferred via Cut&Paste, the obsolete site maybe deleted. The history vanishes; one has to store the list of contributors instead of. -- Axum (talk)  Endorsed. See also phab:T29629 and phab:T30064. Helder 12:20, 22 November 2015 (UTC)

Votes

1.  Support.--Nahum (talk) 19:40, 1 December 2015 (UTC)
2.  Support  Trizek ^{from FR} 22:12, 1 December 2015 (UTC)
3.  Support Seems reasonable. Stevie is the man! ^{Talk • Work} 23:15, 1 December 2015 (UTC)
4.  Support Necessary per our licenses, and useful to spread the idea of F/LOSS. The opposite would also be great (User X has contributed 57% of article A, 21% of article B, and so on). --Pgallert (talk) 18:27, 2 December 2015 (UTC)
5.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
6.  Support Halibutt (talk) 22:50, 4 December 2015 (UTC)
7.  Support Juetho (talk) 11:58, 8 December 2015 (UTC)
8.  Support --Waldir (talk) 15:18, 8 December 2015 (UTC)
9.  Support Why have we waited so long even to suggest this? Daniel Case (talk) 19:30, 8 December 2015 (UTC)
10.  Support Wugapodes (talk) 18:45, 11 December 2015 (UTC)
11.  Support Yes, this does seem like something that should have been put in place a long time ago. Mz7 (talk) 04:39, 12 December 2015 (UTC)
12.  Support Nicely done will be non-trivial (comparing to some other proposals) and can save much time, lead to better compliance with license terms. aegis maelstrom δ 11:04, 14 December 2015 (UTC)

Machine learning to identify sockpuppets

Tracked in Phabricator:
Task T107600

Use machine learning and text mining to detect potential sockpuppet accounts. See Sockpuppet evidence from automated writing style analysis.

MER-C (talk) 22:31, 18 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed We can use every tool that we can get for hunting socks. ⋙–Berean–Hunter—► ((⊕)) 14:38, 20 November 2015 (UTC)

Votes

1. This is something WMF Legal would benefit from as well IMHO. -- とある白い猫 ^chi? 19:53, 30 November 2015 (UTC)
2.  Support Per personal experience on other sites; also, if one can make an antivandal bot (like enwiki's en:User:ClueBot NG) making a sockpuppet catcher bot should be possible to.Jo-Jo Eumerus (talk) 20:16, 30 November 2015 (UTC)
3.  Support --Isacdaavid (talk) 02:28, 1 December 2015 (UTC)
4.  Support Stevie is the man! ^{Talk • Work} 23:17, 1 December 2015 (UTC)
5.  Support Binksternet (talk) 06:15, 2 December 2015 (UTC)
6. Neutral for now due to the effort required for implementation -- there are easier ways to cut block evasion and sockpuppetry as detailed above. This will be, eventually, become a part of a defence in depth approach to combat sockpuppetry and is still worth doing... once all the obvious problems with core MediaWiki software are sorted. Might be a good research project. MER-C (talk) 15:33, 2 December 2015 (UTC)
7. Neutral better to use machine learning for good edit versus vandalism. Slowking4 (talk) 02:50, 3 December 2015 (UTC)
8.  Oppose Just because two editors write the same doesn't make them the same person. Krett12 (talk) 16:18, 3 December 2015 (UTC)
   • Funny ... that's the first line you read from any sockpuppet account when you raise the subject. Daniel Case (talk) 19:32, 8 December 2015 (UTC)
   • Statistically speaking AI in this area is used even by Law Enforcement to prosecute criminal cases, most noteworthy is the en:Enron Corpus which was a product of the investigation of the en:Enron Scandal where en:Text mining was used. In a nutshell AI was able to identify authors of fake email accounts to the real people accounts to identify who was engaged in the illegal activities. Naturally, processing emails and on wiki edits are different tasks but that is more of a matter of feature selection and fine tuning. Such a tool would probably be better than humans in identifying patterns and such (humans cannot review every revision, machines can for example). Such a tool would also identify lobbyist groups and other such coordinated efforts of POV pushing (meatpuppets) which are also a problem. -- とある白い猫 ^chi? 03:54, 29 December 2015 (UTC)
9.  Support Wouldn't be perfect, but would identify something for us to take a closer look at. Daniel Case (talk) 19:32, 8 December 2015 (UTC)
10.  Support Any help in identifying possible socks is a good thing. AlbinoFerret 18:29, 10 December 2015 (UTC)

OTRS permissions checker

Tracked in Phabricator:
Task T88620

For a normal user is impossible to know if an OTRS permission is true or false. For example this template suggests to contact an OTRS volunteers. This is a very fragile system. I propose to expose a special page (or anything you want) where you can add a ticket number and get as response a boolean value. The system should also be easily machine readable, in order to have an automatic report of false/wrong/vandalised ticket.

--AlessioMela (talk) 10:12, 10 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed it could be very useful when there is no OTRS volunteers available. Restu20 10:16, 10 November 2015 (UTC)  Endorsed even if OTRS volunteers are available it is infeasible to check a large number of tickets. --Incola (talk) 10:32, 10 November 2015 (UTC)  Endorsed --Yann (talk) 16:27, 10 November 2015 (UTC)  Endorsed provided this was setup to handle queries such as "does OTRS ticket 9999 relate to user:ZZZZ" more detail and you start to breach privacy, less detail and all you know is that this is a valid OTRS ticket though it may be unconnected to the person who is using it. WereSpielChequers (talk) 19:08, 10 November 2015 (UTC)  Endorsed Yes some automatic tool to even check if the ticket even exists or if it is valid( any images approved to use it), or what other images use this ticket number, or to check what is ticket number if only "ticketID" or ticket URL is known (like here File:002 Dafna Kaplan Taliya Finkel 5x5.jpg). There is a lot we could check without need to access the actual ticket details. --Jarekt (talk) 19:43, 10 November 2015 (UTC)  Endorsed --Shizhao (talk) 02:26, 11 November 2015 (UTC)  Endorsed - certainly useful to lessen OTRS workload, and give non-OTRS members a direct access to some basic non-private status information. Such a feature should give at least some basic status info like "ticket does not exist", "ticket request received", "ticket is still in process", "ticket reviewed and OK", "ticket reviewed and rejected". None of these anonymous general infos breaches privacy (imo). GermanJoe (talk) 16:40, 11 November 2015 (UTC) Using OTRS for permissions is backwards and a compromise on our overarching goal of growing the free knowledge ecosystem. OTRS must be discouraged more and more in favour of public solutions such as self-marking. See also commons:Commons:OTRS#When contacting OTRS is unnecessary. Either way, all the requests above are plainly impossible with OTRS and would require a new ticket system. Nemo 09:44, 13 November 2015 (UTC)  Endorsed Trust in the validity of OTRS tickets is essential. Blue Rasberry (talk) 16:31, 30 November 2015 (UTC)

Votes

1.  Support Definitely. Nyttend (talk) 21:55, 30 November 2015 (UTC)
2.  Support We need information to be shared as much as possible, when possible.--Alexmar983 (talk) 09:48, 1 December 2015 (UTC)

    Support Seems reasonable. Stevie is the man! ^{Talk • Work} 23:21, 1 December 2015 (UTC)

   Removed support per Earwig's reply below. This proposal is unnecessary. We should be able to trust OTRS members. Stevie is the man! ^{Talk • Work} 18:59, 4 December 2015 (UTC)

3.  Support but see "Earlier discussion and endorsements"--Jarekt (talk) 03:48, 2 December 2015 (UTC)
4.  Support When I sometimes delete a talk page on which an OTRS ticket is present (because of the deletion of the article itself), the only way I can notify people who would like to restore the article in the future is to explicitely specify the OTRS ticket number in the deletion log. This is not very satisfactory. Litlok (talk) 08:32, 2 December 2015 (UTC)
5.  Support, I already witnessed people adding OTRS permission tickets without getting the permission — NickK (talk) 15:22, 2 December 2015 (UTC)
6.  Support--Manlleus (talk) 15:29, 2 December 2015 (UTC)
7.  Oppose so as far as I can see, you want the information sent in the ticket to be made visible? Unfortunately, not going to happen any time soon, it's against the access to non-public data policy. If the ticket has been added by an OTRS volunteer, then it's as good as valid in any case. Mdann52 (talk) 18:01, 2 December 2015 (UTC)

   Is there a direct way to confirm that the editor adding the ticket is an OTRS volunteer? :) Stevie is the man! ^{Talk • Work} 18:38, 2 December 2015 (UTC)

   @Stevietheman: OTRS/Users is usually a good place to start (every OTRS agent has access to a list of former accounts as well, so if it isn't listed there, feel free to ask :) ) Mdann52 (talk) 14:36, 5 December 2015 (UTC)

   @Mdann52: no, I don't want to make any information visible. I just want a system that tell me a certain ticket is valid. Without showing anything else. --AlessioMela (talk) 19:56, 2 December 2015 (UTC)

   There is already an edit filter (on enwiki) (642) which verifies that the person adding an OTRS permission template is an OTRS member. — Earwig ^talk 02:12, 4 December 2015 (UTC)

   @The Earwig: Do you think this filter is easy to use to check after, for example, one year if a ticket is valid? How about if the OTRS member made a mistake copying the ticket number? Isn't better to have a stronger system directly connected with the OTRS database? --AlessioMela (talk) 21:41, 4 December 2015 (UTC)

   I'm not really sure, hence my lack of a vote on this one yet. It would be frustrating to go through the filter logs to see whether an edit from a year ago was tagged, but the idea is that it lets us deal with nefarious additions as they happen. As for mistakes copying the ticket number, I've never heard of this being a problem; is it?

   There's a more central issue with this. In order for a verifier to be useful, it needs to do more than just check if the ticket exists, because anyone can copy an OTRS template from another page and pretend it applies. There needs to be a check that the permission (a) applies to the resource in question, and (b) is a properly handled ticket. I don't think a non-human tool can do this, based on how OTRS works internally (I'm an OTRS member, but not of permission queues, so I am not 100% informed of procedures). We can require all OTRS members to log permissions in a secure area, but is this really necessary? It seems like that would take more work than just asking an OTRS member if you suspect something is wrong. — Earwig ^talk 22:13, 4 December 2015 (UTC)

   So as someone who has done work on permissions, it's a mess to be frank. There are so many different outcomes that there is little or no way to connect this directly with the database (the only useful flags on there are the ticket "closed as", which are all set to the same sort of thing anyway). This seems like a neat idea - however, there are no better ways of implementing this compared to what we have now in practice. Mdann52 (talk) 14:36, 5 December 2015 (UTC)

8.  Support Kvardek du (talk) 10:35, 5 December 2015 (UTC)
9.  Support --Waldir (talk) 15:21, 8 December 2015 (UTC)
10.  Support --Edgars2007 (talk) 09:07, 12 December 2015 (UTC)

Paragraph blaming tool

Research:Wikiwho Provenance Api.

Sometimes vandals insert a piece non-sensical information in the middle of a big article which stays covert for years. This is specially true in the Portuguese Wikipedia, which doesn't have enough task force for a immediate vandalism response. It would be great if there was a tool to "blame" a paragraph, like there is in GitHub. I mean, a tool that, given a paragraph (or a set of paragraphs), points out the last edition (or, perhaps, all the editions) in which this paragraph showed up in the differential. Sometimes it wouldn't work due to merges and displacements, but probably would for the 99%.

--Usien6 (talk) 03:06, 11 November 2015 (UTC)

Earlier discussion and endorsements
We have the same problem in English. Currently some edits are looked at by dozens of people and others are looked at by none. The solution is to introduce flagged revisions such as they have on DE wiki, that way you know which are the unchecked edits by newbies, volunteer time is wasted less and every newbie edit is looked at at least once. WereSpielChequers (talk) 06:58, 11 November 2015 (UTC) Dear @WereSpielChequers: Actually the Portuguese Wikipedia does implement revision flagging, but it doesn't seem to be 100% effective... --Usien6 (talk) 13:22, 13 November 2015 (UTC) No system is going to be 100%, be glad you have a better system than we have on English. WereSpielChequers (talk) 08:11, 17 November 2015 (UTC) At dewiki we have de:Benutzer:Schnark/js/artikel-statistik.js (to be activated on user-level in common.js), which pretty much has this functionality. There are also other tools available, try for instance X!’s Article-Blamer (on Tool Labs) and WikiBlame. —MisterSynergy (talk) 07:15, 11 November 2015 (UTC) Dear @MisterSynergy: I wasn't aware of these two tools, thank you! But, still, it would be nice if it was a native MediaWiki feature. The wikiblame.php tool doesn't seem to work. The artikel-statistik.js works, but have a poor user interface, though, and runs pretty slowly in article with high revision count. --Usien6 (talk) 13:22, 13 November 2015 (UTC) If there is a high revision count to process, then it’s always very difficult to blame particular words or sentences to their authors. I would be surprised if a native Mediawiki extension would be significantly faster than the JS solution artikel-statistik.js. Anyway, I now endorsed your proposal :-) —MisterSynergy (talk) 14:08, 13 November 2015 (UTC)  Endorsed Related: phab:T26617 ("Implement diff on sentence level (instead of per paragraph block)"). cscott (talk) 19:33, 11 November 2015 (UTC) Notice where I said "line" I actually meant "paragraph" --Usien6 (talk) 13:22, 13 November 2015 (UTC) Usien6: Do you mean that you'd like to change "Line blaming tool" to "Paragraph blaming tool"? You can feel free to change it, if that's more accurate. -- DannyH (WMF) (talk) 15:30, 16 November 2015 (UTC) @DannyH (WMF): Done ! --Usien6 (talk) 16:02, 17 November 2015 (UTC)  Endorsed —MisterSynergy (talk) 14:08, 13 November 2015 (UTC)  Endorsed --Pine✉ 05:17, 16 November 2015 (UTC)

Votes

1.  Support; this tool should also be useable for old revisions (i.e a user adds gibberish in one place in a paragraph; an other user then edits an other part of the paragraph. Find the first user form the latest revision before the second user's edit). עוד מישהו Od Mishehu 16:19, 30 November 2015 (UTC)
2.  Support Orlodrim (talk) 20:30, 30 November 2015 (UTC)
3.  Support --Isacdaavid (talk) 02:31, 1 December 2015 (UTC)
4. Neutral http://wikipedia.ramselehof.de/wikiblame.php --Purodha Blissenbach (talk) 14:53, 1 December 2015 (UTC)
5.  Support --Usien6 (talk) 20:11, 1 December 2015 (UTC)
6.  Support I probably wouldn't have to use this much, but when I need it, it will be very handy. Stevie is the man! ^{Talk • Work} 23:24, 1 December 2015 (UTC)
7.  Support Rdrozd (talk) 00:33, 2 December 2015 (UTC)
8.  Support – Ham II (sgwrs / talk) 20:58, 2 December 2015 (UTC)
9.  Support - I know there are already things along this line, but until they are really obvious and right in the face of every user who potentially can look up how something odd got in an article, they won't really get used. And it's far too common for people to fix one sentence by a vandal and leave another somewhere else, or to merely soften it in a way that camouflages that it is a bogus claim because they don't realize it was just abuse. Wnt (talk) 23:38, 2 December 2015 (UTC)
10.  Support YBG (talk) 06:29, 3 December 2015 (UTC)
11.  Support --V111P (talk) 10:21, 3 December 2015 (UTC)
12.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
13.  Support I can't count the times I would have liked to have had this. Daniel Case (talk) 19:33, 8 December 2015 (UTC)
14.  Support - Good idea if it can be implemented the way Wnt outlined above. - Pointillist (talk) 14:03, 9 December 2015 (UTC)
15.  Support - I'm not sure it's exactly what's proposed here, but the ability to enter some text and find when that text appeared in the page would be much more efficient than the current method of looking at each diff or version of page. DexDor (talk) 20:44, 9 December 2015 (UTC)
16.  Support Could be helpful in tracking odd claims; a simplified version of a more complicated (but probably needed) tool showing who wrote what, eg with background colours. aegis maelstrom δ 11:08, 14 December 2015 (UTC)
17. Comment This is something ORES provides. It only provides a score for the entire revision though. -- とある白い猫 ^chi? 04:06, 29 December 2015 (UTC)

RFC tools

Tracked in Phabricator:
Task T31923

Perhaps there is a better way to capture ideas from the community for things like this document and RFCs. They tend to be very "Meta-Wiki" community focused outcomes, and become a bit complicated at times for less tech-savvy people (which sometimes may defeat the purpose). Perhaps there are tools that can increase language and cross-wiki participation. There might also be lessons from the IdeaLab project that could be applied.

Earlier discussion and endorsements
Endorsed This is really an area, where I agree we could truly use 'community' tech. Delection discussion/Voting/RFC'ing is a core element of our community workflows, that is sorely unsupported by tools. —TheDJ (talk • contribs) 18:17, 19 May 2015 (UTC)

Votes

1. Comment Does anyone have any examples of RFCs where such tools could have had a positive impact? I'm trying to better visualize this proposal. Stevie is the man! ^{Talk • Work} 23:30, 1 December 2015 (UTC)

Shared variables for AbuseFilter

for example, to store regex with bad words, and such variable could be used in many filters

--AS (talk) 22:20, 15 June 2015 (UTC)

Earlier discussion and endorsements
phab:T57472?--Qgil-WMF (talk) 10:53, 16 June 2015 (UTC) I'm not sure what global filters are, but I think I mean different thing: be able to declare a variable, which would be available in each filter. --AS (talk) 10:13, 18 June 2015 (UTC) See also: phab:T20067: Decision tree for AbuseFilters phab:T68032: AbuseFilter should allow storing regex matches into variables phab:T49512: Switch AbuseFilter to using Lua Helder 13:52, 9 July 2015 (UTC)  Endorsed MER-C (talk) 10:18, 16 November 2015 (UTC)

Votes

1.  Support John Vandenberg (talk) 02:01, 1 December 2015 (UTC)
2.  Support --Alexmar983 (talk) 09:52, 1 December 2015 (UTC)
3.  Support --Usien6 (talk) 20:16, 1 December 2015 (UTC) // The code of most filters miserably lacks separation between data and logics, killing maintainability.
4.  Support Helder 23:41, 1 December 2015 (UTC)
5.  Support --AS (talk) 13:19, 2 December 2015 (UTC)
6.  Support --Edgars2007 (talk) 09:08, 12 December 2015 (UTC)
7.  Oppose I would strongly recommend against this. You will find that such a generic list brings collateral damage. I would instead suggest using machine learning tools that will weight bad words and other content based on the existing content/wiki. ORES can provide this. -- とある白い猫 ^chi? 04:03, 29 December 2015 (UTC)

Special:NewPagesFeed in every language

Tracked in Phabricator:
Task T50552

Hello. I wish the following page Special:NewPagesFeed in the english Wikipedia to be available in all wikipedias in every language (and especially in my language, the french language). It is a very useful page for every contributor and for the reviewing of new articles. Thank you !

--Consulnico (talk) 14:25, 10 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed בנימין (talk) 15:04, 10 November 2015 (UTC) To enable page curation, this should be fixed. --Edgars2007 (talk) 16:20, 10 November 2015 (UTC)  Endorsed --Superjuju10 [Aubline available to you], 22:24, 10 November 2015 (UTC)  Endorsed Thibaut120094 (talk) 23:29, 10 November 2015 (UTC)  Endorsed--Shizhao (talk) 02:29, 11 November 2015 (UTC)  Endorsed Amir (talk) 10:12, 11 November 2015 (UTC)

Votes

1.  Support This is a very important workflow tool built but not yet finished as it only works on English Wikipedia. I would very much like to see the Community Tech team work with one or two wikis to set it up adequately for those communities needs, which will iron out some bugs, and help other wikis follow in their footsteps. John Vandenberg (talk) 02:13, 1 December 2015 (UTC)
2.  Support--Shizhao (talk) 09:41, 1 December 2015 (UTC)
3.  Support Sadads (talk) 15:54, 1 December 2015 (UTC)
4.  Support --Temp3600 (talk) 16:44, 1 December 2015 (UTC)
5.  Support Papuass (talk) 17:20, 1 December 2015 (UTC)
6.  Support Stevie is the man! ^{Talk • Work} 23:35, 1 December 2015 (UTC)
7.  Support Spencer (talk) 01:10, 2 December 2015 (UTC)
8.  Support per John Vandenberg and other commenters. That several of the supporters of this edit primarily in languages other than English should add some bonus points to this. Risker (talk) 04:25, 2 December 2015 (UTC)
9.  Support--Barcelona (talk) 12:02, 2 December 2015 (UTC)
10.  Support Very useful over at en.wp, I'm sure the other wikis could benefit as well.  DiscantX 12:27, 2 December 2015 (UTC)
11.  Support--Manlleus (talk) 15:29, 2 December 2015 (UTC)
12.  Support Mule hollandaise (talk) 19:00, 2 December 2015 (UTC)
13.  Support – Ham II (sgwrs / talk) 21:02, 2 December 2015 (UTC)
14.  Support SantiLak (talk) 10:44, 4 December 2015 (UTC)
15.  Support Avgr8 (talk) 21:04, 4 December 2015 (UTC)
16.  Support, we could really use something like this on pl.wiki (and we can help ironing out the bugs) Halibutt (talk) 22:53, 4 December 2015 (UTC)
17.  Support --Yeza (talk) 16:49, 5 December 2015 (UTC)
18.  Support - Bcharles (talk) 01:08, 9 December 2015 (UTC)
19.  Support Matěj Suchánek (talk) 21:05, 9 December 2015 (UTC)
20.  Support --Edgars2007 (talk) 09:08, 12 December 2015 (UTC)
21.  Support - Superb tool at en. SBaker43 (talk) 03:55, 14 December 2015 (UTC)

Suggesting AbuseFilter by machine learning

Recently we should manually write down Extension:AbuseFilter's pattern (string match, regex, etc). It is a very hard for non-technical user (just a admin of a wiki) and consumes technical user's time.

I propose a machine suggestion for AbuseFiliter's pattern to reduce such difficulties and cost. For example, when I put marks on some revisions or users, I can get the suggested pattern generated by machine learning which extracts points in common among the specified revisions or user's contributions.

I don't have any concrete methods or implementations because I'm specialized in neither the machine learning or natural language processing. But I heard it's not impossible.

--aokomoriuta (talk) 23:56, 9 November 2015 (UTC)

Earlier discussion and endorsements
Endorsed See also Research:Revision scoring as a service and Objective Revision Evaluation Service. Helder 11:30, 10 November 2015 (UTC)  Endorsed -- とある白い猫 chi? 11:42, 14 November 2015 (UTC) I hesitate to endorse this. It's interesting to do and I support it, but I'm not sure its feasible. MER-C (talk) 22:06, 18 November 2015 (UTC)

Votes

1. Neutral --Usien6 (talk) 20:22, 1 December 2015 (UTC) // The idea is indeed genial but, until someone comes up with the working thing, it's just sci-fi...

   It's somehwat related to Research:Revision scoring as a service. /Johan (WMF) (talk) 09:34, 8 December 2015 (UTC)

   I would argue that Revision scoring may make AbuseFilters obsolete as we would no longer need to rely on simple word matches & regexes to identify problematic behavior. I would argue that current AbuseFilter implementation acts like a sledge hammer. While we wield it with great care, collateral damage is inevitable - more so on some smaller wikis where good content is often blocked due to the poor management of local AbuseFilters as the person setting it up simply may not be editing anymore. Main problem with AbuseFilters is that it is non-trivial to setup and maintain even for seasoned regex users as a simple syntax error can block a wide margin of edits which may remain unnoticed for quite some time. -- とある白い猫 ^chi? 04:00, 29 December 2015 (UTC)

Work queues

Using the extensive collection of en:WP:TC (cleanup templates) we have in articles, I would like to see each Wikiproject have a dedicated page to show what tasks (and how many) can be done to articles within a WikiProject scope. (I would also like to see such a page for the entirety of each language Wikipedia.) I describe a version of this idea at Grants:IEG/Automated inventory pages for all WikiProjects. My inspiration for this idea is from http://www.wikihow.com/Special:CommunityDashboard Thank you.

Biosthmors (talk) 19:56, 19 May 2015 (UTC)

Earlier discussion and endorsements
We need work queues. A system, that shows articles that are in a certain group, for a certain reason and that need to be '(pre)processed' in a certain way. The person who can crack the task of solving this in a generic enough and expandable enough way, usable for multiple wiki types, in a way that editors can build new queues themselves, will be my ULTIMATE hero. Inspiration: WikiGrok + NewPagePatrol + WikiHow dashboards, Wikipedia:Huggle and Wikipedia:STiki, etc. I consider this the key to our long term survival. —TheDJ (talk • contribs) 18:29, 19 May 2015 (UTC)  Endorsed And I saw this comment after I posted mine below about "Inventory pages". Agreed. Biosthmors (talk) 19:57, 19 May 2015 (UTC)  Endorsed +1 --AS (talk) 16:45, 27 May 2015 (UTC) The Parsoid team would like something like this as well for wiki linting (finding and correcting wikitext errors or deprecated usage). Cscott (talk) 18:23, 11 November 2015 (UTC)  Endorsed--Shizhao (talk) 08:04, 12 November 2015 (UTC)  Endorsed Helder 12:02, 12 November 2015 (UTC)

Votes

1.  Support Though this sounds like the work already in place by Bambots for English: https://tools.wmflabs.org/bambots/cwb/ . Its probably scalable to other language editions, Sadads (talk) 15:57, 1 December 2015 (UTC)
2.  Support per Sadads. These cleanup listings are wonderful. Stevie is the man! ^{Talk • Work} 23:40, 1 December 2015 (UTC)
3.  Oppose - The extensive disfiguration of articles in the English Wikipedia is something that should be strongly prevented. Please do not try to enforce this frustrating way of working, for both users and readers, upon other wiki's. RonnieV (talk) 13:20, 2 December 2015 (UTC)
4.  Support per Sadads - SantiLak (talk) 10:44, 4 December 2015 (UTC)
5.  Support - Would need to focus on a dozen or so tasks at any given time that could use new recruits to help out. Could have customizable set of "widgets" that user has interacted with. Bcharles (talk) 01:49, 9 December 2015 (UTC)