Wiki143:Bots/Requests for approval/Archive 1

Please add new archives to Archive 2.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Tawkerbot

Tawkerbot is a python wikipediabot from http://meta.wikimedia.org/wiki/Using_the_python_wikipediabot. It will be running subst: on a list of userpages that I have created and checked that have templates without the subst tag.

I'm taking lists of user pages that have not used subst: on common warning templates and adding the subst in. I will be using a text file of configs.

= Config params given

As per the instructions, Tawkerbot will be run with the commands

python replace.py -file:articles_list.txt "{{test2}}" "{{subst:test2}}

with the articles_list.txt being a file created by myself and the template being the template that needs to be edited.

Sorry that I started running the bot, I misunderstood the bot procedures, thankfully I had 20 people on the IRC room trying to tell me I hard errored.

Update: Tawkerbot presently has an option for auto reverting the squidward mass vandal, I know it is a little step up from the reasons I applied for the flag however, I'm not expecting any complaints (nobody on IRC expressed any concerns besides the innocent page revert which I explained) so it is enabled in case squidward returns. Tawkerbot will only auto revert if one follows the exact conditions of the attack, it should be virtually impossible for an innocent edit to be caught by this process

Tawker 02:35, 13 February 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

LDBot

LDBot is a self named bot (Lightdarkness) which is to be created to do a few tasks. It is based off PHP and the Pywikipedia framework. The purpose of the bot would be to maintain Wikipedia:List_of_non-admins_with_high_edit_counts. The bot would check the edit counts via the toolserver of each user once a week (Spaced out in equal incriments, as to not cause stress on the server) and once every other day, update the counts at the previously mentioned page. The bot is being designed to make the stat tracking of that page automated, so it doesn't require user upkeep.

The Bot will not strain wikipedia servers at all. It will make no more than one request to the toolserver (or a wikimedia server) per hour, with updating to take place once every other day. If this behavior were to ever change, I would first bring it up here, or to AllyUnion to be sure that the bot is still within operating limits.

I've recieved the support of an active admin on that page, and another admin, which is why I've started development on this bot. Development will occur in my Usernamespace, and willnot edit the main page until the bot is approved, and testing is complete. --lightdarkness ^(talk) 03:09, 17 February 2006 (UTC)

Hvae you asked the people who run the tool servers if checking all edit counts once a week would stress their servers? Martin 09:35, 17 February 2006 (UTC)

I haven't specificly asked, I just came to the conclusion that it was common sense because I know several users who check their edit count using Interiots tool several times PER hour. I'm also very certain that Interiot uses the SQL function COUNT() which reduces the amount of processing power, but I will ask Interiot and the other toolserver admins if indeed it will be no trouble. --lightdarkness ^(talk) 12:05, 17 February 2006 (UTC)

A question, currently some users bold their user names, for Users in bold have shown an interest in becoming admins.. What do you think about this, will you add this feature.--Ugur Basak 13:13, 17 February 2006 (UTC)

Of course! Syntax for editing the users interest is as follows: "LDBot mod Lightdarkness <bold>" for making a user bold, if a user is not interested, you'd use the following: "LDBot mod Ugur_Basak (Not Interested) <strikethrough>" Where a comment is given within the parenthises, and the format type is given in < >. --lightdarkness ^(talk) 14:06, 17 February 2006 (UTC)

This sort of thing would be much more efficient to implement on the toolserver itself (eg. a slight modification of this SQL query), and it would probably much less work to just pester someone who has a toolserver account than to write the code for the bot. Even better, wikisign.org has a database dump that's currently 10 days old, and you can run that query there. All that aside, a hit or three an hour isn't going to be noticable, but there may be better ways to do it. --Interiot 17:18, 17 February 2006 (UTC)

The thing about it being on the toolserver though, is that users can't add their own comments on if they want to be an admin, no interest, links to previous nominations, ect. You do bring up a good point though, and I'll ask you about it on your talk page. --lightdarkness ^(talk) 21:36, 17 February 2006 (UTC)

A single-user count is available here, but the tool specifically says "Please note, that mass automated querying of the edit counter (for example, to generate lists of user sorted by edit count) is not allowed.". eg. it's better to pester someone with a toolserver account to get the raw data that you want, with a single query, since that will only take 3 - 5 minutes to generate. --Interiot 22:01, 17 February 2006 (UTC)

But the thing is, the stats will be constantly updated over a period of time, I wouldn't think that one query per hour would be trouble, but I'm very open to ideas as far as what type of things toolserver access would attribute to this project. --lightdarkness ^(talk) 22:46, 17 February 2006 (UTC)

Isn't it easier for the report to be generated on the toolserver, with LDBot copying the results once per day and overwriting the local Wiki page? Or am I misunderstanding something? Titoxd^{(?!? - help us)} 01:28, 18 February 2006 (UTC)

That would pretty much do the same thing, I'm just using my personal server to do all of the data processing, but with toolserver access I could just get the counts and update there, but I'm not sure the toolserver has access to a mysql database to write to (which I'm now using on mine for storage), but I'll look into it. --lightdarkness ^(talk) 03:12, 18 February 2006 (UTC)

The database in the toolserver does not have write access to the Wikimedia databases, so I don't think that is possible. However, the same effect can be done by making the raw query in the toolserver (which is much less expensive than querying Kate's Tool or doing anything like that), and then have LDBot read the file and process it slightly for posting in the local Wikipedia page. Titoxd^{(?!? - help us)} 05:50, 19 February 2006 (UTC)

I knew that the toolserver can't write to the wikipedia database, but the way I'm storing all the edit counts is on my own database, once I query Interiots tool for them. This way, I can space them out rather than gathering all the data at runtime. Even if it were on the toolserver, being run all at once would probably take quite a bit of processing power, and I wouldn't want to disrupt the toolserver. --lightdarkness ^(talk) 05:55, 19 February 2006 (UTC)

After re-reading all of the comments here, I thought I would just sum everything up. Interiot stated that 2-3 hits per hour wouldn't affect the toolserver, but suggests that there could be a better way of going about it. However, the current layout of WP:NA has a lot of information, regarding notes about users seeking adminship, links to previous noms, ect, that the toolserver cannot gather by doing a simple query. Which brings me to my conclusion. The concensus on the talk page of WP:NA suggests everyone is for the bot implimentation. Due to the limitations of the toolserver for this layout, I'll just be querying the toolserver once every hour for an updated editcount of the user whose editcount was updated the longest ago. Every night, my bot will grab commands from a specialized page in LDBot's usernamespace for commands to add, modify, delete users from the master list, and then LDBot will blank the page, signaling that the commands have been processed. Then, if any updates are required, LDBot will update the pages of the appropriate sections. I hope this wasn't too long, but this is the best way I can see doing it while automating the process and keeping the output the same. --lightdarkness ^(talk) 21:48, 19 February 2006 (UTC)

It would be simple enough to generate a report on the toolserver in a machine-readable format to provide this data; this could be kept up to date via a cron job; Interiot hinted above, I think, that he'd be willing to help with this. The thought process I'm in here is that this method is going to produce outdated information.

Nevertheless, approved for a trial run. Rob Church 01:12, 30 March 2006 (UTC)

Thanks Rob, I appreciate it. The bot is going to start it's trial run Friday evening. I'd also like to request one more thing from here. It's been noted that AFDBot hasn't been operational since mid-March, and AllyUnion hasn't been around to fix it. I've offered to write another feature for LDBot to update the AFD pages as appropriate. I've put a note about it on WT:AFD, and am waiting to see if there aren't any objections. If after 5 - 7 days there aren't any objections on that page, would anyone have a problem with LDBot taking over this task? At least until AllyUnion gets back? --lightdarkness ^(talk) 13:33, 6 April 2006 (UTC)

No objections, it would be great if you could take over that task. Martin 14:01, 6 April 2006 (UTC)

• Please note that the WP:NA functions of the bot have been discontinued, as the toolserver isn't getting updates from en_wiki. Have applied for a bot flag for the purpose of AFD. --lightdarkness ^(talk) 03:18, 13 April 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

JoeBot

Hello everybody! I would like your permission to run my little JoeBot. It replaces common misspelled words, like 'begining' (beginning). I sort of started doing it on my own account and then realized that if I wanted to continue doing it I would need a bot account because of the sheer massive amount of these misspellings. So here I am! Wikipedia is getting so large that little mistakes fall through the cracks, and I'd like JoeBot to be able to be there to catch em'.

The bot will be manually run by yours truly, using the mighty fine AutoWikiBrowser by Martin. I'll be using it with the JoeBot account for a trial week to earn my spurs. It will only check the english wikipedia. JoeSmack ^Talk 18:39, 19 February 2006 (UTC)

And yes, I am truly aware of regional spellings and pay special attention not to replace these. (example: 'color' and 'colour') JoeSmack ^Talk

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Closed with no consensus. Martin 09:04, 11 April 2006 (UTC)

Bobblebot

I would like to use a bot to delink date links.

The task is to reduce linking of solitary months, solitary years, etc in accordance with the Manual of style. It will not delink date links used for date preference. Each edit will be checked visually.

For example Economy of Algeria has many solitary year links, including 14 to 2004. The policy and popular misunderstandings are explained at:

• Wikipedia:Manual_of_Style_(dates_and_numbers)#Date_formatting
• Wikipedia:Manual_of_Style_(links)#Dates_and_numbers
• Wikipedia:Make_only_links_relevant_to_the_context#What_should_not_be_linked

For example:

• The short-lived ABC Cable News began in 1995; unable to compete with CNN, it shut down in 1997. Undaunted, in 2004 ABC launched a news channel called ABC News Now.

will become:

• The short-lived ABC Cable News began in 1995; unable to compete with CNN, it shut down in 1997. Undaunted, in 2004 ABC launched a news channel called ABC News Now.

See this diff working towards that: [1].

Several editors were using Martin's excellent AutoWikiBrowser to do this task. That was a good solution that just got better and better. The shared experience of many users produced a 'whole is greater than the sum' process of logic improvements. I would be more than happy to share or give away the date delinking task to AutoWikiBrowser if it is permitted to do it.

A previous application on 15 December did not succeed. User:Sam Korn suggested that I should reapply and make clear that I look at each edit. So this is my re-application. Thanks. bobblewik 19:48, 25 February 2006 (UTC)

• Oppose, Bobblewik has said in the past that he checks each edit yet his talk page has plenty of examples of where he has clearly not taken enough care to check that his bot script has not done something silly. I don't believe this will be any different. I am also opposed in principle to using bots to "improve" articles. Articles are organic things and rigid scripts rarely help. Talrias (t | e | c) 19:54, 25 February 2006 (UTC)

  By the way, what has changed since the last application? Talrias (t | e | c) 20:01, 25 February 2006 (UTC)

  Basically, it was my perception (and Bobblewik has taken my advice, so please shoot me first :-) that most complaints were more due to concerns that the edits would not be monitored. Is there any reason why Bobblewik should not be allowed another chance? Sam Korn ^(smoddy) 20:44, 25 February 2006 (UTC)

• I have no problem remove 13 links to 2004 in the first example but do not agree that the links are irrelevant in the second example. Worse, I have seen no effort on your part to determine which links are relevant - only mass removal. I understand you do not like these links. I also understand that some other people do. I don't see why a bot is required in this controversy. Rmhermen 19:57, 25 February 2006 (UTC)
• So long as Bobblewik undertakes that he will look very carefully at complaints about mistakes and will examine edits carefully, I see no reason why this should not happen. Support Sam Korn ^(smoddy) 20:44, 25 February 2006 (UTC)
• I support a bot that makes these types of changes, as i did in Decemeber. i hate seeing over linked articles with respect to dates. David D. (Talk) 20:54, 25 February 2006 (UTC)
• Support. I can't see what the damage could be. No information could be lost. As it stands no one could go to a date to find information anyway. Caution should always be exercised in using bots. --Wetman 20:58, 25 February 2006 (UTC)
• Strong support. There are so many articles which overlink dates, I often fix these manually and comment about it in peer reviews, etc. it would be great to have a bot get rid of these overlinkings. --Fritz Saalfeld (Talk) 21:06, 25 February 2006 (UTC)
• Support, provided that the changes only delink dates and don't otherwise affect their formatting. (I'm particularly concerned about [[1941]]-[[1944|44]] becoming 1941-44, and not 1941-1944.) —Kirill Lokshin 21:09, 25 February 2006 (UTC)

• Support/Oppose. Bobblewik's edits are manually checked, therefore they are at most robot assisted. Since we are all fallible, there is no reason these edits should not show on recent changes so a bot-flag should not be granted. Bobblewick is restricting his changes to 2 per minute - slower than some of his critics. If giving him bot-status will allow him to continue editing, then I support his request. If on the other hand, he should be permitted to edit at two changes per minute (or a faster rate) in any case, then I would oppose the bot status, and say let him get on with the editing. If someone has a problem with any of his edits, they can discuss them or revert them individually, and follow the normal dispute resolution procedure if that doesn't work. Rich Farmbrough 21:17 25 February 2006 (UTC).
• Support, let the bobblebot role - get rid of needless date links. Vsmith 21:22, 25 February 2006 (UTC)
• Support. Even if it delinked every year in Wikipedia it would probably be doing more good than harm, considering how bad date overlinking is. Kaldari 21:35, 25 February 2006 (UTC)
• Support. My only concern is that there seems to be an army of "robots" who are just going to mindlessly continue to turn every year and date into a link. It seems like there needs to be a way to get the people who make all these useless links to stop doing so. --JWSchmidt 21:39, 25 February 2006 (UTC)

• Support. I do want to see those year links reduced. -- Donald Albury (Dalbury)^(Talk) 21:54, 25 February 2006 (UTC)
• Support. As long as we are agreed that overlinking of dates is a Bad Thing, i.e. that the MOS isn't in dispute, it makes sense to grant Bot status to any program that will bring the Wikipedia closer to the MOS. Consider the possible good vs. the worst possible harm, as Kaldari did, above. It's hard to imagine a scenario in which net harm would be done.—GraemeMcRae^talk 22:05, 25 February 2006 (UTC)
• Support - All Bobblewik is trying to do is conform to policy, which is a noble goal. We should be thankful we have someone who is willing to do this nitty gritty work; I certainly don't enjoy going through hundreds of pages and doing it manually. Even though he is going AWB-assisted, it is manual in the end, so kudos to him. As an answer to someone who said he shouldn't get the bot flag because these are manual: he's had people repeatedly blocking him because he's "editing too fast" for a non-bot. If he doesn't get this bot flag he's just going to face the same kind of misunderstanding and repeated disruptions of his work. --Cyde Weys 22:16, 25 February 2006 (UTC)
• Support, I've made my views on this clear before. Stephen Turner (Talk) 22:25, 25 February 2006 (UTC)
• Strong Support. As others have said, conforming to the Wikipedia policies is a good thing, and it helps us get this sort of maintenance/MoS work done quicker. — Wackymacs 22:37, 25 February 2006 (UTC)
• Support. – Quadell ^{(talk) (bounties)} 22:48, 25 February 2006 (UTC)
• Support. Unnecessary links are a nuisance and bobblewik wants to do something about them. He has taken enough flak for his efforts and deserves the support of the community. Stroika 22:57, 25 February 2006 (UTC)
• Support --Duk 23:08, 25 February 2006 (UTC)
• Oppose - Works on the nerves of many people (also this whole "crusade" idea, so comparable to the "conversion to BC/BCE" crusade, to the "conversion of references" crusade, etc - the proponents of these crusades all advocated application of guidelines as a justification of their crusade, all these were stopped by ArbCom leaving them on probation without permission for further conversion ***neither by bot nor manually***). This bot proposal would mean Bobblewik is "approved" to come in on any article (say, for instance, Furniture music), and decide which are the significant dates, and which aren't. Sorry, Bobblewik *may* be aware of which the significant dates are for several ranges of articles, but (s)he should not get a ticket to be allowed to meddle with any article (s)he pleases. Further, I take it as a principle that anyone who goes around inviting only parties who previously expressed sympathy to take part in a vote, should better not be granted anything. --Francis Schonken 23:47, 25 February 2006 (UTC)

• What are you talking about, only parties who previously expressed sympathy, Bobblewik contacted some of his harshest critics, including admins who have blocked him before. Go take another look. --Duk 00:16, 26 February 2006 (UTC)

• Comment. Bobblewik has just spammed over 50 user talk pages (with the edit summary "Date links") about this bot account application. I'm assuming Bobblewik was using AWB, because a rate exceeding one edit per 10 seconds was reached.--Commander Keane 23:26, 25 February 2006 (UTC)

  Bobblewiki is not presently enabled to use AWB, plus all edits made with it reference the AWB page in the summary. Martin 23:32, 25 February 2006 (UTC)

  I've done that rate with page moves and talk page notices. Look at successful admin candidates' edit rates immediately after their promotion. 10/minute is not difficult if all the edits are identical, and if you have a reasonable connection. Sam Korn ^(smoddy) 23:36, 25 February 2006 (UTC)

  OK, let's stop this vote, and start an RfC on Bobblewik, what were we waiting for all the time? --Francis Schonken 23:47, 25 February 2006 (UTC)

  An RFC on what, formatting pages to conform with existing policy? I don't think that'd get very far. --Cyde Weys 06:03, 26 February 2006 (UTC)

  OK, my apologies to Bobblewik for the AWB assumption. I didn't realise that edits could be done that fast manually, and I'm glad the the AWB coutermeasures are still enabled.--Commander Keane 23:51, 25 February 2006 (UTC)

  Comment. Spamming or not, can users kindly refrain from having revert wars over this message on my talk page? The orange new message box keeps flashing up on my screen and it's really annoying to find there's no actual new messages. - Randwicked Alex B 06:34, 26 February 2006 (UTC)

• Oppose per Wikipedia:Administrators' noticeboard/IncidentArchive73#Bobblewik and other experiences with user. --M@thwiz2020 01:53, 26 February 2006 (UTC)
• Strong support: trivial date links are a scourge on WP that have somehow crept in on the back of the mechanism for formatting full dates. I can't imagine why anyone would object. I guess the only circumstance in which a year date might be useful is for ancient dates/centuries, but even so, that's not a good enough reason to stop the bot. Tony 04:43, 26 February 2006 (UTC)
• Support. Fully automated is ok too - no text is ever lost. BTW, I think it is outrageous that an admin would block Bobblewik for 2 weeks(!) for making good, valid edits (to date links), following the Manual of Style, without revert warring, as was described on the Administrators' noticeboard link that Mathwiz2020 pointed out above. -R. S. Shaw 04:58, 26 February 2006 (UTC)
• Support, please get rid of the "sea of useless blue" as commented above. But the bot should have had a better name. Tempshill 06:01, 26 February 2006 (UTC)
• Oppose. The edits have no consensus, thus a bot is not appropriate. Ambi 06:22, 26 February 2006 (UTC)

  Are you sure there is no consensus? David D. (Talk) 00:52, 27 February 2006 (UTC)

• Oppose, because I'm not so sure that all the changes being made are positive. Specifically, I'm thinking of this diff; although I agree that some of the delinkings were good style, some of them removed links to years in which a number of events pertinent to the article occurred, and I can easily imagine a reader clicking on the links for useful context. This suggests to me that the delinking should be done more slowly and thoughtfully, not quickly and automatically. Robth^Talk 06:31, 26 February 2006 (UTC)

  Although I support this bot, Robth's example is a good one. It might be advisable for Bobblewik to concentrate on more recent dates, post-1800 or post-1900, say. Such date links are much less likely to be useful, and earlier ones may be useful. Of course, any cut-off point is arbitrary, and may lead to strange results if applied strictly, so I would make this a recommendation rather than a condition. Stephen Turner (Talk) 09:15, 26 February 2006 (UTC)

  Ditto per Stephen. good point (and example). --Quiddity 00:13, 11 March 2006 (UTC)

• Strongly Support any and every means to get rid of unnecessary date links and nearly all of them are unnecessary. This is just a method of implementing the current guidelines, as published. We should all be implementing the guidelines until and unless an agreed upon change is made to them. Not noticing changes made in the past or not agreeing with changes made in the past is no reason to stop implementation of the guidelines. Hmains 06:38, 26 February 2006 (UTC)
• Absolute Support. Bobblewik has proven an excellent and dedicated user who will be able to handle the task of getting rid of unnecessary links after identifying them. Antonio Perrito Martin 09:34, 26 February 206 (UTC)
• Support - removing the "sea of blue" is certainly helpful, and, as far as I am aware, in line with the MoS and consensus more generally. Why should every year in every article be linked? It helps no one. Yes, he makes mistakes; but, in my experience, his improving edits massively outnumber any unhelpful ones. Surely it is better that the worthless links are stripped out, and then more targetted ones added back?-- ALoan (Talk) 09:57, 26 February 2006 (UTC)
• Strong support. Almost all such links are pointless and link-cluttering. I have faith in Bobblewik to identify the few exceptions. He is only working for a better Wikipedia. Neonumbers 10:09, 26 February 2006 (UTC)
• Strong support The overlinking is pointless, and that it goes uncorrected perpetuates it, I know I started linking dates for no reason other than I saw it was being done, and assumed it was wiki policy, it was only in seeing bobble's edits that I then bothered to look at the actual MoS to see that I was wrong. Correcting the overlinking will help unclutter pages and make new users aware of what the MoS actually says regarding wikifying dates. Gheorghe Zamfir 10:50, 26 February 2006 (UTC)
• Neutral. Separating bobblewik's "bot edits" from his "real edits" will be a step in the right direction, though I disagree with the principal of many of the types of edits he is interested in performing, only one of which has been mentioned in this discussion. — Feb. 26, '06 [13:06] <freakofnur_xture|talk>
• Support. — Matt Crypto 13:09, 26 February 2006 (UTC)
• Support. older ≠ wiser 13:45, 26 February 2006 (UTC)
• Oppose. While I support the contention that multiple data links in an article are unneccessary, unless very widely separated, date links per se provide a useful mechanism for establishing historical context for events. I resist the 'automatic' de-linking of dates, and contend that this should only be done in the overall context of reviewing an article, and not on a 'fly-by' basis. Noisy | Talk 14:50, 26 February 2006 (UTC)
• Weak opppose. I sometimes use links to years and dates to search for events that happened on a particular date. I also like years being linked, since that visually differentiates them from other numbers. All of this based on my personal tastes, hence weak oppose. Zocky | picture popups 18:41, 26 February 2006 (UTC)
• Oppose. A solution in search of a problem, and Bobblewik's demonstrated carelessness when supposedly doing this sort of thing manually is likely to be magnified if done automatically. --Calton | Talk 20:58, 26 February 2006 (UTC)
• Strong support. KillerChihuahua^?!? 23:51, 26 February 2006 (UTC)
• Support so long as the edits are well-checked, and if a bot-flag is needed (see Rich Farmbrough's vote above). ··gracefool |☺ 00:11, 27 February 2006 (UTC)
• Neutral The idea of the bot to assist in making wikipedia more consistent is worthwhile. However, there doesn't seem to be consensus on what the appropriate level of date-linking should be. I'd like to see that sorted out, first. -- Ch'marr 00:15, 27 February 2006 (UTC)
• Support. I fully support the cutting back of date links (assuming it is done with care, and an understanding of the Article involved). I believe a date should be linked only if it enhances my knowledge of the substance of the Article I am researching. Michael David 00:55, 27 February 2006 (UTC)
• Support –Joke 00:58, 27 February 2006 (UTC)
• Support - links that are not useful to someone reading the article should be removed. Most date links fall under this category. So long as it is monitored by humans to catch the exceptions to this rule then I am firmly behind this attempt to remove unconstructive link clutter. --HappyDog 01:13, 27 February 2006 (UTC)
• Strong Support Excessive blue linking is one of two linking scourges on Wikipedia today. File:VirtualSteve.pngVirtualSteve 03:12, 27 February 2006 (UTC)
• Oppose Most date links are excessive; but, without a clear concensus on what must be monitored (ie, are historically significant years ok to link — and how does Bobblewik propose to sift those from his bot edits), allowing this to go at a fast rate is going to end up in several reverted pages which should be avoided. Neier 11:23, 27 February 2006 (UTC)
• Support a bot to bring articles closer in line with recommended style but I note Rich Farmbrough's discussion above.Thincat 14:01, 27 February 2006 (UTC)
• Strong support. However, the code should be reviewed and tested against all cases to ensure a minimal amount of errors. Gflores ^Talk 14:51, 27 February 2006 (UTC)
• Oppose. Example diff given above by Robth shows why these chamges should not be made by an automated or semi-automated process. android79 15:19, 27 February 2006 (UTC)
• Support, with the proviso that Bobblewick should be careful to look for years that may be of interest to readers of the article, and leave them linked. As others have suggested, ancient dates and the first appearance of a year in an article are more likely to be useful to the reader. Otherwise, year links are a plague and I welcome an effort to remove them.--Srleffler 13:04, 28 February 2006 (UTC)
• Support. Dates are massively overlinked and distract from valuable links. Drastic action is needed. Haukur 16:05, 28 February 2006 (UTC)
• Oppose why remove information that does no harm? If someone took the time to manually link stuff they probably had a reason. -Ravedave 19:05, 28 February 2006 (UTC)
• Weak support - good idea to get all pages into general conformity, but having some links gives context to many articles: I found it very useful for early history, and when The Guardian of 24 October 2005 had a panel of experts review articles, one positive point was "And when I click on click on the year 1922, I get a page telling me what else happened that year. [TS] Eliot is at the centre of a whole web of other references." So please try to keep enough linked years to give context. ...dave souza, talk 19:23, 28 February 2006 (UTC)
• Support, so long as all edits are monitored by a human and not ALL date links are removed (i.e.- typically leave one link per year per page, and ones that are actually significant). EWS23 | (Leave me a message!) 04:34, 1 March 2006 (UTC)
• Support - all in favour of anything to improve consistency of formatting in articles. --Ali@gwc.org.uk 12:44, 1 March 2006 (UTC)
• Oppose - excessive year linking looks ghastly, but frequently the context of an article means a year should stay linked. I feel uncomfortable letting a bot loose on what will be very high volume changes. -- Ian ≡ talk 07:48, 3 March 2006 (UTC)
• Oppose - links should not be repeated but I don't believe anyone would agree that no year should ever be linked to. It appears the aim of this bot is to remove all year links without discrimination, including the first occurence in an article, which I think would be severely to the detriment of many articles. Worldtraveller 13:08, 3 March 2006 (UTC)
• I think it's a terrible idea to delink years automatically, and besides, Bobblewik spammed me. anthony 15:59, 4 March 2006 (UTC)
• Strong support: trivial date links are a scourge on WP that have somehow crept in on the back of the mechanism for formatting full dates. I can't imagine why anyone would object.Jclerman 13:20, 5 March 2006 (UTC)
• Support, plan is to operate in accordance with guidelines; if guidelines change then bot can be adjusted.--A Y Arktos 19:18, 5 March 2006 (UTC)
• Strong Support - He even states "Each edit will be checked visually." invalidating half the oppose votes rationales. --Quiddity 22:47, 5 March 2006 (UTC)

  I didn't want to get into point/counterpoint here, but I don't think it is fair to say that people's voting rationale is "invalid". Bobblewik's made this claim from the word go, yet even a cursory look at his talk page shows numerous examples of where he hasn't bothered to check his edits, leaving his mistakes in place - and this isn't just the case with his date modifications. Astonishingly he's even left mistakes in a section barely a few sentences long. He's made these promises in the past but he's demonstrably not keeping them. Talrias (t | e | c) 23:39, 5 March 2006 (UTC)

  Is permission even needed if you are going to check every edit manually? Can't that be done using your regular user account?

• Comment - when and how should this poll be closed? – Quadell ^{(talk) (bounties)} 14:51, 6 March 2006 (UTC)

I would like to know too. bobblewik 15:34, 8 March 2006 (UTC)

I think one legitimate oppose is all that should be required to deny a bot operating approval. We got that at 19:54, 25 February 2006 (UTC). I suggest we close the comments now.--Commander Keane 15:40, 8 March 2006 (UTC)

• Oppose. While I am not opposed to Mr. Wik himself, I am really tired of all the controversy surrounding the stupid bot antics to begin with. ... aa:talk 02:09, 9 March 2006 (UTC)

• Oppose this user is reckless in his delinking of dates. Jooler 01:53, 10 March 2006 (UTC) - example TWICE on the same article within the space of two weeks! - [2] and [3] Jooler

• Oppose - Although some date delinking done by Bobblebot is beneficial, some edits, such as to Abu Ghraib prison (diff: [4] (removing links to dates significant as part of the US invasion of Iraq) clearly detracts from the completeness of the encyclopedia. I will support this bot if more careful supervision if enforced. - Tangotango 06:50, 10 March 2006 (UTC)

Application withdrawn (see section below). bobblewik 07:03, 10 March 2006 (UTC)

Depending on the how I counted the votes and the close date, I counted:

• Support: 70 to 80%
• Oppose: 20 to 30%
• Neutral: ~5%

As the proposer, I have a vested interest. If somebody independent wants to count, that would be welcome. bobblewik 16:56, 13 March 2006 (UTC)

Wikipedia on the whole operates by consensus, not by voting. However, the bot policy is that if there is a significant objection, then the bot request is denied - because bots are potentially damaging and we want all bases covered. Talrias (t | e | c) 17:00, 13 March 2006 (UTC)

Thanks for the reply. The count has no hidden message. It was not a request for information about policy, or a challenge to my own decision to withdraw. Some people are interested in the count. I agree with your point that consensus and counts are two separate things. bobblewik 18:24, 13 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Approved, Martin 09:06, 11 April 2006 (UTC)

Bot flag added, +sj + 20:07, 25 April 2006 (UTC)

Bot flag for DFBot

DFBot has been maintaining the RFA summary and the AFD summary for a number of weeks now, but I have been too lazy to come back here and ask for a bot flag. Are there any objections to getting one? Dragons flight 17:07, 21 February 2006 (UTC)

I've seen no complaints. Bot flag approved. Rob Church 01:12, 30 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Closed with no consensus. Martin 09:04, 11 April 2006 (UTC)

Bobblebot

I have just been trying out Martin's AutoWikiBrowser. I would like to register User:Bobblebot as a bot account to use it. The task is to reduce linking of solitary months, solitary years, etc in accordance with the manual of style.

If another bot is already doing this task, please let me know. It is a huge slow task (for me anyway) and I would rather do something else. Bobblewik 18:06, 15 December 2005 (UTC)

Um, reduce linking? I don't quite understand. --AllyUnion (talk) 11:32, 16 December 2005 (UTC)

Yes. For example Economy of Algeria has many solitary year links, including 14 to 2004. The policy and popular misunderstandings are explained at:

Wikipedia:Make_only_links_relevant_to_the_context#What_should_not_be_linked

Wikipedia:Manual_of_Style_(dates_and_numbers)#Date_formatting

For example:

• Text with links to solitary years: The short-lived ABC Cable News began in 1995; unable to compete with CNN, it shut down in 1997. Undaunted, in 2004 ABC launched a news channel called ABC News Now.
• Text without links to solitary years: The short-lived ABC Cable News began in 1995; unable to compete with CNN, it shut down in 1997. Undaunted, in 2004 ABC launched a news channel called ABC News Now.

See this diff working towards that: [5]. Bobblewik 13:32, 16 December 2005 (UTC)

A bot can't distinguish between when it should or shouldnt be linked, it needs a human to watch over it. Plus it's pretty damn easy to do with the AWB! Martin 14:14, 16 December 2005 (UTC)

There is rarely, if ever, a valid reason to link to a year article when other date elements are not present, so removing such links (with proper edit summeries) is IMO one of the safest possible editing tasks for an automated or semi-automated process. Detecting the presence of other adjacent date elemetes (and they must be adjacent dor the link to function as a date preference mechanism, normally the only valid reason for the link) is pretty purely mechanical. DES ^(talk) 22:30, 16 December 2005 (UTC)

If the dynamic date formatting feature can recognize correctly-linked dates, there's no reason a bot can't. These links are pointless, look silly and should be removed by a bot. — Omegatron 22:41, 16 December 2005 (UTC)

Oppose. I don't think bots should do the task suggested by Bobblewik nor I'm a particuarly convinced that Bobblewick bothers thinking about in which cases such links are definitely worth it, e.g. at Talk:Luxembourg_(city). -- User:Docu

Support. There is almost never contextual reason to link a solitary year, if these can be detected by a bot they can be removed by one. Hence, the bot makes perfect sense and should be run. Neonumbers 09:46, 17 December 2005 (UTC)

• I support the use of bots to remove useless links for dates. There should also be an effort to get people to stop making links for every date they see. Is there a policy on this or just this style guideline: "simple months, years, decades and centuries should only be linked if there's a strong reason for doing so"? --JWSchmidt 17:07, 17 December 2005 (UTC)
• i support the removal of date links. Does anyone use them? They are a distraction and i would be happy to see them go. David D. (Talk) 17:09, 17 December 2005 (UTC)

• The MoS says generally do not use them, not never use them. Martin 19:13, 17 December 2005 (UTC)

• I definitely support. The more linkcruft we get out of Wikipedia the better. --Cyde Weys _talk^contribs 19:19, 17 December 2005 (UTC)

• Strong support—Trivial chronological links, including such moronic items as 2005, 20th century and 1980s, degrade WP. The issues, in my view, are (1) the dilution of high-value links, (2) the reduction in ease of reading (any reading psychologist will tell you that), and (3) the messy appearance, particularly where linking is otherwise relatively dense. Please eradicate this scourge, far and wide. Most people see the light when you refer them to Wikipedia:Make only links relevant to the context, Wikipedia:Manual of Style (links)#Internal links and Wikipedia:Manual of Style (dates and numbers)#Date formatting. The post-bot relinking of one or two dates for which there might be a strong reason to link is a small price to pay. Tony 23:30, 17 December 2005 (UTC)
  • Good synopsis Tony. i agree that if the bot makes a few mistakes they can be relinked by hand. However, i have rarely seen a date link that is useful. David D. (Talk) 23:33, 17 December 2005 (UTC)
• Strong support as well for the above reasons. 99% of the time, the linked years are unnecessary. If the bot is done correctly, it should be fine. Gflores ^Talk 23:56, 17 December 2005 (UTC)
• I'm not sure this is a good idea, since, as Martin points out, it says to generally not use them. In fact, in the section on dates of birth and death (which immediately follows), it does use bare years when giving the format for when only the year is known. So removing the links in all cases seems like a bad idea. --Mairi 00:14, 18 December 2005 (UTC)
• Support. To Mairi: it should be possible to identify birth and death years from [[Category:xxxx births]] and [[Category:xxxx deaths]] and being in the first sentence.Susvolans ⇔ 17:05, 18 December 2005 (UTC)
• Support Date links have been rendered useless by overlinking. --Wetman 10:06, 19 December 2005 (UTC)
• Very very strong support This is a great idea! Not to mention saving so much of my time as I explain (once again) only necessary appropriate links, paste link to MoS/Date formatting, change tons of leeetle links....did I say this is a great idea? If there is a date which needs a link, fine we put it in, no big. The ratio must be a thousand to one. Did I say how great I think this is? Allow me to say it again: This is a great idea, wonderful bot, thank you! KillerChihuahua^?!? 10:24, 19 December 2005 (UTC)
• I Support the use of bots to remove useless links. Keep up the good work!! Armindo 11:40, 19 December 2005 (UTC)
• Comment. Bobblewik has been adding a link to this conversation to edit summaries. Example.--Commander Keane 11:52, 19 December 2005 (UTC)
• Support. I'm not a big fan of using bots for anything, but this seems to be working very well. Kafziel 12:21, 19 December 2005 (UTC)
• Support - I have to admit I had it in mind that I had to link the year on its own (as it was a separate link anyway from the day month combination). Only when the bot chanced across Royal Society did I find I was in error! Scottkeir 12:24, 19 December 2005 (UTC)
• Support – The improvement from removing these overwhelmingly useless links is worth the inconvenience of reverting the small number of mistakes it will make. – Smyth\^talk 12:28, 19 December 2005 (UTC)
• Question. Will the bot cover infoboxes? I understand that certain infoboxes include incomplete dates. ╫ 25 ^ring-a-ding 12:33, 19 December 2005 (UTC) ╫
• Oppose. This is annoying enough. If Bobblewik cares that much, he can do it manually. Ambi 12:36, 19 December 2005 (UTC)
• Support. Who the hell linked all these dates in the first place? - Randwicked 12:44, 19 December 2005 (UTC)
• Oppose Needs to be done by hand; some links to years should be kept. For early years, what links here function is extremely useful in expanding articles. Warofdreams talk 12:48, 19 December 2005 (UTC)
• Comment. I confess myself guilty of overlinking until recently. I'd vote support provided the solitary birth and death years linked at biographical articles stay like that. I'd like to see someone change solitary year links to more contextually relevant ones (for example, 2005 to 2005 in film in movie/actor articles) rather than simply removing all links; that's what the policy aims at. --Pablo D. Flores (Talk) 12:56, 19 December 2005 (UTC)
• Though I support the discussion of the issue in the Manual of Style, I'm not sure what "harm" comes from linking dates or how it "degrades" Wikipedia. Sometimes it's sort of interesting to be one click away from what else happened in 1755, even if the relevance to the article is not immediately apparent. It's not a big plus, but neither is it a big minus. --Dystopos 16:31, 19 December 2005 (UTC)
• Strong support Having a bot for this kind of cleaning up is a great idea. There are so many over-linked years, especially in lists like discographies or filmographies... --Fritz Saalfeld (Talk) 16:47, 19 December 2005 (UTC)
• Limited Support I agree, lots of date links are really of little value. But it seems to me there is a value in having the ability to see what else happened on that date (or in that year) in history. This is more interesting for the more historical (older) years than more recent ones (say pre-1950) and for events of at least some significance - Marshman 20:28, 19 December 2005 (UTC)
• Oppose. This needs to be done by hand. Removing linked dates makes it difficult to organise articles per "x in music". I fail to see the long-term use for this bot. As an alternate idea, what about setting up a bot to detect articles that have an overage of linked dates, and collecting them for user review, in some sort of cleanup/wikiproject? -Tim Rhymeless (Er...let's shimmy) 04:46, 20 December 2005 (UTC)

(moved from User talk:Bluemoose/AutoWikiBrowser by Bobblewik - please reformat comment): This bot is the bane of most history articles! J. D. Redding 20:46, 19 December 2005 (UTC) [... beginning to understand why most people are uneducated about history ..] (Ps., please don't let this loose on a timelime or similiar article... )

• Is there a reason why it's being done while we are discussing if it should be done and in which cases it might be done? -- User:Docu
  • One reason (as eluded to by J. D. Redding above) to link to years is that is provides landmarks in a history article. For example, if you are reading History of Australia but are interested in events of the 1850s, the blue linked dates are easy to see (Note: it is not my personal belief that date linking should be used, so so get cranky at me about it).--Commander Keane 01:20, 20 December 2005 (UTC)

Aha, link=highlight.

My own theory is that this is a huge unintended consequence of using square brackets to perform two entirely different functions: links, and date preferences. If date preferences could be done a different way, date preferences would not feed the link all dates myth. Bobblewik 02:21, 20 December 2005 (UTC)

• Support. Removing all those spurious year links would be good. It would be good to have a syntax the bot will not remove a link for -- does the piped syntax serve? -R. S. Shaw 05:59, 20 December 2005 (UTC)

• • You said reduce, not eliminate. Some dates must be linked. You should set the bot to eliminate repeated linked years or months across the text, not exterminate them. So, the idea is good, but not so good as I thought in the beginning. Armindo 21:31, 20 December 2005 (UTC)

• Oppose. As a bot, false-positive matches are very likely. For example, Mii-dera links to 1117; and the 1117 page has a factoid about Mii-dera. Links such as the 1117 link from Mii-dera should not be deleted, because it can provide historical context to the event concerning the temple. Neither User:Bobblewik or the author of User:Bluemoose/AutoWikiBrowser seems concerned about this problem, although these types of links can be detected via the Whatlinkshere page. For example, on Special:Whatlinkshere/Mii-dera, 1117 is listed, so it seems clear that there is a reason to keep the 1117 link on the Mii-dera page. Since most year articles mention events of that year, the Mii-dera example is far from the only case where this could happen. Neier 12:41, 23 December 2005 (UTC)

• You clearly havent read the discussion, "the author" (myself) has already objected to a bot being used for this for roughly the reasons you point out. Martin 12:45, 23 December 2005 (UTC)

• Sorry. I was going from the response on your talk page about how much bandwidth it would take to implement the check into the bot. I didn't notice that the comments above also came from you or that Martin is the sig for Bluemoose. So, apologies for thinking you didn't care about the problem. Neier 13:52, 23 December 2005 (UTC)

• No problem ;) Martin 14:03, 23 December 2005 (UTC)

• Support. I support this effort to reduce overlinking. SlimVirgin ^(talk) 22:08, 23 December 2005 (UTC)

• Strong oppose: No link is ever "useless." That is determined by the eye of the beholder. --AllyUnion (talk) 12:27, 25 December 2005 (UTC)

• This is not only not what a bot should be used for, it's not what you should spend your time doing on Wikipedia either. All these edits are incredibly annoying, they often remove useful wikilinks in appropriate areas, and they make the headings more confusing by deleting the spaces in between the equals signs and the words. I strongly oppose any automation of these tasks which more often than not make it worse. Talrias (t | e | c) 02:57, 27 December 2005 (UTC)

It might be useful in this debate to distinguish between comments about the merits of:

• Manual of Style guidance
• a bot proposal to edit articles consistently with the Manual of Style.

For example, Talrias raised a perfectly valid question suggesting that he prefers spaces in headings. I don't care either way and had not thought about it. Now that I check the Manual of Style, it does not use spaces. So the criticism of making edits in accordance with the Manual of Style does not seem fair. If the Manual of Style needs changing, then we can change it easily enough. There are plenty of changes I would like in it. Please can we could separate our judgement about the merits of a bot proposal from our disagreements with guidance in the Manual of Style. Just a thought. Bobblewik 18:15, 27 December 2005 (UTC)

Fine, I oppose the automation of tasks like this because bots get it wrong, making the article worse. I've seen cases where dates have been unwikilinked from in image captions. That's not useful and no human would delink them as "excessive". I find the intended target of this proposed bot highly unnecessary and "fixing" a nonexistent problem. Talrias (t | e | c) 18:47, 27 December 2005 (UTC)

I think you are reinforcing my thoughts that this is an issue for the Manual of Style. Dates in captions are not mentioned in the Manual. If you think they must be linked, then it is easy enough to add that to the Manual. As far as 'no human would delink them' is concerned, I have delinked plenty and I am a human (despite occasional accusations to the contrary). We should be debating these points of style in talk:Manual of Style, not here. I am not targetting you, but a lot of the discussion on this page has been people disagreeing with the content of the current Manual of Style. A bot proposal should be criticised on its merits, not on the basis that the Manual of Style that it follows is inadequate. Bobblewik 19:26, 27 December 2005 (UTC)

No, you are not reading what I'm saying. I'm saying removing excess linkage (per the manual of style) is good, but the way your script/bot has been doing it in the past is bad. Therefore, I oppose the use of a bot flag for these "cleanup" edits. I don't think this is something which should be automated. I don't think I can make it any clearer than this. Talrias (t | e | c) 22:49, 27 December 2005 (UTC)

What is the bot doing that a human being who was following the MoS would not do, apart from being faster? SlimVirgin ^(talk) 03:57, 28 December 2005 (UTC)

Getting it wrong would be a fair summary. Talrias (t | e | c) 04:00, 28 December 2005 (UTC)

Yes, of course, but what would "getting it wrong" amount to in this case? SlimVirgin ^(talk) 07:35, 28 December 2005 (UTC)

I've already explained this above, to you personally, and on a number of other pages. I've provided links to those pages above and in other places. I find your consistent badgering of me annoying. Please stop it. Talrias (t | e | c) 16:04, 28 December 2005 (UTC)

• Oppose. An automated procedure makes these reversions easy to make, but it makes more mistakes with wikilinks than a proper manual review. Policy says that: "...should not be linked ... Months, years, decades or centuries, unless they will clearly help the reader to understand the topic." -- and a quick glance in an editor (AutoWikiBrowser) can not be sufficient at considering that. Also I consider the Manuals recommendation (Manual of Style...Date formatting) not be 100% a law, as also a (recent?) note there states it to be controversial. feydey 02:39, 29 December 2005 (UTC)

• Support. While there may be a very few limited instances where a year link may be helpful for readers to understand the context of an article, the vast majority of such links are just clutter. older≠wiser 21:01, 29 December 2005 (UTC)

A further example of how this bot screws up is this edit to List of Prime Ministers of the United Kingdom, which removes unlinks the dates the particular Prime Minister in question is in office for. Talrias (t | e | c) 21:33, 29 December 2005 (UTC)

No dates were removed in that edit as far as i can see. >ots of dates were unlinked. I would have made the exact same changes manually, had I come upon that article. DES ^(talk) 21:38, 29 December 2005 (UTC)

How did he screw it up? it was much better without the over linking. Martin 21:39, 29 December 2005 (UTC)

Sorry, I didn't proofread. I meant of course unlinked. The links are of course useful so you can then go and find out what else happened in that year in question, in the same manner you would click on the link to the Prime Minister's article page to find out more about the Prime Minister. It is clearly better to have more links in an article than less, because it gives people ease in browsing to related articles. Removing them is an aesthetic difference and this should not be forced on other people - consider changing your own style preferences rather than forcing them on other people. Talrias (t | e | c) 21:43, 29 December 2005 (UTC)

1. Strongly oppose; while there may be some basis to removing date links, it's still being debated. In addition, I don't believe that running it as a bot, even a manual bot, allows the user to know what pages they're editing, and whether the date links on that page might actually be useful. Ral315 (talk) 21:47, 29 December 2005 (UTC)

• Oppose. A bot getting it wrong IMO does more harm than a link of possibly questionable merit. --IByte 22:24, 29 December 2005 (UTC)

• Oppose. I agree that there are too many year links in some places, but I don't think that using a bot to remove them is a good idea. It's something that needs more judgement than any automated system can give -- sannse (talk) 22:46, 29 December 2005 (UTC)
• I think not. Many historical subjects will have years in which actual things happened but not "other date elements". This should never be done unconditionally, especially for the tiny amount of "value" it provides. Demi ^T/_C 22:47, 29 December 2005 (UTC)

• Oppose Per Sannse for the most part. There doesn't seem to be any real reason to go through with this from what i'm hearing. karmafist 22:53, 29 December 2005 (UTC)

• Strong oppose. Style issues should be handled on a discretionary article-by-article basis, as a matter of editor judgement, rather than something cranked out by an automatic bot. More importantly, the Manual of Style is not absolute gospel, and personally I think date links are useful for cross-indexing of information by chronology. If you look at it from the perspective that date links do no harm, and clearly show some merit for finding information about dates specified in the article, I cannot see merit in a bot for the purpose. --NicholasTurnbull | (talk) 00:27, 30 December 2005 (UTC)

• Oppose. There is a contention here that date link clutter is harmful. If so, it is a very marginal harm. On the other hand they do provide an article with context, even if that is a very marginal benefit. On the whole, this seems like a wash to me and so whether or not to keep them should be left to the discretion of the article's contributors. I am opposed to anyone, bot or not, mounting a campaign to systematically remove such links. Dragons flight 00:42, 30 December 2005 (UTC)

• Qualified support. I think that the semiautomatic use of AutoWikiBrowser for reducing the date linking is fine. In the example of 14 instances of links to "2004" in a single article, delinking 13 of them with the assistance of a semiautomatic bot (requiring human intervention) seems to be a good thing. However, instead of eliminating the first occurence of the link to a year, I have often found it more useful to change the link to something more relevelant to the article such as [[2004 in music|2004]], [[2004 in sports|2004]] or [[2004 in books|2004]]. A bot could not do that (at least not without some level of AI) and I would rather leave the link for a human to fix in the future if the person manning the AutoWikiBrowser doesn't take the time to do so. -- DS1953 ^talk 01:07, 30 December 2005 (UTC)

• Oppose. A solution in search of a problem. If it is a problem in an article, just do it; if it's too much trouble, it wasn't much of a problem, was it? --Calton | Talk 03:09, 30 December 2005 (UTC)

• Oppose using a bot to do such. And please don't interpret my vote as either support or oppose of unwikifying/wikifying dates; this vote solely opposes the use of a bot to do such, but does not imply endorsement (or disendorsement) of doing so manually. (In other words, don't read into my vote too much! :-) ) Thanks! Flcelloguy (A note?) 03:15, 30 December 2005 (UTC)
• Oppose, I agree with the idea to reduce overlinking, but I think that this is a task that needs to be done manually in order to ensure that relevant links are not removed (see also my comment at Wikipedia talk:Manual of Style (dates and numbers)#Edits_which_just_unwikify_stuff. JYolkowski // talk 03:21, 30 December 2005 (UTC)
• Oppose — Dan | talk 17:11, 31 December 2005 (UTC)
• Oppose. Removing multiple date links is a good thing, but in general I agree with the opinions expressed above that year links provide a valuable tool for investigating context in articles on historical matters. When making such edits, it is very necessary to see the layout of the article as a whole, and I suspect that a bot-assisted mechanism would not provide the detail. User:Noisy | Talk 20:41, 31 December 2005 (UTC)

Can we close this disucssion now? It seems clear that there is no consensus for User:Bobblewik to run a bot. The votes that continue to pile on aren't of any beneift and are disruptive to this page.--Commander Keane 22:49, 31 December 2005 (UTC)

That might be a good idea: this is a deeply flawed poll. Take Note: "AutoWikiBrowser is not a bot: any edits made using this software are the responsibility of the editor using it." Kim Bruning 16:49, 3 March 2006 (UTC)

Also take note: There is a button (for bots) in the autowikibrowser that will make the bot go automatically. ILovEPlankton 15:16, 4 April 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

Bot flag for SmackBot

Since the above discussion about SmackBot has been concluded, I created a new section below --Francis Schonken 10:47, 26 March 2006 (UTC)

Mulder416sBot

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

I would like to request to run a bot (AWB) to simplify tasks that I come across that are way too repetitive, or would otherwise take foreever. Mainly mass template substitution. -Mulder416 00:52, 28 February 2006 (UTC)

Bot status is granted --Walter 20:20, 24 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

User:Slobot

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Approved. Martin 09:17, 11 April 2006 (UTC)

I'd like to request permission for running the pywikipedia bot Slobot. It will not do much edits here. It will mainly work on nds.wikipedia and only add interwiki links to en. --::Slomox:: >< 13:46, 2 March 2006 (UTC)

Approved for a trial run for one week. Rob Church 01:13, 30 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

FairuseBot

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Approved. Martin 09:19, 11 April 2006 (UTC)

This will be a bot for maintainence tasks relating to the use and abuse of fair-use images. The initial task will be removing any image in Category:Disputed fair use images that's been tagged for a few days from articles it's in. Future tasks may include going through Category:Fair use images and tagging images with no source information as Template:Tl, and other tasks as determined by Wikipedia:WikiProject fair use and Wikipedia:Fair use review.

I'll be using OrphanBot's code as the basis for this bot's code.

The reason for using a separate bot account, rather than having OrphanBot do the editing, is that working with fair-use-tagged images is more contentous than no-source or no-license images, and I want to keep OrphanBot's activities separate from those of FairuseBot. --Carnildo 07:07, 3 March 2006 (UTC)

Approved for the week trial run with one caveat; let me know how it goes. ;-) Rob Church 01:14, 30 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

Tawkerbot2

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Tawkerbot2 is a fork of the anti-vandalism aspects of Tawkerbot as it was suggested that vandal fighting be removed from Tawkerbot (1)'s duties. Tawker 10:26, 3 March 2006 (UTC)

The bot runs various anti-vandalism checks on suspicious pages flagged by a set of algorithms that has been closely QA'd to ensure there are no false positives (the release of the algorithms will be considered upon request). It will run without a bot flag so any edit it can make can be double checked. Tawker 10:50, 5 March 2006 (UTC)

I imagine checking even a small fraction of recent changes could be load intensive. Where does it get the data to check the edit? And I assume it will remain indefinitely user-assisted (apart for sqidward), since a bot will always make some mistakes when looking for vandalism?--Commander Keane 11:52, 5 March 2006 (UTC)

Source data comes from pgkbot the IRC bot used by the CVU and once it grabs an alert from there, it runs an extensive check on the page (simply being flagged by pgkbot will NOT result in an automatic revert, it only touches article space edits and it will not revert twice. In addition, any chanop on #wikipedia-en-vandalism as well as any sysop can kill the bot -- Tawker 12:52, 5 March 2006 (UTC)

This bot appears to be doing a trial run (approved?). Out of a current 7 edits, the 5th was a mistake - the bot reverted a good edit[6]. --Commander Keane 12:40, 5 March 2006 (UTC)

Noted, and fixed -- Tawker 12:52, 5 March 2006 (UTC)

The bot has multiple shutdowns and we're working on a couple more to make it even more safe.

Sample output (what it shows on console) is here - if anyone wants to monitor the bot please send me a message.

We just rewrote the bot and tested it for a while but somehow a codechange was commited when the bot was in live mode (over debug mode where it simply reports what it would do), I've locked out the live codebase from the development version so that better not happen again, I stopped the bot as soon as it happened and I fixed all of it's "damage" -- 02:15, 6 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

Approved. Martin 09:08, 11 April 2006 (UTC)

User:Terminatorius - automated blanking of the vandal anonymous IP talk pages

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

The problem

It is usually good to give for a vandal a warning, adding one of the standard warning templates to the talk page (vandals then frequently stop they activity without blocking). However if the vandal uses the anonymous IP, another user (not vandal) may later see that warning, being unpleasantly surprised. In Lithuanian Wikipedia there was a discussion that the person who left the warning to the anonymous user should delete it after several days, but remembering IP's and removing warnings manually takes too much time. Also, the page with 20 previous warnings and the new warning at the bottom may be less impressive to a vandal. From the troll concept, such page feeds the troll, as seems showing a lot of attention to vandal (multiple warnings, threats, sad words from the innocent users of that IP and so on).

There are talks on Wikipedia that the found vandal warning templates must be automatically substituted by they content, and there are some bots doing this. We think that there is no need to keep old warnings and old anonymous talk pages in general: the page can be just blanked instead, saving even more resources. The real discussions (not warnings) should go on the talk pages of the registered users, or on the article talk pages.

The solution

The proposed bot blanks the anonymous vandal IP talk pages that are older than 240 hours (since the last editing). It does not blank the normal user talk pages and also does not blank pages without the standard vandal warning templates.

This bot

• Obtains the list of IP pages where the vandal warning template is included (via "what links here"). We do not use the subst: in anonymous pages as they do not last longer than two days and vandal seems not deserving needed Vikipedia resources. There are anonymous IP talk pages with such unsubstituted templates in English Wikipedia also (see [7], for instance).
• Checks the page name, it must be the anonymous user talk page (User talk:N.N.N.N).
• If these two matches, checks the page history (visits the history page). The page must be older than 240 hours (10 days since the last editing session).
• If all previous conditions are true, replaces the page content to empty, also explaining the action in the edit summary.

The bot was tested in Lithuanian Wikipedia, first under strictly manual control and later in automated mode. It typically runs daily and just for several minutes. The program is written entirely in Java, without using any additional libraries.

The expected running frequency could be once a day, and (if required) it can be a limit of the deleted pages per session.

The bot could probably be useful for doing the same work in English Wikipedia. The bot operator is Audriusa 20:51, 14 March 2006 (UTC)

• Oppose the possible misunderstandings can be easily explained by to anyone who accidentally gets one of these templates and indeed several templates are solely around to warn people of the possibility. There is also a great need for other users and admins to be able to see warnings left on people's pages so they can determine A) whether an Template:Tl is warranted to welcome the person and/or whether the IP in question has a history of vandalism. Pegasus1138^{Talk | Contribs | Email} ---- 19:39, 15 March 2006 (UTC)

• Oppose While I accept Andriusa's good faith, my main problem with his rationale is that not all IP's are transient. Many IP's, probably most, are consistent vandals or non-vandals. For consistent vandals, warnings indicate their lack of innocence (first-time vandalism is presumed to be a test until warned) and lack of credibility on dubious edits. Conversely, for transient or shared IP's, my problem is that the bot is removing notices that IP's are AOL, for which warnings are pointless, as well as contact information about shared IP's at schools etc, sockpuppets, etc. Wuzzy 00:41, 16 March 2006 (UTC)

• Oppose. Some IPs are chronic vandals, and when I'm dealing with them it's very useful for me to be able to see that. When I'm going to an IP talk page to add a vandalism warning and I see that that IP has 20 previous warnings stretching back over six months, I handle it differently than if the IP has received one or no warnings. I'm sure that many other RC patrollers work in the same way. Hbackman 01:03, 16 March 2006 (UTC)

• Strongly Oppose as per the above. Elf-friend 07:09, 16 March 2006 (UTC)

• Oppose. Per above. Some IPs are static and therefore warnings should not be blanked. Anyway if no new warnings are added other users of that IP will not get 'new messages' so there isn't too much risk of them being coming across them. Editors can manually remove old warnings when adding new messages to an IP talk page if appropriate (what is appropriate depends a lot on the IP and therefore should be left to an editor not a bot to decide). In addition warning templates should be subst. Petros471 09:50, 16 March 2006 (UTC)

Consensus has been reached

I am convinced by the negative arguments and do not plan to continue with this idea. The discussion on this bot can be closed. The idea of time-dependent editings (not necessarily blankings) probably can be reused by someone suggesting something completely different in Wikipedia:Bot requests. Audriusa 08:17, 16 March 2006 (UTC)

Just for completeness, adding a reference to User:Tawkerbot which has a similar blanking function and ran into similar problems. Femto 15:06, 16 March 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.

Ganeshbot

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

I've created Ganeshbot to create city/town articles in India based on on data provided by Census India 2001. I converted the data into a comma-seperated file. Bot will read the file line by line and create article stubs. Please see examples, Aadityana and Aambaliyasan, that I had created in the sandbox using the bot. It will be manually run by me. There are 5161 towns listed. It should take a couple of hours to complete. This is similar to the User:Rambot that created U.S city/town articles. Thanks, Ganeshk (talk) 08:33, 26 March 2006 (UTC)

A few questions:

• Do you have more information you can put into the articles besides the name and rough location? Part of the strength of the Rambot articles is the amount of information they contain: basic geography, a basic demographic profile, and frequently a map showing where the place is.
• How fast is the bot editing? To keep from overloading the servers, bots shouldn't edit more than once every ten seconds. Creating 5161 articles should take at least 15 hours.

--Carnildo 09:08, 26 March 2006 (UTC)

Reply:

• The excel file had the following columns,

City Name,Urban Status,State Code, State Name, District Code and District Name

Let me research if there is anyway possible to expand further.

• I have added 30 seconds delay between each edit. Couple of hours was just a guess. Out of 5161, many exist already. I have not run it on the entire file yet. So I don't have a time estimate.

- Ganeshk (talk) 09:22, 26 March 2006 (UTC)

Per suggestions by User:Carnildo, I had made changes to the bot. Could you please look at Aadityana, Aambaliyasan and Kodumudi and give approval for the bot?

Changes:

• Added population count
• Added Geo-coordinates and Altitude

I would like to run it for all towns.

Thanks, Ganeshk (talk) 08:20, 1 April 2006 (UTC)

I think it looks good, I would give it permission, but I think it might need greater community approval first, as it is such a large project. Maybe you could mention it at the village pump? Martin 12:56, 1 April 2006 (UTC)

We should only be happy to have something like this, and eventually for all the countries in the world. Hearily support, except perhaps the demographics could be moved to another paragraph, like Rambot did. But if this is all the info you have then this is probably fine. --Golbez 20:05, 1 April 2006 (UTC)

Martin, Thanks for the tip. I put up a message on the village pump.

Golbez, I have added two new sections, Geography and Demographics. Please check Kodumudi. Thanks, Ganeshk (talk) 22:00, 1 April 2006 (UTC)

How do you feel about exploiting Geographic references like Rambot, instead of adding the same reference to all the articles? Melchoir 21:52, 1 April 2006 (UTC)

Melchoir, Could you please explain with an example on how Rambot used Geographic references? Thanks, Ganeshk (talk) 22:00, 1 April 2006 (UTC)

Done. Added geographic reference similar to Rambot. I used Template:Tlx format since the number will not be constant. - Ganeshk (talk) 22:40, 1 April 2006 (UTC)

Please don't go ahead Bot articles are horrible, especially the awful rambot ones about American localities. A couple of human-written sentences are always better and they will all get done in the end. I see huge long term problems with the rambot articles. They are going to get very out of date, but when new census data is ready how are they going to be updated, especially those where people have added proper content? Is is going to be wiped along with the old rambot bilge, or is the rambot data going to be left in place forever. Please be patient and wait for Wikipedia to be written by people not machines. Hawkestone 23:03, 1 April 2006 (UTC)

I have to disagree; having the stubs help greatly, I find the information generally useful, and I prefer to have machine-made articles than none at all. Many of the Rambot articles have since been improved, sometimes vastly so, by editors. Treat them as stubs. --Golbez 23:53, 2 April 2006 (UTC)

The examples provided look pretty good, as for becoming out of date, this is applicable to all articles, regardless of how they were created. Creating the articles like this is good, as it provides a base for humans to build on and create much better articles, as is the case with many of the rambot articles. Martin 23:17, 1 April 2006 (UTC)

It should round to the nearest percent, rather than truncate (to avoid the gender proportions totaling to 99%). It should give the literacy rates for males and females, rather than the proportion of the literate people. TimBentley (talk) 19:21, 2 April 2006 (UTC)

I fixed the rounding issue. I rephrased the the literacy line. If you feel it still does not sound right, Could you please write the exact line how it should show using Kodumudi as an example. Thanks, Ganeshk (talk) 23:32, 2 April 2006 (UTC)

Why is the bot not getting approved now? Any more issues left? deeptrivia (talk) 12:31, 28 April 2006 (UTC)

June 2006

I have added a locator map feature. For towns that have longitude and latitude values, a locator map will be automatically loaded and dot placed for the city. With this last change, I feel the bot is complete. Please check Kodumudi. Can the approval group please give it a bot flag? - Ganeshk (talk) 07:07, 24 June 2006 (UTC)

Looks good to me. --Carnildo 01:10, 25 June 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.