Bots noticeboard |
---|
This is a message board for coordinating and discussing bot-related issues on Wikipedia (also including other programs interacting with the MediaWiki software). Although this page is frequented mainly by bot owners, any user is welcome to leave a message or join the discussion here. If you want to report an issue or bug with a specific bot, follow the steps outlined in WP:BOTISSUE first. This not the place for requests for bot approvals or requesting that tasks be done by a bot. General questions about the MediaWiki software (such as the use of templates, etc.) should be asked at Wikipedia:Village pump (technical). |
Archives |
---|
|
Contents
Bot unblocks
Hi there, we currently have two blocked bots with pending unblock requests. To a lay admin, it's unclear whether these bots can be unblocked outright at this point, or if we should get the nod from BAG first, or at least from a bot-experienced admin. Can someone who is more qualified handle these?
- DiBabelYurikBot (BRFA · contribs · actions log · block log · flag log · user rights) - blocked for light testing on non-test pages (an honest mistake by trusted user), but still pending BAG approval
- Citation bot (BRFA · contribs · actions log · block log · flag log · user rights) - blocked for issues which have apparently been resolved, but still undergoing maintenance.
Thanks in advance, ~Swarm~ {sting} 19:21, 1 May 2019 (UTC)
- Often for blocks of bots any admin can feel free to unblock when they believe the original problem won't reoccur when the block is lifted, see Wikipedia:Blocking policy#Blocks in temporary circumstances (second bullet). This can be as simple as the operator stating that they've disabled the offending code. On the other hand, if it was blocked as an unapproved bot there's usually no reason to unblock until BAG approves a trial as there's usually nothing the bot account is allowed to do that is prevented by the block.
BTW, if the operator is an admin, that may even implicitly allow the operator to do it themselves. Exactly how clear it has to be before that applies is arguable, safest is to only do so if the blocking admin explicitly said something like "feel free to unblock when that's fixed". Anomie⚔ 21:06, 1 May 2019 (UTC)
- As far as Citation bot goes, it appears the block was for making inappropriate edits (that would have been inappropriate if made by a human editor as well) - assuming this is the situation, it needs to be resolved first. It appears there is still active discussion occurring about that point? — xaosflux Talk 22:08, 1 May 2019 (UTC)
- You are incorrect, the complaint was that a very few links to copywrite infringling citeceerx references were added. That is long since fixed. The second complaint is that the bot does edits on its own without a human being getting credit. AManWithNoPlan (talk) 22:13, 1 May 2019 (UTC)
- It is a minor complaint that doesn't cause any harm to the encyclopedia, but it is an annoyance. IMO that shouldn't be enough to maintain the block, but I'm pretty involved in the discussion, so I'm not keen to give the thumbs up myself. Someone else could though. Headbomb {t · c · p · b} 23:53, 1 May 2019 (UTC)
- @AManWithNoPlan: can you elaborate? That bot is operated by Smith609, who is responsible for any edit made under that account. Is there a complaint that it is making edits outside of its approval, or that the approval needs to be revisted? — xaosflux Talk 00:20, 2 May 2019 (UTC)
- People are able to request that the bot edit a specific page, but the bot does not force users to reveal their identities. AManWithNoPlan (talk) 00:27, 2 May 2019 (UTC)
- OK? Is there a specific BRFA task # where you expect this to be occuring? Tt looks like it is using an ancient (in wiki time) passed-on BRFA (Wikipedia:Bots/Requests for approval/DOI bot) - which doesn't seem to require that. As long as Smith609 is taking responsibility for the edits, this doesn't seem to be a specific violation of anything, is there more to this? — xaosflux Talk 00:41, 2 May 2019 (UTC)
- People are able to request that the bot edit a specific page, but the bot does not force users to reveal their identities. AManWithNoPlan (talk) 00:27, 2 May 2019 (UTC)
- @AManWithNoPlan: can you elaborate? That bot is operated by Smith609, who is responsible for any edit made under that account. Is there a complaint that it is making edits outside of its approval, or that the approval needs to be revisted? — xaosflux Talk 00:20, 2 May 2019 (UTC)
- It is a minor complaint that doesn't cause any harm to the encyclopedia, but it is an annoyance. IMO that shouldn't be enough to maintain the block, but I'm pretty involved in the discussion, so I'm not keen to give the thumbs up myself. Someone else could though. Headbomb {t · c · p · b} 23:53, 1 May 2019 (UTC)
- You are incorrect, the complaint was that a very few links to copywrite infringling citeceerx references were added. That is long since fixed. The second complaint is that the bot does edits on its own without a human being getting credit. AManWithNoPlan (talk) 22:13, 1 May 2019 (UTC)
The bot has been approved many different times as features were added. Always approved to run as a pure bot. AManWithNoPlan (talk) 02:42, 2 May 2019 (UTC)
- @AManWithNoPlan: OK - so can you point to what task you think is malfunctioning? — xaosflux Talk 03:26, 2 May 2019 (UTC)
- I take it that part of the contention is that some people want a feature request added - @Headbomb: I think you are one of the requesters of this? I'm not seeing how this is a showstopper. If Smith609 is allowing his bot to make "bad edits" then, sure it should be stopped (and putting a name of who they made the bad edit on behalf of isn't really fixing that core problem is it?). Smith609 could work on fixing that lots of ways, including by only allowing certain users to trigger the bot. Are there some examples of bad edits that the bot is making for reference here? — xaosflux Talk 03:49, 2 May 2019 (UTC)
- @Xaosflux: I am indeed one of those that requested that. The reasons mostly being that when it was adding CiteSeerX links, this would have been useful to WP:TROUT users that didn't review the bot's edits. This would still be useful for trouting people that don't review the bot's edits, or to help users that try to use the bot, but there is no outstanding/egregiously problematic behaviour in terms of actual edits (at least as far as I can tell). Headbomb {t · c · p · b} 03:56, 2 May 2019 (UTC)
- I take it that part of the contention is that some people want a feature request added - @Headbomb: I think you are one of the requesters of this? I'm not seeing how this is a showstopper. If Smith609 is allowing his bot to make "bad edits" then, sure it should be stopped (and putting a name of who they made the bad edit on behalf of isn't really fixing that core problem is it?). Smith609 could work on fixing that lots of ways, including by only allowing certain users to trigger the bot. Are there some examples of bad edits that the bot is making for reference here? — xaosflux Talk 03:49, 2 May 2019 (UTC)
- I have several concerns about this bot relative to BOTPOL. First that its operator of record appears to be an absentee landlord: before the bot was blocked they had not edited any page related to the bot for several months (and ditto before that). There was a recent, contentious, RfC about the bot's behaviour in which they were pinged multiple times, but did not participate at all (despite editing elsewhere on the project). I have several times asked them (on their talk page and on the bot's talk page) to confirm whether they are in fact still the bot's operator, and to address WP:BOTACC, second para, and WP:BOTCOMM. There has been no response, beyond removing the question with the edit summary
Archive aggressive comments
. During the RFC the bot's proponents argued that since the bot was user activated it was not the bot that was responsible for the problematic edits (but it didn't identify the user activating the bot). Once the bot was blocked for, among other things, not identifying the activating user, the bot's operator suddenly chimed in claiming it was, in fact, the operator that was responsible for the edits (see also this thread at WT:BOTPOL). The net result seems to be that nobody takes responsibility for the edits.
- Second that this bot relies on a 11 year old BRFA for a task to "Adds DOIs to citations provided using
{{cite journal}}
". But the bot has changed extensively in the decade since that BRFA and its maintainers now appear to operate on the assumption that it has de facto approval to do "anything at all that is related to citations" with no need for a new BRFA (in the above RFC they were asked several times for the BRFA that authorized mass-removal of valid citation template parameters, with the only response a suggestion that it was "grandfathered in" due to the bot's age, and besides, the bot's maintainers didn't think anyone ought to use those parameters in any case, nevermind CITEVAR). There are several concerns surrounding responsiveness to the community and only making uncontroversial edits. I think the bot's various tasks need to be analysed (documented), and which ones have actual authorisation sorted. Existing BRFAs: 1, 2, 3, 4, 5, 6, 7, 8, 9 (last one in 2011).
- As for the two direct reasons the bot was blocked… As I understand it the addition of the problematic links has been stopped and this reason for the block resolved. In fairness it should also be noted that just how problematic these links are (or are not) and how our various policies applies to that issue is not clear cut, and the bot's de facto maintainer has taken steps to get that question resolved through community discussion at Wikipedia talk:Copyrights#CiteSeerX copyrights and linking with no real conclusion (ie. the community has failed to provide the necessary guidance).
- The second reason is the issue of not attributing edits to the user responsible for them, which has not been resolved. The bot now has some mitigating functionality that, when I looked at it, at least included a form field to voluntarily provide a user name that the bot would then insert into the edit summary (see https://tools.wmflabs.org/citations/). I've not paid close enough attention to tell whether this is the sum of the changes made to address this issue (I would hope the operator would answer that question). Usually this might have been "good enough" for practical purposes, but given the problems outlined above, the bot's mode of operation, and the potential for abuse (for example, during the RFC there were—alleged, not substantiated—claims that there was a dramatic uptick in removing the citation parameters in question, in an apparent effort to get them removed before the community could prohibit that behaviour: i.e. anonymous mass edit-warring using the bot), my conclusion is that in order to actually meet the requirement for attributing edits to the user making them, the bot must implement some form of actual authentication. It doesn't absolutely have to be OAuth, but it needs to be something that achieves the same effect. It should probably also not allow blocked or non-logged-in users to use the bot. But to be absolutely clear: I am here talking about potential for abuse, not any actual ongoing abuse (the bot is currently blocked, after all), and the requirements of BOTPOL.
- PS. Apologies for the wall of text. I know some people dislike that, but if I could have written this shorter I would have. --Xover (talk) 14:29, 2 May 2019 (UTC)
- some times it takes a lot of text. AManWithNoPlan (talk) 14:39, 2 May 2019 (UTC)
- Thank you for pointing out that the general community responce to my attempts at various times to start discussions has resulted in less than the sound of crickets. The whole citeceerx issue was never really resolved other than the bot stopping the addition since it is not a high value feature anyway. AManWithNoPlan (talk) 14:48, 2 May 2019 (UTC)
- some times it takes a lot of text. AManWithNoPlan (talk) 14:39, 2 May 2019 (UTC)
- "There has been no response, beyond removing the question with the edit summary
Archive aggressive comments
". Sure, if you ignore the response, and the patently obvious evidence that Smith609 operates the bot. Headbomb {t · c · p · b} 15:35, 2 May 2019 (UTC)- Easy to miss a response when it's ten days later. Fewer such assumptions might not go amiss. ——SerialNumber54129 15:40, 2 May 2019 (UTC)
- You appear to be confused. They have indeed edited, sporadically, in the interrim. As I wrote above. However the first diff you provide (which message was one of the ones included in the diff I provided) is a response to the blocking admin, not to my query, and addressed a completely unrelated issue. My query was a reply to that message and they have not provided any kind of response to that, much less actually addressed the query. The third party presuming to answer for the bot's operator (your second diff, message also included in my original diff) may well find the answer blindingly obvious, but I am not posessed of such powers of mind reading. By pure happenstance I am perfectly capable of interpreting various logs at Github, but that does not seem a reasonable requirement for resolving a question addressed to a bot's operator. Their ability to periodically click the one button it takes to merge a pull request on Github is also completely orthogonal to the question asked: do they, in fact, consider themselves the bot's operator—with the attendant responsibilities set out in BOTPOL—and can they address WP:BOTACC, second para, and WP:BOTCOMM. I do not feel that this is an unreasonable question to pose to a bot operator, and by implication of your argument it should not be a hard one to answer. And yet I have now literally waited months without an answer (the irony...). --Xover (talk) 16:34, 2 May 2019 (UTC)
- Citation bot (talk · contribs) is operated by Smith609 (talk · contribs), as it made evidently clear by the prominent {{bot}} template featured on its user page, as required per WP:BOTACCOUNT. Headbomb {t · c · p · b} 17:14, 2 May 2019 (UTC)
- That's a start. Let's get it to adhere to the rest of BOTPOL shall we. ——SerialNumber54129 17:35, 2 May 2019 (UTC)
- All the more remarkable then, that after I have asked three times, over three months, they have still not managed to affirm this, much less respond to the rest of my question. In fact, you may feel free to consider the repetiton here the fourth time, in the fourth month, that I have asked the question. --Xover (talk) 18:01, 2 May 2019 (UTC)
- Citation bot (talk · contribs) is operated by Smith609 (talk · contribs), as it made evidently clear by the prominent {{bot}} template featured on its user page, as required per WP:BOTACCOUNT. Headbomb {t · c · p · b} 17:14, 2 May 2019 (UTC)
- I could be wrong about this but my understanding was that user-initiated tools are not subject to bot policy. This came up when a well-known user-initiated tool by a well known and respected bot op was causing problems, and the operator was not responding to fix requests. A BAG member told me that it was outside the responsibility of the BAG group, they could not block it. Since this is similar to what happened with Citation bot, I'm thinking I was given bad information. Are Tools that edit Wikipedia (on behalf of a user and triggered through a web interface) subject to bot policy? -- GreenC 20:52, 2 May 2019 (UTC)
- For tools like WP:TWINKLE or WP:AWB, the answer is usually no, unless there are WP:MEATBOT concerns. Headbomb {t · c · p · b} 21:05, 2 May 2019 (UTC)
- I mean tools like Citation bot and others like it, of which there are many, where the user initiates through a web interface and the bot edits on their behalf and the user (not the bot operator) is responsible for the edit -- this being the key difference from classic bots. -- GreenC 23:05, 2 May 2019 (UTC)
- For tools like WP:TWINKLE or WP:AWB, the answer is usually no, unless there are WP:MEATBOT concerns. Headbomb {t · c · p · b} 21:05, 2 May 2019 (UTC)
- "There has been no response, beyond removing the question with the edit summary
I am going to throw this out there. The bot is approved to edit any pages it wants to (as long as it does not have a no robots tag). In fact historically it did this and even did certain categories without being asked. So, the fact that we prioritize pages a human wants us to look at is a reduction in our authorized activities. I realize this is a highly technical interpretation of the rules, but some people want to follow the letter of the law, here you go. AManWithNoPlan (talk) 13:52, 3 May 2019 (UTC)
- A quick reminder that I (like most people involved) actually have a day job, family, other volunteering, etc. Please everyone remember this during discussions, etc.. AManWithNoPlan (talk) 02:23, 4 May 2019 (UTC)
Please see this: https://en.wikipedia.org/wiki/User_talk:Citation_bot#The_current_block_is_not_well-founded_on_the_policy AManWithNoPlan (talk) 17:03, 29 May 2019 (UTC)
Inactive bots - May 2019
Per the bot policy activity requirements, the following bots will be deauthorized and deflagged in one week. These bots have not made an edit in 2 or more years, nor have their operator made an edit in 2 or more years.
BOT_user_name | BOT_last_edit | Oper_username | Oper_lastedit | Notes |
---|---|---|---|---|
User:StatisticianBot | 20161104065821 | User:Dvandersluis | 20161019 | |
User:DYKReviewBot | 20161030205038 | User:Intelligentsium | 20161030 | |
User:DefconBot | 20160902053125 | User:A930913 | 20160403 | |
User:Mr.Z-bot | 20160830220939 | User:Mr.Z-man | 20160821 | |
User:BracketBot | 20160719215737 | User:A930913 | 20160403 | |
User:DrTrigonBot | 20150617013726 | User:DrTrigon | 20160626 | |
User:DixonDBot | 20130329214425 | User:DixonD | 20170204 | |
User:MGA73bot | 20130202213645 | User:MGA73 | 20160925 | |
User:Lucia Bot | 20121116225341 | User:Beria | 20170501 | |
User:Ryan Vesey Bot | 20120928012455 | User:Ryan Vesey | 20170308 |
Should an operator wish to maintain their bot's status, please place "keep" and your signature in the notes column above for your bot. Deauthorized bots will need to file a new BRFA should they wish to become reactivated in the future. Thank you, — xaosflux Talk 03:27, 14 May 2019 (UTC)
- Required user talk notices left. — xaosflux Talk 03:33, 14 May 2019 (UTC)
- Discuss
- With no comments by the operators, the above accounts have had their bot flag removed. Primefac (talk) 20:22, 23 May 2019 (UTC)
Naming conventions
Okay, so before I go ahead and say something that either ends up causing more confusion, and/or making it seems like I'm the absolute authority in this case...
Background: I noticed at this BRFA that there was a question about keeping the absolute numerical numbering of bot tasks even though there are three different bots being numbered (i.e. "Task 13" is run as bot II's first task), and the above Task 38 might be run by bot III (even though it's the first bot run by that bot).
I suppose my question is, should we allow this sort of "absolute" numbering between different bots run by the same operator? Does it matter? As an arbitrary example, the following could be a set of task requests:
- GenericBot
- GenericBot 2
- GenericBot 3
- GenericBot II 4
- GenericBot 5
- GenericBot II 6
I genuinely don't have a position on this (and said so in this discussion), but as mentioned above I'd like to not give advice that's either contrary to what should be done and/or what people expect. Primefac (talk) 20:20, 23 May 2019 (UTC)
- In my experience, each bot gets its own list of tasks. See User:BattyBot, for example. In the case that you linked, a single operator is proposing to run multiple bots (you might call these bots "siblings"). In that case, in order to preserve everyone's sanity, I would love to see a single User page for the set of bots, and a single list of tasks, even if different tasks are performed by different Bot-siblings. Just include a column in the table that shows which bot performs each task. In short, I think it would be the most sane option to do it as you have shown in the list above. – Jonesey95 (talk) 20:55, 23 May 2019 (UTC)
- There's a very small amount of past precedent: AnomieBOT, MusikBot, ClueBot and SoxBot restart their numbering when they go to "II", but Cyberbot doesn't. I suppose "It doesn't really matter" is a valid outcome of this discussion. Primefac (talk) 21:06, 23 May 2019 (UTC)
- @Jonesey95: Does User:DannyS712 bot/tasks meet your needs? Its transcluded by all of my bots --DannyS712 (talk) 21:08, 23 May 2019 (UTC)
- Case by case is the way to go here. First, it is OK to "skip" task numbers (especially as some may not get approved) so having Bot,Bot 2, BotB 3, Bot 4, BotB 5 isn't a problem that some were "skipped". In general, it is a good idea to have the BRFA subpage name==the bot actual name - over very long periods of time we've seen bots be renamed, bots change operators, etc. None of this is a "big deal". Within an actual BRFA, the actual bot name of whatever it is should always be used. If someone has a suite of bot accounts, using a centralized page name (that is redirected from the other accounts) is also OK. — xaosflux Talk 00:23, 24 May 2019 (UTC)
SineBot inactive since 2019-04-30
SineBot's latest contribution is from 2019-04-30T06:38:33. See User talk:Slakr#SineBot down.
@Masumrezarock100: I'd say the issue is not urgent enough to require other users' assistance. As the bot's code sadly is not available to the public, other users can not help by running the bot in the meantime.
~ ToBeFree (talk) 16:44, 26 May 2019 (UTC)
- @ToBeFree: Is there any alternate bot to sign comments? It's a pain now that those unsigned comments are not automatically signed. Teahouse suffers the most. Masum Reza📞 16:49, 26 May 2019 (UTC)
- {{Xsign}} makes it relatively easy compared to other templates, but I agree that SineBot is extremely useful. I miss it dearly. ~ ToBeFree (talk) 17:09, 26 May 2019 (UTC)
- User:Anomie/unsignedhelper.js is great for unsigned comments. Primefac (talk) 12:40, 27 May 2019 (UTC)
- {{Xsign}} makes it relatively easy compared to other templates, but I agree that SineBot is extremely useful. I miss it dearly. ~ ToBeFree (talk) 17:09, 26 May 2019 (UTC)
Issue with a bot, not finding the bot maintainer's response satisfactory, not sure what to do now.
See Wikipedia talk:Username policy#Utility of reports by DatBot. This bot reports users at UAA for being possible sock puppets. 99% of the time they are not blatant violations of the username policy, which is absolutely the only thing UAA is for. I was under the impression that bot tasks all had to be approved so I'm curious as to whether this specific task was approved and if so, why?
When questioned about it the reply from the bot's maintainer was along the lines of "I think I might remember why I did this and since many of the users wind up blocked it's clearly working."[1] I think this ignores several pertinent facts:
- There is no evidence that the UAA reports are in any way what leads to eventual blocks, let alone global locks, of these accounts
- If you don't seem to know why you coded the bot to do certain things, when those things are challenged it seems reasonable to change the bot's behavior
- It's a blatant misuse of UAA to report socks there, it is in now way a forum for anything other than blatant violations of the username policy
I would therefore appreciate input in the discussion at WT:UPOL from BAG members about how to proceed here. I don't have anything against DatGuy and I'm sure his bots do lots of helpful things, but this particular thing does not seem helpful and just creates noise in an administrative area that regularly experiences backlogs. Beeblebrox (talk) 19:41, 30 May 2019 (UTC)
- @Beeblebrox: can you place a couple of Diff's here for edits you think this bot is making that you don't think have an approved BRFA? Then the operator can be called to identify if there is a BRFA task approved for them or not. If there is, but things have changed over time you can ask for a re-evaluation of the BRFA here as well. As this is all "back page" type stuff, I don't suggest blocking while this is sorted out. — xaosflux Talk 20:08, 30 May 2019 (UTC)
- All the reports are like this: [2]. The creation of an account trips an edit filter and the bot reports it to UAA. Typically these accounts have no edits, which is something we normally discourage reporting at UAA unless it is something like hate speech or attacks on specific people, all the information it provides is "possible sockpuppet creation" which isn't in any way a username violation. Beeblebrox (talk) 20:37, 30 May 2019 (UTC)
- As it happens, it appears that DeltaQuadBot is not making reports right now, so for the last few days all the action WP:UAA/B has been this bot making reports and admins removing them without blocking because literally none of them are based in any way on the username policy. Beeblebrox (talk) 21:21, 30 May 2019 (UTC)
- For the record, the relevant BRFA is this one, as a takeover of this one. Headbomb {t · c · p · b} 21:30, 30 May 2019 (UTC)
- @Headbomb: is my reading of that that these UAA reports should only be getting triggered from filter 579? — xaosflux Talk 22:00, 30 May 2019 (UTC)
- And if so, @Beeblebrox: can you think of a better place for "Possible sockpuppet account creations" to go (or do you think this is useless)? — xaosflux Talk 22:02, 30 May 2019 (UTC)
- Without context I don't know how anyone can evaluate these reports other than the maintainers of that specific edit filter. It doesn't mention who's sockpuppet the reported user might be, and it reports them before they've actually done anything so there's nothing to go on without running a (completely unjustifiable) checkuser on them. If there was some way to post these reports to pertinent pages at SPI so that people who actually knew what the context was could review them that would be something, but they serve no purpose at all at UAA. Looking at that BRFA, it seems it was going to provide context by linking to relevant SPI pages, but it doesn't do that in practice. But even if it did, this just doesn't belong at UAA, it belongs at SPI if ti belongs anywhere. Beeblebrox (talk) 23:26, 30 May 2019 (UTC)
- And if so, @Beeblebrox: can you think of a better place for "Possible sockpuppet account creations" to go (or do you think this is useless)? — xaosflux Talk 22:02, 30 May 2019 (UTC)
- @Headbomb: is my reading of that that these UAA reports should only be getting triggered from filter 579? — xaosflux Talk 22:00, 30 May 2019 (UTC)
- For the record, the relevant BRFA is this one, as a takeover of this one. Headbomb {t · c · p · b} 21:30, 30 May 2019 (UTC)
- As it happens, it appears that DeltaQuadBot is not making reports right now, so for the last few days all the action WP:UAA/B has been this bot making reports and admins removing them without blocking because literally none of them are based in any way on the username policy. Beeblebrox (talk) 21:21, 30 May 2019 (UTC)
- All the reports are like this: [2]. The creation of an account trips an edit filter and the bot reports it to UAA. Typically these accounts have no edits, which is something we normally discourage reporting at UAA unless it is something like hate speech or attacks on specific people, all the information it provides is "possible sockpuppet creation" which isn't in any way a username violation. Beeblebrox (talk) 20:37, 30 May 2019 (UTC)
- OK, so it seems this is "easy" to turn off (at User:DatBot/filters) - @DatGuy: looks like there is pushback on this one report job, do you have any issues with it being removed? (Perhaps one day a better filter can be made to feed that task). — xaosflux Talk 23:44, 30 May 2019 (UTC)