Wikipedia talk:Bots/Archive 2

From Wikipedia, the free encyclopedia

Archive 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.

Contents

Formatting bot

Looking through the days of the year pages, I have noticed that quite a few contain such ugly things as hyphens instead of dashes, and the use of the "conventional quoting style". I have been going through fixing these, but there are simply too many to do manually. Would anybody mind if I wrote a bot to do this? It should be a straight-forward text replacement (eg. s/ - / – /). —Wereon 15:16, 24 May 2004 (UTC)

Strongly object. Have you seen Wikipedia talk:Manual of Style/Dashes. Many people do not want unreadable markup like — in the wikitext. Angela. 15:58, May 24, 2004 (UTC)

Japanese author bot

I am planning to put hunderes of stubs for Japanes authors using bots. Any objection? The format should be like Hozumi Shigeto -- Taku 04:47 Feb 22, 2003 (UTC)

Nice format. I like the way you have the Japanese characters in the same parenthesis as the birth/death dates. One minor issue: a space needs to go after the comma in the dates. --mav
I don't like the way the Japanese characters are in the same parenthesis as the dates. They are different things altogether, so it seems strange to link them in that way. And it doesn't follow the standard "(date - date)" pattern that seems to have been agreed for biographical articles. I prefer the format used at, e.g. Kokichi Mikimoto. -- Oliver P. 16:04 Feb 22, 2003 (UTC)
The article contains little more than List of Japanese authors, only the Japanese characters. These could be added in the list. It seems better that a separate article about an author is made at the moment that there is more info on him/her. Except when it is probable that soon more info will be added to a large percentage of the stubs. - Patrick 10:33 Feb 22, 2003 (UTC)
I agree that it makes more sense to do a list, and when someone decides to flesh out a particular author they can create a new page. --snoyes 16:34 Feb 22, 2003 (UTC)

What is the benefit of adding all these authors using a bot? Is it truly contributing to the content of wikipedia, or just cluttering the database? If people feel that they need to add content about a particular topic, then surely to benefit readers, it should have at least a little content. Kabads

I don't think they're just "clutter", since these people are presumably all going to get articles, so pages on them are going to be in the database eventually anyway. These bot contributions are just providing slots into which people can put stuff. We all want more encyclopaedic content to be added, of course, so I think the most important question is: are people more likely to add encyclopaedic content on an author who (a) has no entry at all, or (b) has a very brief entry with almost no content? I suspect that the latter is more likely to spur people on to add information, but it's debatable, of course. (I still don't like the format, though!) -- Oliver P. 16:51 Feb 22, 2003 (UTC)

First of all, stub articles have inadequate information but the purpose of them is to provide a good starting point to write an article. Finding out the birth date and death date is a tedious job and particularly putting Japanese characters are difficult to job for those who can't type Japanese. Secondly, we already have the List of Japanese authors. It means soon or later we will have articles for all of them. There is no reason to postpone making articles unless we are unsure we need such a article. Yeah, we need a little more at least, where he/she was born and died. I think I can put such information because I have it.

Yeah, maybe bots are a little too overwhelming. I will put stubs by hand if we can agree with having stub articles for Japanese authors.

Taku, Have you read Wikipedia:Bots? It's a useful page on this issue. -- sannse 16:41 Feb 22, 2003 (UTC)

Yes. I believe my bot is useful, harmless and not server hog.

I guess another benefit of using a bot is to guarantee consistent format for the articles. --Tomos 20:30 Feb 22, 2003 (UTC)

Japanese city bot

I would like to use a bot to make a stub of Japanese city. The format should be like Funabashi. I will wait to use it until enough dicussion is done. I welcome any sort of comment, really any. -- Taku 04:18 Mar 17, 2003 (UTC)

Hmm, that's awfully short. Could you include a little more data? Population would be a start, and such stats are likely widely available. --Brion

Yes, it is less than adequate stub. I put population and I will put area. Unfortunately I couldn't find out the complete set of stat data like census in USA. So we have to get each data like population individually. What else do you think we need? -- Taku 22:27 Mar 17, 2003 (UTC)

Yea that's too short. I think you could add information like location (latitude, longitude...), climate (max/average/min temperature, rainfall...), geographical features (nearby rivers/mountains...), history (when is the city formed, who formed it, important historical events...), just to name a few. Maybe you could use the US cities articles as reference since they are also added by a bot (AFAIK). --Lorenzarius 15:30 Mar 18, 2003 (UTC)

Of course, I wish I could. If someone has good stats in English, I will appreciate. But I doubt there is such. Maybe we need to add history or such by hand later. Anyway we can't expect artciels like US cities articles. -- Taku 16:14 Mar 18, 2003 (UTC)

IMHO, Why can't we expect articles of similar quality? We don't necessarily need history, but if we're going to run a bot, then we should be able to find all the information possible. At the very least, Lorenzarius' first few suggestions are very good; population alone is not enough, nor the city designation. Atorpen

Taku: for population, how about this? [1] Tomos 20:05 Mar 18, 2003 (UTC)
for more demographics, [2] (files are in .xls format, and you would perhaps need japanese font)
and land areas [3] Tomos 21:21 Mar 18, 2003 (UTC)

First of all, thank you for all of those who gave me comments. Particularly pages Toms gave seems quite useful making an city article bot will add decent probably. I am working on a converting Japanese into English and writing the script of bot. Feel free to keep discussing formatting or whatever. -- Taku 02:02 Mar 21, 2003 (UTC)

I updated Funaba article. I think it is still short but adequate stub. For some reason, I can't access [4]. If someone can download the excel file and upload it somewhere or send it to me, I will appreciate. -- Taku 04:30 Mar 22, 2003 (UTC)

Moved from Wikipedia:Village pump

I would like to create a bot that gets info from various U.S. Department of State websites, and makes articles. I have no experience making this type of program, could someone point me in the direction as to what I need to learn, or a where to start? MB 20:26 27 May 2003 (UTC)

If you have experience with Perl, it's easy. Even if you don't, it's easy as long as you're somewhat good with computer languages. For example, on page 708 of the first edition of the Perl Cookbook (Christiansen & Torkington) it shows how to grab a web page with two lines of code:
use LWP::Simple;
$content = get('http://www.wikipedia.org');
... at which point the variable '$content' contains the HTML of the requested page. More sophisticated robot work requires a little more code, but it's surprisingly easy. Jordan Langelier
If you have no experience writing bots please be careful that you don't unleash some kind of terminator onto the servers. Always test your work on a sandbox system (install an HTTP server locally with static copies of some example pages), and don't leave it going for hours and hours without checking what it's up to. CGS 22:40 27 May 2003 (UTC).

Automated content is generally disliked here: the value of Wikipedia comes from the fact that human beings interested in each subject have written and edited the articles. If you really feel that you must auto-create, you can test the bot on my server first; mail me and I'll give you all the info. LDC

Wapcaplet has asked me to help upload approximately 5000 images for U.S. locations within states. Here is an example:

Image:Map_of_Alaska_highlighting_Aleutians_East_Borough.png

I have an auto-uploader program debugged and ready to go. I have registered an account User:The Anomebot for this purpose. Could someone please approve this, and add the Anomebot to the list of registered bots? -- The Anome 11:30 12 Jul 2003 (UTC)

Have I understood correctly that it was Wapcaplet who modifed them? If so, it might be nice to replace the word "modified" with "modified by Wapcaplet". Credit where credit's due. :) -- Oliver P. 11:34 12 Jul 2003 (UTC)
I've been modifying most of them, but James also created several of them, set the standard for formatting, and kindly provided web hosting space to house the completed images. I figured a generalized attribution was best. This is a community effort, after all :) However, if you like, since James and I are the only ones (so far) who have been working on them, "modified by Wapcaplet and James F." would be fine too. Oh, and of course I second the bot request! -- Wapcaplet 13:43 12 Jul 2003 (UTC)

Perhaps the attribution should read

"Public domain map courtesy of The General Libraries, The University of Texas at Austin, modified to show counties. Released under GFDL. See Wikipedia:U.S. county maps."

Then we can put the credits on the Wikipedia:U.S. county map page, where there is plenty of space to credit contributors. This would also neatly link them back to a single page. In fact, I'll do that now. -- The Anome 13:54 12 Jul 2003 (UTC)

I believe my IP address has been banned because I swamped the server with a script. (It was retrieving and not updating, but being automated I suppose it qualifies as a bot anyways. I thought it was slow enough, but apparently not.) Strangely, though, my IP address does not show up in Special:Ipblocklist. How do I go about verifying that is what happened, and perform appropriate apologies and grovelling to get re-instated? Thanks for any advice. -- Amillar 19:16 15 Jul 2003 (UTC)

Auto Loading (Batch) to Wiki

I am looking for a program to load a mass of articles (batch load) to Wiki.
I try to add new articles to the newly born Hebrew Wikipedia.
Can U pl. give me some directions? -- Dod1 12:17 18 Jul 2003 (UTC)

Bot and Images

Is it the User:The_Anomebot that is slowing down the server? -- User:Docu

I doubt it - it's only uploading one image every four minutes. The server is just slow sometimes. --Camembert
Why can't we hide its edits? -- Tarquin 17:36 19 Jul 2003 (UTC)
Apparently it's not registered as a bot yet. I suppose it needs a developer to fiddle with something (sounds like the sort of thing Brion would usually do, but he's away, I think). --Camembert
From User talk:The Anome it seems that the bot has indeed been given bot status, but image uploads aren't hidden, even when done by bots. Evercat 17:52 19 Jul 2003 (UTC)
Ah yes, I was going on the bot's user page. Hopefully somebody can fix that. --Camembert
Read somewhere else, can't remember where, or who from, that it is set as a bot, but the code for hiding bot edits doesn't hide picture uploads yet. כסיף Cyp 18:05 19 Jul 2003 (UTC) (Damn edit conflict... This reply would have sounded more intelligent if the couple of posts above weren't there...)

For most of the day (19 Jul 2003), it ran much faster: one image per minute [5] -- User_Docu

Even one edit a minute is no faster than a human could go (though admittedly, you'd have to be really committed to keep it up for that long). But I admit, I don't know for sure whether it's slowing the server down. It's not doing anything really remarkable if it is, though (User:Rambot went much much faster at its peak, though it was just uploading articles, not images). --Camembert 17:46 19 Jul 2003 (UTC)

Image uploads are probably more complex, possibly the length of Wikipedia:Upload_log slows down things as well. Maybe an administrator can archive it? -- User:Docu

Wikipedia:Upload_log is very slow to load, and currently just a blank page:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META content="text/html; charset=windows-1252" http-equiv=Content-Type></HEAD> <BODY></BODY></HTML>
Might still be possible move it to an archive, by creating a suitable page-renaming url, but I'm not sure that's a good idea... כסיף Cyp 18:46 19 Jul 2003 (UTC)

The problem persisted after User:The_Anomebot stopped, thus it's likely that it was something else. In the meantime, the bot had just been more efficient getting to the site. -- User:Docu

I couldn't figure out where else to put this question. Is there a client for Wikipedia that allows you to edit articles without a web browser? Specifically, I would want to be able to make edits to an article directly as with a wordprocesser. I think it would be relatively simple to implement. You would use it by clicking exactly where you would like to make a change, and then just type. Ezra Wax 23:16, 11 Aug 2003 (UTC)

There has been some discussion about having some sort of interface, but so far the bots just use the same methods that are used as if you had done the edit manually. I think sometime it will eventually get implemented but it isn't a high priority. Maybe once wikimedia starts taking donations we can purchase the feature. -- Ram-Man 03:01, Aug 12, 2003 (UTC)
Or, of course, the people who want the features could write them. C'mon! If you're already writing your own bot, surely you're capable of doing the work to make your own job easier in the future? --Brion 04:37, 12 Aug 2003 (UTC)
You would think that, except that my bot is a hacked up job which has some limitations (such as a much too small maximum article size). It also has other issues, but for now it does the job. It would be great to implement a better solution, but for now it does what it needs to do fairly quickly and so I leave it at that. I move slow enough as is with a real life to live! But maybe in the future I *will* do something more. -- Ram-Man 11:04, Aug 12, 2003 (UTC)

The client would require a custom editor. Possibly scintilla could be used. There are also some links there to custom editors. Possibly mozilla's Html editor could be customized to do the job. Ezra Wax 13:59, 12 Aug 2003 (UTC)

Bot for generating Rambot-like articles on Israeli cities

I am planning to use a bot to create Rambot-like articles on Israeli cities. The discussion thread on wikiEN-l starts here. The example article is at User:AdamRaizen/Ramla. Comment there or at User_talk:AdamRaizen or User_talk:IsraBot. -- AdamRaizen 14:08, 2003 Sep 15 (UTC)

I like some of the different types of data in the above article. The U.S. Census bureau data gets old to look at after a while. Good work! -- Ram-Man 22:29, Sep 19, 2003 (UTC)

From the Village pump

This may be a stupid question but; when Google and sites like Onelook update themselves regarding to wikipedia content, does it have an significant effect on the server performance and response time? -- Skysmith 11:52, 19 Sep 2003 (UTC)
Well-behaved, polite, gentle spiders like Google's don't seriously degrade performance, but we get our share of visits from jerks who hit every link on the site four at a time with no delay between hits. When these are found, they get banned. --Brion 05:16, 20 Sep 2003 (UTC)

Gazetteer time?

Moved from Wikipedia:Village pump on Thursday, September 25th, 02003.

With the Israbot adding to the output of Rambot and others a lot of 'statty' settlement articles are being created.

IMO these articles are non-encyclopedic and are cluttering the main namespace (I'd also say they are valueless additions, but there you go ;). I think it is time for a gazatteer.wikipedia.org, or similar, to hold these articles. TwoOneTwo 21:16, 21 Sep 2003 (UTC)

I like these articles. Without the one for my hometown I wouldn't be here. Ark30inf 21:28, 21 Sep 2003 (UTC)
Hey, if you have enough extra time on your hands, why not take one or two of them and turn them into more than gazetteer entries. Improve them, don't bitch about them... -- Cimon Avaro on a pogo-stick 21:44, Sep 21, 2003 (UTC)
I was thinking of taking some databases on French communes, but I think that sending in 36000 communes, most of which are unimportant little villages, is excessive. Indeed, a gazeteer would perhaps be better adapted. We should also agree on some easy-to-reprocess format. David.Monniaux 00:24, 22 Sep 2003 (UTC)
There are about equal # of Rambot-US place-articles too (30,000), so I say: "add the French ones away!" At least then we'd then have some US-world balance. Unless, of course, that we remove puny US entries, then we should not add French entries as well. But it looks Rambots are here to stay, so bring the French in! --Menchi 00:36, 22 Sep 2003 (UTC)
We might as well use a bot to add an article for each known star in the universe (coordinates, colour, estimated blow-up time, SETI scan results). That would boost our article count. Of course just as a starter, feel free to add all you know about any of those zillion stars :) Or add articles about every chemical compound known to mankind, or about every McDonalds restaurant in the world. But what is the point? I'd say let us be careful with converting the Wikipedia into a raw data container. Most of these data are covered well in some other place, where experts/afficionado's know to find them. Maybe use a bot only to add such data to exisiting handwritten articles. Erik Zachte 10:32, 22 Sep 2003 (UTC)

If you've got some comments on Rambot (to stay or to delete), drop a note @ User talk:Rambot/Delete. --Menchi 00:36, 22 Sep 2003 (UTC)

I'm not arguing about Rambot. In effect I'm arguing from Erik Zachte's pov. The articles created are formatted demographic and economic data - nice if you want to do a comparison or create colourful maps but in no way encyclopedia articles. They have value but they do not belong in the main encyclopedia namespace - as evinced by the fact that so few of the machine-created articles have been edited.

They are distorting the value of Wikipedia content, especially article counts and size counts into giving an unwarranted impression of the depth of Wikipedia. I'm not saying delete I'm saying move, like the sep11 material they are specialised and not generally usable. With a separate namespace for the raw articles, as and when they are improved they can be moved back. TwoOneTwo 14:29, 22 Sep 2003 (UTC)

I think that it would not be good to start limiting the types of content that appears on wikipedia. Rather, we should look at ways of adding meta-information to certain kinds of pages. It is possible that this can be done in a non-labor intensive manner. Then searches can be done with some of the meta-data attached. For instance, I like the random page feature. Yet, I do not kike getting back so many town/city entries. But I would rather be given a "random search with no town/city entries" than remove the town and city entries from wikipedia, which are useful for other reasons. Also, putting domain name qualifiers in front of wikipedi.org (eg gazetter.wikipedia.org) only gives us a single dimnsion of meta-data. The human mind is able to process much more complex types of meta-data and we should not deliberately dumb down wikipedia by imposing a rigid meta-data structure. We should put in a flexible meta-data structure and use that. RayKiddy 18:43, 22 Sep 2003 (UTC)

An encyclopedia is not an almanac! I think we should take a stand on this. An encyclopedia article is human-written and gives context, weighs relevance, etc. Bot-entries are not articles, they're data. If I want raw data on every county in the US for example, I go to an almanac; if I want context, filtered by a knowledgable human, I go to an encyclopedia. In an encyclopedia, I can rely on the fact that someone has taken the time to include the important facts and leave out the cruft. That's why I went there, and not to the almanac. I'd like to see Wikipedia is not an almanac added to the basic tenets of Wikipedia. Axlrosen 22:37, 24 Sep 2003 (UTC)

Disambiguation bot

Rob Hooft has written a bot to simplify Disambiguations. Given the name of a disambiguation page, it gets the pages that link to that page, shows the disambiguators, and lets the user choose one of them or 'none', then changing that page.

It has been used with success on nl:, and I would like to ask whether it would be okay to use it here as well? Andre Engels 09:01, 28 Sep 2003 (UTC)

How does it change the page? I mean, if a link is to [[English]], does it change the text to [[English language]] or put it as a hidden link: [[English language|English]]? I would much prefer the second - that's what I tend to do when manually fixing links to disambiguation pages. The first option can lead to some very strange sentences. -- sannse 09:07, 28 Sep 2003 (UTC)
It does the second. Andre Engels 09:19, 28 Sep 2003 (UTC)
As an example, I have used the robot on 11 (out of a larger number) from the pages linking to 'Java'. See http://en2.wikipedia.org/w/wiki.phtml?title=Special:Contributions&target=80.129.93.237 for the results. Andre Engels 09:51, 28 Sep 2003 (UTC)
This looks as though it could be really useful. If the technical types give the go-ahead I would be very interested in learning how to use this (I might take a bit of teaching - I don't know the first thing about bots). I do a lot of manual disambiguation fixing, so something that makes the work easier would be welcome. -- sannse 10:43, 28 Sep 2003 (UTC)

Navigation for Swiss Federal Council members

According to the discussion at Talk:List_of_members_of_the_Swiss_Federal_Council I wrote a bot to update the navigation of members of the Swiss Federal Council. The bot has been tested on my local copy of the data set and works fine. For every change I manually enter the start and end year and accept the changes. Any objections to this? -- Patrice Neff 14:18, 29 Jan 2004 (UTC)

No. Give it a try, preferably run it slowly and check your talk page once in a while. -- User:Docu
The wikitech-l mailing list would be a better place to ask for permission as a developer needs to enable your bot edits to be marked as such. Angela. 20:36, Jan 31, 2004 (UTC)
I doubt someone will do this before it actually ran for a couple of entries. As there are just approx. 150+ entries, it's likely that it's done before a developer gets to it. -- User:Docu
My impression was that you were supposed to ask for this before you start running it. Angela. 22:02, Feb 2, 2004 (UTC)
Probably Patrice should be asking someone instead of writing text here. -- User:Docu
I'm not currently running the bot yet, so I don't have to ask anyone at this moment. I have subscribed to the list Angela cited and will ask for permission there before running the bot. Additionally it might be good to include this information on Wikipedia:Bots. I thought that asking on this talk page is equal to asking for permission, which obviously it is not. Patrice Neff 05:03, 3 Feb 2004 (UTC)
Some things have been added to Wikipedia:Bots now, based on a discussion on the policy on IRC. Angela. 05:34, Feb 3, 2004 (UTC)
I suppose Angela now agrees that Patrice can go ahead.
BTW I find the times given a bit fast. Last time Rambot uploaded maps for US counties (6 per minute), he/she/it was the main user able to edit on the site. As image uploads are more work intensive, the software changed, and the site works faster, they rates may be ok. -- User:Docu

bot building

is there a way to build a bot?

I'd like a bot that could track (somehow) party standings from official websites that keep track of these, and automatically update pages with these party standings. (political pages)

is there a way to make such a bot? if so, respond in my talk page

Pellaken 06:32, 5 Feb 2004 (UTC)

Heraldic bot

I'm attempting to include the coats of arms of a very large number of towns, cities, boroughs, districts and counties in England and Wales on the towns' Wikipedia pages. (See User:Marnanel/CivicHeraldry_pages for more information, background, and what has been done so far.) I've done all the entries for one county by hand: there were about 17 images to upload and describe, and around the same number of pages to include the images on. It was a pretty mechanistic process and took a couple of hours. Since this is only about 5% of the full corpus, I'd like to request permission to use a bot to do the uploading and renaming of the pictures automatically, and possibly then the inclusion of those pictures on the relevant pages semi-automatically (though this would perhaps be easier done by hand anyway). I'm a pretty competent programmer, and it shouldn't be too hard for me to do while keeping within the rules. Marnanel 03:58, 4 Mar 2004 (UTC)

That sounds a useful reason to have one. I believe Eloquence has a script for uploading images so you might want to check with him before writing something new. Angela. 01:39, Mar 5, 2004 (UTC)
Thank you! I've asked him and got hold of the script. Do you need to know the username to mark it as a bot or something like that? I thought of running it as HeraldicBot. Marnanel 19:31, 5 Mar 2004 (UTC)
Is it OK to start this going now? Marnanel 03:45, 8 Mar 2004 (UTC)
I guess so. :) Angela. 19:50, Mar 8, 2004 (UTC)

Unregistered taxo-box spacing bot

195.50.12.118 and User talk:195.50.12.115 are unregistered bots which have been attempting to add spacing to all taxoboxes accross the Wiki. They have been blocked, but I fear they shall be back soon from another IP. Fennec 20:56, 6 Apr 2004 (UTC)

Interwiki bot

I want to use a bot in order to add a bulgarian interwiki on the year (1800-2009) and calendar pages. It will wait at least 60 seconds between edits. --Borislav 04:15, 8 Apr 2004 (UTC)

Please let your bot run with at least 30 seconds wait intervals between each edit. When we have a set of contributions from it from which we can see that it doesn't have bugs and doesn't cause harm, we can give it the "bot" flag and you can remove the wait interval. — Timwi 19:57, 13 Apr 2004 (UTC)
The bot is working on the calendar articles with about 40 sec intervals. After them it'll edit also the year pages. --Borislav 11:56, 17 Apr 2004 (UTC)
The bot has now made over 500 edits and received no complaints, so I have marked it as a bot. It will no longer appear in recent changes. Angela. 21:23, Apr 18, 2004 (UTC)

Request for a bot?

It is, perhaps, pointless to do so, but I would like to request a bot. The amount I know about computer programming could be written in large letters on the back of a stamp, so I can't write it myself. The pointers to WikiProjects on the talk pages of articles (see, for example, Talk:Urban Hymns, Talk:Chimpanzee, Talk:Mount Airy, Maryland) could be created using ‹The template Msg has been proposed for deletion here.›  {{Template messages}} tags by adding it to the talk pages of everything linked to from the list of albums, list of primates, list of cities in Maryland, etc, updated perhaps every month. There seems to be much support for these messages, and using ‹The template Msg has been proposed for deletion here.›  {{Template messages}} means they could be easily changed as needed, also providing much more visibility for the WikiProject as well as for the very idea that the article itself can be changed by any casual user. It also strikes me as being very likely simple to program -- anybody willing to do it? Anybody object to it being done? Tuf-Kat 04:56, Apr 8, 2004 (UTC)

Apparently, Mt. Airy, Maryland doesn't have a talk notice for some reason, and the Wiki is being too slow for me to go find one that does, and there is no list of primates, but this bot would provide an incentive to keep such lists updated... More important for list of albums than list of primates, since primate speciation occurs significantly more slowly than the release of new albums. Tuf-Kat 05:05, Apr 8, 2004 (UTC)
Maybe the pywikipediabot could do this. It might be best to ask on their mailing list. Angela. 17:07, Apr 8, 2004 (UTC)

Request to use human-assisted spelling bot

I made a script that downloads pages from my list of articles with common misspellings, fixes any of the common misspellings, and then lists other potentially misspelled words, which I can correct with the click of a button. I review every change before it is submitted, but the script allows me to correct pages faster and more accurately as compared to changing articles by hand.

I'm not sure if I need permission for this, but I thought I'd ask to be safe. My script should be:

  1. harmless because I review and discard any incorrect fixes
  2. useful because I can easily correct not only misspellings on the list, but any misspellings I notice
  3. not a server hog because it is limited by how fast I can approve changes. It should use less overall bandwidth than fixing every article by hand because I only request the edit page for the article, rather than loading the article and then following the "edit" link.

I could even allow others to help because the script runs off my personal computer through a web interface. Thanks. Wmahan. 21:55, 2004 Apr 8 (UTC)


There have been objections to spelling bots in the past. See Wikipedia talk:Bots/Archive 1 and the Spelling Day? thread on Wikien-l. It might be best for you to ask on the mailing list (WikiEN-l if you are just doing this on the English Wikipedia, or Wikipedia-l if you want to run it in other places as well). Angela. 22:07, Apr 8, 2004 (UTC)

Thanks for the reply. I don't think that the primary objection to spelling bots, namely that they perform erroneous changes, applies here because every change is reviewed by a human before being submitted. So this method shouldn't make any changes that I wouldn't make by hand; it only speeds the process up. I've emailed WikiEN-l as you suggested for confirmation. Wmahan. 03:34, 2004 Apr 9 (UTC)

I think Angela's point is that for example, sometimes valid foreign words happen to be misspelled English words. Human knows the article is talking about completely non-English stuff but bots cannot be aware of contexts. I am pretty sure it happens extremely unlikely but it might happen. And it is really tedious and hard to spot errors made by bots because people assume they don't make mistake. So I object to spelling bots. -- Taku 06:17, Apr 9, 2004 (UTC)

I'm claiming that my bot will be different from others that might run amok in that I will be aware of context when I review each change. I agree that it is easy to make a mistake in approving a bot's changes; that's why I must explicitly request any correction that doesn't appear on the list of common misspellings, which has been reviewed for potential pitfalls. If this bot is approved, I will be conservative and not make changes that I'm unsure about.

I admit that there is a possibility, though small, that the bot will make an incorrect change, and then I approve it. My points here are that there are that 1) because I am responsible for the changes, anyone who notices the mistake is free to yell at me on my talk page; and 2) one can interpret the maxim be bold in updating pages as an argument that even if I make a few mistakes in making many corrections, I'm justified by the overall benefit to Wikipedia.

To summarize, I share your concern about assuming that bots are perfect, but I won't make such an assumption. Thus I think that your blanket rejection of spelling bots is a little too cautious. Wmahan. 15:14, 2004 Apr 9 (UTC)

Addendum: I thought I should let you know that Jimbo agrees with the idea, see WikiEN-l. But I'm not trying to bypass anyone's opinion, and I'm willing to listen to any further objections or comments. Wmahan. 18:25, 2004 Apr 9 (UTC)

Please do. As far as being bold goes, how well you do will be the best way of judging whether it's good or bad. If you get it wrong regularly, people will object.:) Jamesday 14:17, 13 Apr 2004 (UTC)

Sounds good to me. Let's give a shot. After all, the whole point of wikipedia is any mistake would be corrected soon or later. -- Taku 15:07, Apr 13, 2004 (UTC)

Update: I have made about 1500 corrections using the above method, receiving one complaint and several thanks. The complaint was regarding three articles where I changed barbeque to barbecue. I had checked dictionary.com prior to that change, but alas, it failed to include the alternate spelling. Thanks for everyone's comments and support! Wmahan. 04:24, 2004 Apr 25 (UTC)

Authorising bots

I haven't been paying much attention to the approval of bots, but to me this statement looks wrong: "You must get permission from an administrator before using a bot." What is the relevance of administrator status here? As I understand it, administrators have extra technical capabilities, but no extra decision-making authority. Here's a suggested policy that doesn't rely on setting up a cabal: "Before running a bot, you must get approval on Wikipedia talk:Bots. State there precisely what the bot will do. If a rough consensus emerges that it is a good idea, wait a week to see if there are any further objections, and if there aren't, go ahead and run it." What do you think? -- Oliver P. 18:06, 14 Apr 2004 (UTC)

I see no reason it shouldn't be changed to that. Angela. 01:56, Apr 15, 2004 (UTC)
IIRC "administrator" originally meant a developer, but I think Oliver's idea for a guideline change is fine. --mav 06:25, 15 Apr 2004 (UTC)
I would say, discuss here, and also with a developer. Sometimes developers may know things about the Wikipedia software that we don't, in terms of loading and suchlike. Martin 21:47, 15 Apr 2004 (UTC)
Okay. Thanks for the feedback. I'll change the guidelines to say "rough consensus" and "approved by a developer". -- Oliver P. 00:21, 16 Apr 2004 (UTC)
The reason it says administrator not developer is because I have absolutely no interest in approving bots, and said so on IRC. If someone runs a bot without approval, it falls to sysops to block it without hesitation. If someone wants the bot flag set, then a steward can handle that. There's nothing about approving bots which requires technical knowledge. -- Tim Starling 08:21, Apr 16, 2004 (UTC)

ChuqBot

I (User:Chuq) have registered a user User:ChuqBot to use to edit links to avoid redirects - specifically, change all links to Hobart, Tasmania or Hobart, Australia to Hobart. I plan to user pywikipediabot to do this. Any tips or advice for a first time wikipedia bot/script user? --ChuqBot 14:11, 24 Apr 2004 (UTC)

I don't want to sound negative, but it seems that if most other articles refer to the city as Hobart, Australia, then that should be the name used as the title. I would guess that the typical reader who doesn't live in Hobart wouldn't already be familiar with a city of ~200K people. The old name also seems more consistent with other city names. So why go to the trouble of renaming the page and changing all the links? Wmahan. 18:54, 2004 Apr 24 (UTC)
Most other articles refer to Hobart, Australia because that was the name of the article up until it was changed yesterday. See Talk:Hobart for reasoning - the other three Hobart's are tiny. --Chuq 23:49, 24 Apr 2004 (UTC)
I see your reasoning, but I was questioning the wisdom of renaming the article in the first place. I wanted to suggest that it might be creating unnecessary work to rename the article when the title seems well established, and then go to the point of using a bot to change all the links. It does make sense to have Hobart redirect to Hobart, Australia because it is by far the largest. I just think that the fact that you need a bot to rename the page is a warning that it may be overkill, especially since the benefit is debatable. But I'm not strongly opposed to your idea, and am willing to defer to the judgment of the others if there is a consensus in favor of your bot. Wmahan. 02:38, 2004 Apr 25 (UTC)
The disambiguation policy suggests Hobart should be used for the most linked to Hobart, and should not be a disambiguation page when the other Hobarts are much less well known. I think this is a good use for the bot. Angela. 03:05, Apr 25, 2004 (UTC)
I think we all agree that Hobart should lead the user to a page about the city in Australia, rather than a disambiguation page. To me the the question was when an article's title (without redirects) should be just City rather than City, Country. I was probably biased by the City, State notation used for cities in the US. After perusing the article names for other cities in Australia (Sydney, Melbourne, Wollongong, etc.) I see how Chuq's idea is consistent, and therefore withdraw my reservations. Wmahan. 03:51, 2004 Apr 25 (UTC) P.S. Just found Wikipedia talk:Naming conventions (city names), which discusses the issue. Sorry if I'm behind.
Even if in this case, it's a good idea to move the page about Hobart, Australia to Hobart and update the links (I wonder why it wasn't placed there in the beginning), in general, it's not necessarily a good idea to up eliminate all redirects, as the redirect may be more precise than just the direct link to the (current) page title. -- User:Docu
I don't intend to delete the re-directs - just correct links so they go direct to the article --Chuq 10:46, 25 Apr 2004 (UTC)
A better sample may be Munster as there are too many possible/probable choices, it's generally preferable to link to the article through a redirect. -- User:Docu
Retrieved from "http://en.wikipedia.org../../../b/o/t/Wikipedia_talk%7EBots_Archive_2_692b.html"