Wikipedia:Counter-Vandalism Unit/RfC (Proposal)

From Wikipedia, the free encyclopedia
Jump to: navigation, search
CVU.svg
There is consensus among the participants in this discussion that the bot be approved. However, some valid concerns about function and usefulness are raised; moreover, there is so little discussion here that I would be hesitant to draw any firm conclusion. Moreover, it seems to me that this RfC was a way to gain momentum for Wikipedia:Bots/Requests for approval/Cyberbot II, and if that was the goal it has failed. In other words, though the RfC appears to endorse the proposal, a decision should be made elsewhere (for the sake of redundancy, at Wikipedia:Bots/Requests for approval), and it hasn't yet. Calling the bot request dead in the water may be an overstatement, but considering that only 13 editors have responded to the proposal one might conclude that this does not bode well. Drmies (talk) 17:43, 26 July 2012 (UTC)
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

This discussion is intended primarily those who are involved or are members of the CVU WikiProject. It is also directed to dedicated vandal fighters.

The purpose of this RfC is to determine whether a notifications bot should be developed to notify subscribed users of elevated vandalism. Because vandal fighters have to proactively check vandalism levels, to decide if they want to help fight it or not, some of them may forget to check it. Receiving a notification of elevated levels by triggering the New Messages bar would be an ideal way of notifying the subscribed users immediately.

The bot currently supports the following features:

  • The bot only notifies the subscribed users in this list.
  • The bot will notify when the DefCon reaches a certain level.
  • The bot will only notify you after a certain amount of time has elapsed and the defcon is still at a certain level.
  • You can have the bot notify you as vandalism subsides. (Overrides the time setting)
  • You choose exactly where you want to be notified.

The bot will in the future have the following tasks implemented:

This RfC is broken into two sections. The first section is the community's input on the task itself. You may comment and discuss the task there. Please indicate your stand by supporting, opposing or remaining impartial. The second section if for functions that should be implemented in the bot's task if passed. Editors are encouraged to support, oppose, or comment each individual tasks.

Community input[edit]

Support[edit]
  1. Support As long as there is a system where it is only made through dummy edits. Ryan Vesey Review me! 02:33, 5 July 2012 (UTC)
  2. Support - With the existence of ClueBot, there are many experienced and willing counter vandals who have moved on to other projects, be they content creation, XfD, or other maintenance tasks. By having a system to alert users that Vandalism is particularly high and that their hands would be useful, we can curb a serious influx. The biggest problem facing in regards to Vandalism today is that, during times of peak activity, it is unlikely that all the vandalism will be caught by the bot and those RCPatrollers who are around. So, unless the vandalism occurs to a frequently watchlisted page, it just sits there. This bot may do well to solve that issue. Achowat (talk) 13:01, 5 July 2012 (UTC)
  3. Support I agree, ClueBot does a fantastic job, and beats me to most of the reverts, but I would still like this to notify me of the then-current status of vandalism. Perhaps the same way that the {{User:X!/RfX Report}} keeps real-time stats as to an RfA nomination, this bot could produce a similar status update. Vertium (talk to me) 23:18, 5 July 2012 (UTC)
  4. If it's only making small dummy edits rather than full-fledged sections, and is fully opt-in, I don't see the harm. Though, as Monty states below, it's unclear how effective it actually would be, since vandalism would probably no longer be a big enough problem by the time most of the editors notified actually got around to editing. elektrikSHOOS (talk) 04:38, 6 July 2012 (UTC)
  5. sounds reasonable. If AIV gets backlogged, it may take a human note at WP:AN to draw in help, but this could be useful. Of course we may want a proportionate response so a slightly elevated level alerts a few, and a really bad outbreak notifies the maximum number of helpers. Cluebot gets a lot but if reverted it does not battle the vandal, so an alert on this sort of behaviour would be good. I also notice spambots work in spurts, and may also involve crosswiki damage, so notifying those on other major wikipedias could be in order. Graeme Bartlett (talk) 12:22, 6 July 2012 (UTC)
  6. Support It's opt-in and reasonable. Electriccatfish2 (talk) 22:09, 6 July 2012 (UTC)
  7. Support A useful addition to monitoring. ʘ alaney2k ʘ (talk) 17:45, 26 July 2012 (UTC)
Oppose[edit]
  1. Per everything I said in the BRfA. Rcsprinter (converse) 16:14, 5 July 2012 (UTC)
    I fail to see an argument that you put forth in the BRfA that is applicable here. You stated that consensus hadn't been reached, that is what this is for. You also stated that you prefer to check manually; however, there are editors who would prefer not to. In addition, as you stated in the BRfA, the manual form doesn't get updated very often. Assuming that the parameters are valid, I believe the concern that it would spam people every minute is no longer relevant. I can limit the bot to notify me no more than once every 2 hours 12 hours etc. As an aside to cyberpower, does it only recognize whole numbers or could I set it to .25 if I wanted to be notified every 15 minutes (I wouldn't do that, but I'm just curious). Ryan Vesey Review me! 16:30, 5 July 2012 (UTC)
    Mind providing a link to the last state of that BRfA so we can understand your reasoning? Hasteur (talk) 18:19, 6 July 2012 (UTC)
    Rcsprinter123 is referring to this. This user is skeptical along with several other people about my abilities to handle a bot given that my first BRFA was a total disaster and also whether this bot is necessary or not.—cyberpower ChatOnline 18:28, 6 July 2012 (UTC)
  2. Oppose I disagree, ClueBot does a substandard job, and reverts grammatically correct changes. I have watched ClueBot revert expert changes, both in content and grammar. It uses a substandard protocol. I am against this tool. Tempdoc (talk) 23:32, 6 July 2012 (UTC)
    Do you you even know what this RfC is for?—cyberpower ChatOnline 23:37, 6 July 2012 (UTC)
Neutral[edit]
  1. As long the notifications are fully opt in, I don't see any real harm. That said, I think the vandalism defcon is quite useless. Vandalism usually comes in spurts, such that editors noticing an increased defcon almost always arrive to help out after the vandalism spike has subsided. Its not like there are queues or backlogs that build up (that non-admins can do much with). Monty845 20:32, 5 July 2012 (UTC)
  2. Not convinced that DEFCON serves any substantive purpose in counter-vandalism efforts. But as Monty said, I suppose there's no harm in trying. -Cntras (talk) 09:53, 9 July 2012 (UTC)
Comments and discussion[edit]
  • How will users be notified? I wouldn't want a bunch of messages left on my talk page, but would like it if it made dummy edits and left it in the edit summary. Would the bot be smart enough to leave a dummy edit? Maybe users who want the service could put a hidden notice on their talk page with something like "Cyberbot please make dummy edit here". Then would it be possible for the bot to use regex to ignore current spacing and always add one extra space after the word cyberbot? It would need to be a relatively long string of words like the one I pointed to so there isn't a risk of the words showing up in the course of normal conversation. Ryan Vesey Review me! 00:42, 5 July 2012 (UTC)
    • Something like this bot's small edit and edit summary, perhaps? BencherliteTalk 00:45, 5 July 2012 (UTC)
      Yes that can be implemented. Please add this request below by clicking the link provided.—cyberpower ChatOnline 00:47, 5 July 2012 (UTC)
      • Here's the problem that I see with the way that works. I'm fairly positive that adding a space at the end of a page won't cause a dummy edit. Book talk:Anglesey churches has a hidden comment at the end that gives the bot an area to make it's edit. A hidden comment at the end of someone's talk page would be archived by MiszaBot. In addition, new sections would follow the comment I believe. Ryan Vesey Review me! 00:52, 5 July 2012 (UTC)
        Empty spaces may not be archived by the bots is another problem I see. Perhaps <!--Dummy Edit--> could be appended to the end of a section on the talk page. It won't be left behind like empty spaces would therefore no clutter will build. It also doesn't show up on the talk page.—cyberpower ChatOnline 02:27, 5 July 2012 (UTC)
That would work well enough for me. Ryan Vesey Review me! 02:32, 5 July 2012 (UTC)

Does the bot automatically make modifications based on Wikipedia:Counter-Vandalism Unit/Notifications list or do you manually change it? My worry is that if a user decided they didn't want any restriction on the timespan between edits, and found they were being spammed by too many notices, it would need to be modified quickly. Is it possible for the bot to check that list prior to each run? Ryan Vesey Review me! 02:36, 5 July 2012 (UTC)

The bot always checks the list prior to each run. Any change made is instantaneous. If there isn't any restriction, users are only notified when the DefCon changes and when it's equal to or above the set threshold.—cyberpower ChatOnline 02:39, 5 July 2012 (UTC)

I do acknowledge that there are some issues with the defcon process itself. Namely, the more editors there are patrolling, the more reverts there could be. If there are only a few editors patrolling and a lot of vandalism, it won't appear like there are many reverts. Will this factor in reverts by cluebot? Ryan Vesey Review me! 12:26, 6 July 2012 (UTC)

  • Im mostly neutral in regards to this proposal. The idea has its merits, but the bot itself can actively change the very metric it relies on for reports. If there are many patrols about the amount of reverts will naturally be higher, and if there is a shortage of patrols the amount of reverts may be lower.I would also point out that reverts may not be current - if there are many people using Stiki the " Amount of reverts" metric may be off since STiki serves old edits instead of current edits.
I would equally suggest a setting that allows a user to input his regular editing hours and / or a functionality that checks if a user is currently editing before notifying that user. Seeing the "Don't warn more often then once per X hours" the bot might actually get "stuck" in a timezone during which an editor never edits (Warns user at midnight their regarding vandalism levels -> 12 hours later the user wakes up, bot won't warn for another 12 hours due to setting -> Loop(infinite) ) Excirial (Contact me,Contribs) 19:33, 6 July 2012 (UTC)

Requested implementations[edit]

Click here to request a feature be added to the task

Proposal by Ryan Vesey

Proposal: I would like an option to have the bot use dummy edits and only leave the message in a talk page note. The exact method of doing this would need to be discussed.

Support[edit]
  1. Support as proposer. Ryan Vesey Review me! 02:32, 5 July 2012 (UTC)
Oppose[edit]
Comments[edit]
  • This can be an additional setting that can be implemented into the bot.—cyberpower ChatOnline 02:20, 5 July 2012 (UTC)

Proposal by Excirial

Proposal: The bot should check if an editor is editing before notifying them, or allow a user to define regular editing hours. Otherwise the bot might warn at times a user never edits, with the "Warning cooldown" preventing another help needed notice when the user comes back online.

Support[edit]
  1. Support if this is doable. Theopolisme TALK 15:10, 11 July 2012 (UTC)
Oppose[edit]
Comments[edit]

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