User talk:ClueBot Commons/Archives/2013/March
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
F words with space
Maybe Cluebot could catch cases such as this one? ("i s u c k d i c k" was the string added). -- Magioladitis (talk) 15:56, 28 February 2013 (UTC)
- Cluebot doesn't work on a keyword basis anymore; it uses an artificial neural network. I must say, though, that should have been caught... If somebody has archives for the Cluebot IRC feed, perhaps we can see what score Cluebot gave that edit. Calvin (t·c) 03:00, 4 March 2013 (UTC)
Multiple warnings
Just letting you know that here the bot created a new "March 2013" after there was already one created three days earlier, and jumped back to level 1 warning after a level 2 warning. --ELEKHHT 04:41, 7 March 2013 (UTC)
Clarification for IPs
After I reported a false negative to ClueBot, I looked at some of ClueBot's recent edit and the responses to them. I saw an Anonymous report about an edit made in 2010. Anonymous was commenting about an edit made by 66.183.239.217 (talk · contribs). This IP has made only one edit, back in December 2010, and the edit was clear vandalism. ClueBot caught the vandalism, reverted it, and left a comment on the IP's talk page. My concern is about what happened next. On February 9th, 2013, someone using this IP address came to Wikipedia and apparently saw the new comment message generated by the comment ClueBot left and followed the links until they came to the false negative reporter, where they reported that "The edit was not mine, and I have never edited any page of wikipedia, ever." To me, this brings to light the issue of how ClueBot comes across to the (probably many) people who get this message but didn't make the edit in question. Maybe something could be added to the FalsePositives page with a short explanation of ClueBot and what reporting false negative means with a line or two aimed at IP editors. A message from ClueBot may be the first introduction to the Wikipedia community for editors, potential editors and even readers who are taking a peek under the hood. While it's not able to be as positive, as helpful or as user-friendly in the way that other intros are, it could be improved to at least take into consideration these users. Also an explanation of ClueBot and what reporting means would be helpful for non-newbies as well. I somehow got the impression (I'm not sure how, exactly) that reporting a mistake to ClueBot would undo the Cluebot edit. The bottom of the page did make it clear to me that FP reports are used to improve the bot, but perhaps including a line that explains that reporting an FP does not restore the edit and that editors have to do that themselves would be advisable. Wherever there are institutions with clear processes in place there are editors who are unclear about they can do, what it's acceptable to do, what they should do. 76.171.22.15 (talk) 17:56, 4 March 2013 (UTC)
- ClueBot already provides information about false positives though in the warning message. In my opinion, that's enough. If people aren't going to click on a link to read about them then they're not going to pay attention on the reporting false positives page.--5 albert square (talk) 23:41, 4 March 2013 (UTC)
- I know but that doesn't really deal with the issue I'm talking about. Let's say I've used Wikipedia a lot but never edited and I want to see what it's about. I click over and I have a new message. And that message says that my edit was reverted by a bot and explains what to do if I think a mistake has been made. Well I think a mistake has been made, that isn't my edit and the information that's available to me just keeps reiterating that people think that it is. I'm told that the bot isn't biased against me, and could have been wrong, but there's no information anywhere explaining why the bot thinks that I made the edit or what to do about it. Adding information to either area or the talk page comment that's generated, maybe with a link to another intro page (the benefits of having a username for example) would be beneficial to users who aren't (or are new) editors and it doesn't hurt established editors so why not? 76.171.22.15 (talk) 04:33, 5 March 2013 (UTC)
- There is a statement to that effect ("This is the discussion page for an IP user, identified by the user's numerical IP address. Some IP addresses change periodically, and may be shared by several users. If you are an IP user, you may create an account or log in to avoid future confusion with other IP users. Registering also hides your IP address.") at the bottom of all IP talk pages. A clearer one that might be what you are looking for appears below the notice when an IP is blocked: "If this is a shared IP address, and you did not make the edit, consider creating an account for yourself so you can avoid further irrelevant notices." Also ClueBot always puts its notice under a section header giving the month and year, as well as dating it - so it should be reasonably clear that very stale ClueBot notices are about someone else who previously had that IP. I see your point, but how can it be made clearer? I think the problem may really arise from someone not realizing their IP is dynamic, and the bottom of the page actually does explain that. Yngvadottir (talk) 05:04, 5 March 2013 (UTC)
- The date header doesn't really deal with the problem. If I haven't edited Wikipedia then I not only didn't edit on that date I didn't edit on that article. If I've made a few edits but not that one, then same thing. Not realizing that your IP is dynamic or that you're being recognized as an IP is a real thing. And the problem with the bottom of the page box is that it is so very hard to notice. The page name which includes the IP address is easier, but my eyes go directly to hello. I think the best way to address it would be to add a bullet point to the talk page message and have it link to a section on the talk page that explains briefly about how Wikipedia works for users without an account and gives them a couple links to helpful places. Just a thought. 76.171.22.15 (talk) 07:31, 8 March 2013 (UTC)
- There is a statement to that effect ("This is the discussion page for an IP user, identified by the user's numerical IP address. Some IP addresses change periodically, and may be shared by several users. If you are an IP user, you may create an account or log in to avoid future confusion with other IP users. Registering also hides your IP address.") at the bottom of all IP talk pages. A clearer one that might be what you are looking for appears below the notice when an IP is blocked: "If this is a shared IP address, and you did not make the edit, consider creating an account for yourself so you can avoid further irrelevant notices." Also ClueBot always puts its notice under a section header giving the month and year, as well as dating it - so it should be reasonably clear that very stale ClueBot notices are about someone else who previously had that IP. I see your point, but how can it be made clearer? I think the problem may really arise from someone not realizing their IP is dynamic, and the bottom of the page actually does explain that. Yngvadottir (talk) 05:04, 5 March 2013 (UTC)
- I know but that doesn't really deal with the issue I'm talking about. Let's say I've used Wikipedia a lot but never edited and I want to see what it's about. I click over and I have a new message. And that message says that my edit was reverted by a bot and explains what to do if I think a mistake has been made. Well I think a mistake has been made, that isn't my edit and the information that's available to me just keeps reiterating that people think that it is. I'm told that the bot isn't biased against me, and could have been wrong, but there's no information anywhere explaining why the bot thinks that I made the edit or what to do about it. Adding information to either area or the talk page comment that's generated, maybe with a link to another intro page (the benefits of having a username for example) would be beneficial to users who aren't (or are new) editors and it doesn't hurt established editors so why not? 76.171.22.15 (talk) 04:33, 5 March 2013 (UTC)
Just Trying to help
There's more info in this article and I know one of the articles stated Trap music is associated with violance and money aswell as drug dealing, other than there's more info to help you — Preceding unsigned comment added by 121.219.97.112 (talk) 08:52, 11 March 2013 (UTC)
Using STiki, Continually Flags Admin Edits
I use STiki with the Cluebot NG revision queue, and I've run into many (>5) instances approximately in a row where I've reviewed an article (and found it innocent) that was flagged despite being contributed by an administrator (all articles were quite similar "surname" disambiguation changes and edited by the same admin). I am not sure if I am bringing this to the attention of the proper people, hence not reporting false positives; the changes could certainly have aroused suspicion, and I don't think administrators should be immune from review by any means, but I was wondering if someone could please explain the process to me (I already read the FAQ) as to how this might happen with such frequency and if there is a measure in place that takes into consideration previous flags that turned out to be innocent by a user or a user's rights. --Jackson Peebles (talk) 17:25, 12 March 2013 (UTC)
Sorry
Sorry. Accidental. Weeeeg (talk) 22:16, 13 March 2013 (UTC)
Lynching
Your article is full of incorrect information. It's title is totally WRONG & INAPPROPRIATE. It's neither accurate or factual. Look at the negative comments on YOUR talk page. Such LOW importance, et cetra. And only a jerk would ban me or subjugate me by improperly removing my notations. — Preceding unsigned comment added by MSNanCaroL (talk • contribs) 03:29, 14 March 2013 (UTC)
- Thanks for your message. Presumably you posted it here because ClueBot NG reverted an edit that appeared to be unconstructive. Please understand that ClueBot NG does not own the article, nor does anyone else. You are welcome to improve the article and/or discuss issues on its talk page. Please do not add your comments to the article itself. – Wdchk (talk) 12:10, 14 March 2013 (UTC)
Duplicate L1 warnings
I'm sure this has been discussed before, but I often see Cluebot starting over with level 1 warnings after I've already warned a user in the same month. Example here Andrew327 23:46, 14 March 2013 (UTC)
Just wondering...
I was just wondering how/why Cluebot NG missed this (to a human) obvious vandalism. I don't really understand what an "artificial neural network" is, but I was surprised that this was live in the article for over 30 minutes. Was it the use of a false edit summary? Surely any edit to the name (in bold in the first line of the article and in the infobox) to make it not at all match the article name (ie not just adding a middle name etc) should be a huge vandalism flag. The-Pope (talk) 11:56, 17 March 2013 (UTC)
Question
Cross-posted from my talk:
On an unrelated note, I was wondering about something. Think about a hypothetical situation: A vandal vandalizes a page with things that Cluebot NG would normally detect. Then, they put the
{{nobots}}
template on the page, so Cluebot will not revert it. This hypothetical vandal does this sneakily, so the vandalism sneaks by.
If this hypothetical situation every arises commonly, would there be enough editors doing Recent Changes Patrol to take care of this? Thekillerpenguin (talk) 05:54, 18 March 2013 (UTC)
—Theopolisme (talk) 11:00, 18 March 2013 (UTC)
- Perhaps the bot should check to see if the tag was added in the most recent edit, and if so, revert? 69.255.179.102 (talk) 00:50, 19 March 2013 (UTC)
Another question concerning repeated vandalism
I notice that ClueBot reverted two vandalism edits by the same IP and instead of issuing a Level 2 warning for the second, changed the Level 1 warning to refer to the other article, with this edit. I'm wondering why? (The bot also missed the same IP vandalising at 2 other articles, but that happens.) Yngvadottir (talk) 16:50, 19 March 2013 (UTC)
Pending Changes level 2 on /run
User:ClueBot NG/Run is, as I understand it, a page designed to allow trusted users the ability to stop the bot if it is malfunctioning. It is currently fully protected due to it being a target of autoconfirmed vandalism. This unfortunately makes the page redundant as any editor that can edit it (sysops) can also stop the bot via their block button. Would pending changes level 2, with semi-protection be a viable alternative? If the bot was set to read the accepted version of the page (not the changes pending version), this would allow only reviewers the ability to stop the bot, keeping out autoconfirmed vandalism, and returning the page to its original function. Regards, Crazynas t 23:37, 21 March 2013 (UTC)
Error on Safe Haven (film)
On Safe Haven (film), Boxoffice (magazine) is not the same as Box Office Mojo. Please correct your bot. 108.73.115.113 (talk) 02:58, 23 March 2013 (UTC)
A barnstar for you!
{{/censor}}
Not sure if this is a false positive, but...
On Finley Quaye (edit | talk | history | protect | delete | links | watch | logs | views), the subject of the bio added some information, which was revered by ClueBot. I am not so much complaining about the revert, since I would have done that myself anyway (unreferenced, primary source), but I'd like to understand what triggers ClueBot in this case. Is it the username? He unfortunately picked Finley quaye (talk · contribs), which theoretically shouldn't have been possible, unless he went through the account creation service. But regardless, I'm just wondering, what was the trigger here? §FreeRangeFrogcroak 19:13, 29 March 2013 (UTC)
- Not that I train ClueBot, but I suspect the bot thought of this as random blanking. Yngvadottir (talk) 19:27, 29 March 2013 (UTC)
- Oh you are right, he didn't add any material, he removed it. I didn't pay attention to that. So that's probably it, removing info from a BLP likely did it. Well, I can't say I have a problem with what he did, since all of that was unsourced either way. §FreeRangeFrogcroak 19:36, 29 March 2013 (UTC)
- If you're interested you can check out NG's userpage for details of its implementation. The bot isn't a hard-coded set of rules as such, instead it uses a machine learning scoring mechanism, so for each edit a large number of variables are considered and a single score produced, and if that score is above the set threshold the edit is reverted. Hence it's not entirely straightforward or accurate to pick out any one factor and say "this is why this edit was reverted" Jebus989✰ 20:24, 29 March 2013 (UTC)