User talk:JarektBot/2012

Interwiki to dissimilar page types

Is it proper to using interwiki links instead of {{On Wikipedia}} to link a Commons category to Wikipedia articles (i.e. not categories)? The bot is doing this at a very rapid rate. I know that people have gotten in the habit of doing all the time on Commons categories, but I'm not sure it's correct — that's supposedly what {{On Wikipedia}} is for. Shouldn't the straight interwiki links be for gallery-to-article interwiki, or for category interwikis to Wikipedia categories instead? --Closeapple (talk) 21:41, 10 November 2011 (UTC)

That is because Wikipedia and Commons are dissimilar projects. Information of Wikipedia is organized in Categories filled with articles filled with files. On Commons we have categories (sometimes called meta categories, as in Template:MetaCat) filled with more categories which are filled with files. We do have some galleries but MUCH fewer than categories and often out of date. For example in Category:People by name we have 121k categories and only 4k galleries. As a result of this different organization, often best match for wikipedia Categories are Commons meta-categories and best match for Wikipedia articles are Commons categories. Wikipedia uses en:template:Commons for linking from articles to galleries and categories on Commons. 260k such links are to categories and ~40k to galleries. So linking between Commons categories and Wikipedia articles is in most cases the only option we have. Greetings --Jarekt (talk) 16:40, 11 November 2011 (UTC)
I was going to start a similar conversation when I noticed this. My view is that as long as a certain commons category (e.g. Category:Dacia) has a matching category in en WP (e.g. en:Category:Dacia) or other language WPs (e.g. ro:Categorie:Dacia), the bots and users should use those categories for interwiki links and use {{On Wikipedia}} for article interwikis (this is what the {{On Wikipedia}} documentation suggests as well). If there is no such matching category on any language, then the interwiki links to articles should be used. As far as galleries go, I think they should always have interwiki links connecting to articles on different WPs. The categories from en and other WPs should also link back to commons using commons cat template. This is what I've done for Category:Dacia. Since there are at least 15 languages which have this category (e.g. en:Category:Dacia, fr:Catégorie:Dacie, etc.), it is highly beneficial for all other wikipedias to know about the commons category and the available media. But I faced the problem of manually maintaining a lot of links under {{On Wikipedia}}. My question is, can the bots maintain both, using the suggested rules? --Codrin.B (talk) 16:47, 2 February 2012 (UTC)
When I click on interwiki link in Category:Painters from Italy I expect to land in Wikipedia category like en:Category:Italian painters. But when I click on interwiki links in Category:Marcello Bacciarelli I expect to land in an article en:Marcello Bacciarelli. A bot running mw:Manual:Pywikipediabot/interwiki.py can handle both types and will use either one or the other based on what was found there to start with. --Jarekt (talk) 02:47, 3 February 2012 (UTC)
That seems pretty much what I suggested. The only note, if there is a en:Category:Marcello Bacciarelli (say a category which groups articles like Marcelo Bacciarelli, Works by Marcelo Bacciarelli etc.) I would point to it instead. In lieu of such category, I would point to the article Marcelo Bacciarelli, as you suggested. But what about updating {{On Wikipedia}}? I think it would be great to standardize and implement a set of rules, which a bot could apply when adding\updating interwikis.

babel vandalism

Hallo JarektBot, vandalism: on my user page. Well, not really - but I do not like this change. reason: Looks not good (space on right border (→misalignment with other box), not i18ned). In addition the box background wasn't transparent anymore so that the gray didn't shine through at User_talk:Saibo. Cheers --Saibo (Δ) 01:59, 11 November 2011 (UTC)

Sorry You did not like the change. I usually stay away from people user pages, but switch to new babel style would allow us to retire or simplify maintenance of a LOT of templates. It would also allow us to boxes for much more languages. The topic was discussed here, here, here, and you can find documentation here. I guess the extension a little different look than the template, although when I switched and when I tested it I did not notice them. The broken internationalization of the header and footer is more problematic and will need to be fixed. I will look into it. Greetings --Jarekt (talk) 17:07, 11 November 2011 (UTC)
No problem - I know it was intended as help. :) However, if you compare both versions of my user page you will notice that in "your" version there is a misalignment between the upper box and the babel box. And: although I like the transparent look (on my talk page) a bit more I would adjust to the white block. Or maybe not and simply continue to use the old templates. ;) The background is a advantage of them. Maybe it could be resembled with the "your" babels. However, please do not do such changes using your flagged bot. I didn't notice this through my watchlist (bots hidden) but by wondering about the strange new look and English text on my talk page. And if you do bot changes they should no change the appearance, right? :-) Cheers --Saibo (Δ) 19:37, 11 November 2011 (UTC)
I was under impression that appearance would not change. --Jarekt (talk) 20:36, 11 November 2011 (UTC)
May be only with the bad HTML code I have on my user page. ;-) Have a nice weekend! --Saibo (Δ) 21:22, 11 November 2011 (UTC)
I reverted this because it broke my babel. If you can fix it you may change it again. --MGA73 (talk) 18:24, 12 November 2011 (UTC)

I also reverted the change to my personal page[1] due to the strange indenting it caused. I'd suggest the bot not be allowed to change user pages. --JD554 (talk) 19:55, 12 November 2011 (UTC)

+1. --Saibo (Δ) 20:02, 12 November 2011 (UTC)

+1

table.mw-babel-wrapper {
  margin:0em;
  margin-bottom:0.5em;
  width:250px;
}
 
.mw-babel-notabox {
  margin-left: auto; 
  margin-right: auto; 
  width: 238px;
}

-+ !important → MediaWiki:Common.css ? -- RE rillke questions? 22:53, 12 November 2011 (UTC)

  Comment bugzilla:27793 about displaying babel template text in user language might be fixed in the future. --Jarekt (talk) 18:59, 20 November 2011 (UTC)

If I understand bugzilla:27793, we just have to ask at the Village Pump if everyone agrees to have Babel internationalized and then file the new bug ?--Zolo (talk) 14:26, 25 January 2012 (UTC)
That was done, see bugzilla:32726 --Jarekt (talk) 15:09, 25 January 2012 (UTC)
Oh ok sorry--Zolo (talk) 15:18, 25 January 2012 (UTC)

exclude my user page

Is there a template I can place in my user space to prevent your bot from editing it? -Nard (Hablemonos)(Let's talk) 18:32, 1 January 2012 (UTC)

My bot does not usually edit user pages. I checked yours and the only edit I could find was switch from old-style {{Babel}} template to preferred {{#babel:...}} format, which was a one time run. I am unaware of any templates that AWB or pywikipediabot based bots would use to prevent future edits. However I am planning to stay away from user pages as much as possible. --JarektBot (talk) 13:09, 5 January 2012 (UTC)

Stop

You have edit a user page.--Simon-kempf (talk) 19:01, 18 January 2012 (UTC)

i am no longer editing those. See User_talk:JarektBot#babel_vandalism. Is this edit causing any problems? --Jarekt (talk) 19:03, 18 January 2012 (UTC)

Creator:Jacques_Androuet_Du_Cerceau_(I)

 
Creator:Jacques_Androuet_Du_Cerceau_(I) has been listed at Commons:Deletion requests so that the community can discuss whether it should be kept or not. We would appreciate it if you could go to voice your opinion about this at its entry.

If you created this creator, please note that the fact that it has been proposed for deletion does not necessarily mean that we do not value your kind contribution. It simply means that one person believes that there is some specific problem with it.

Please remember to respond to and – if appropriate – contradict the arguments supporting deletion. Arguments which focus on the nominator will not affect the result of the nomination. Thank you!

Robert.Allen (talk) 10:48, 23 January 2012 (UTC)

"bitmap graphics" and "audio file" became bulgarian and afrikaans

Hallo Jarekt, http://commons.wikimedia.org/w/index.php?title=User:TrygveFlathen&action=history would be nice if you could revert your bot's wrong changes (if no one else has reverted in the meantime for those babel languages at least. Note: babel extension still is not i18ned. ;-) Not really Commons suitable. Cheers --Saibo (Δ) 01:06, 25 January 2012 (UTC)

I will look into it and get back to you. --Jarekt (talk) 14:19, 25 January 2012 (UTC)
I look for other template conflicts and found some more. See output of {{Babel|BG-3|AF-1|ANI-3|VI|win|bot|CID|align=left}} {{#babel:BG-3|AF-1|ANI-3|VI|win|bot|CID}}. --Jarekt (talk) 15:31, 25 January 2012 (UTC)
Babel user information
bg-3
af-1
Template:User ani-3
vi
 This user contributes using Microsoft Windows.
± This user is in fact a bot.
 This user account is secured with a unique committed Identity.
Users by language
Babel user information
bg-3 Този потребител има задълбочени познания по български език.
af-1 Hierdie gebruiker het basiese kennis van Afrikaans.
ani-3 This user has advanced knowledge of Andi.
vi-N Thành viên này xem tiếng Việt là ngôn ngữ mẹ đẻ.
win-N This user has a native understanding of Ho-Chunk.
bot-N This user has a native understanding of Bongo.
 This user account is secured with a unique committed Identity.
Users by language
Ouch, thanks Jarekt. Hmmm... --18:15, 25 January 2012 (UTC)
I will try few things: (1) revert by hand my edits on pages of users using one of the clashing options (2) I requested change to Babel extension to make it case sensitive. If I do not get reply there I will try bugzilla. (3) I know EN wiki has the same issues (see en:User:Jarekt/b). I will alert them about it. --Jarekt (talk) 18:53, 25 January 2012 (UTC)
Okay, thank you! :-) I really wonder that the babel extension still is not i18ned - that is not really suitable for Commons, don't you agree? Cheers --Saibo (Δ) 19:02, 25 January 2012 (UTC)
I think bugzilla:32726 should take care of this. You can add your voice to the bugzilla discussion if you would like it to be processed faster. BTW, I reverted by hand my edits on pages of users using one of the clashing options--Jarekt (talk) 15:47, 27 January 2012 (UTC)
Thank you! --Saibo (Δ) 21:09, 28 January 2012 (UTC)

Swaping title and description for images in Category:Trajan's Column - Cichorius Plates

I noticed that User:JarektBot replaced {{Reliefs of Trajans Column}} with {{Artwork}} for files in Category:Trajan's Column - Cichorius Plates. This could be a good thing since {{Artwork}} is a nice template I've been using myself for archaeological objects and artwork, although I don't know how the replaced template was since is gone now. However, the "title" and "description" are now swapped. If you look for example at File:021 Conrad Cichorius, Die Reliefs der Traianssäule, Tafel XXI.jpg, most of the detail is in the title and not in description. Can the bot swap them? There is a conversation also at User talk:Gun Powder Ma#Commons WikiProject Dacia. Thanks.--Codrin.B (talk) 16:05, 2 February 2012 (UTC)

Adding also multilingual descriptions

I started this discussion at the User talk:MerlIwBot#Adding also multilingual descriptions but I just realized that JarektBot also adds interwiki links (side question: why multiple bots doing the same thing?). So I think the suggestion for this feature applies here as well. Thanks.--Codrin.B (talk) 16:30, 2 February 2012 (UTC)

I agree it would be valuable, but to my knowledge there is no code for this task written at the moment, and I do not have python skills to write one at the moment. Also it seems to me that that would work only for categories with interwikis (~27% of all categories). --Jarekt (talk) 17:01, 2 February 2012 (UTC)

27% is a great number though (where did u find it anyway?) But it should also work at least with categories that have the same name with an English article, which should bring the coverage up considerably. At least this is how I think Sum-it-up works.--Codrin.B (talk) 19:21, 2 February 2012 (UTC)

The number 27% come from the fact that we have 1,862k categories and my recent month long sweep maintaining interwikis in all categories found 1,354k categories without interwikis. I have the list. Be carefull about assuming the is category has the same name as an article in EN wiki than they are related. It is not always the case. For example we have a lot of people's names where category and article refer to a different person. --Jarekt (talk) 19:35, 2 February 2012 (UTC)

Good point, but I would assume that the number of false positives (which could be corrected by hand later), is still far bellow the number of real hits (meaning those hits where the category name really matches the EN article correctly). It would still save a lot of work I think. Or maybe if we want to get fancy some review process can be set in place to review the bot planned updates. But that would give more work to admins. I don't know how many mismatches we have but it should be a small number I think, which should mean that we have to pick the lesser evil. And those mismatches should only prompt users to take some action and bring some consistency between Commons and EN wiki naming.--Codrin.B (talk) 19:52, 2 February 2012 (UTC)

I think the most conservative solution would be to only process categories with interwikis. In my experience people have very low tolerance for bot that makes any mistakes. And it is no fun to be correcting edits by hand. That said I was thinking about writing some bots for boosting number of categories with interwikis. But that is a long term project. --Jarekt (talk) 20:15, 2 February 2012 (UTC)
Well, I added the first category IWs in at least 30000 categories; I see that more and more people start to see the interest and are following. Before adding text descriptions by bot, we have to define a collapse rule. Categories with an intro in 270 languages is not very practical. A more elegant and maintenance-free solution would be the display of the initial paragraphs(s) when hovering over the interwiki links on the left. It would be great if the search engine would include the category IWs in its search, which it doesn't for the moment. --Foroa (talk) 07:03, 3 February 2012 (UTC)

duplicate authority control?

http://commons.wikimedia.org/w/index.php?title=Category:J%C3%BCrgen_Hardt&action=history Is there a way to detect if that template is already present before adding a new one? ;-) --Saibo (Δ) 22:11, 7 February 2012 (UTC)

There is and it is rather simple. But since I was starting with no pages using it, it seems it would be safe not to check.--Jarekt (talk) 02:45, 8 February 2012 (UTC)
Okay, thanks. Also note that there could be {{Normdaten}} already. I am sure you will do what is useful - just wanted to let you know. --Saibo (Δ) 03:24, 8 February 2012 (UTC)

"own" cleanup etc.

Maybe remove |permission = see below at the same time ?-Zolo (talk) 10:36, 6 March 2012 (UTC)

I do not see much of those, but I added a rule to delete those. --JarektBot (talk) 12:19, 9 March 2012 (UTC)

Automatic Creator: creations

Hi! The bot is 'creating' Creator: templates over existing ones. Sometimes it's for the good (see for instance Creator:Francesco Laurana); at other times it deletes data (see for instance Creator:Exekias). Are you sure the bot should overwrite? Jastrow (Λέγετε) 08:31, 9 March 2012 (UTC)

No it was not meant to overwrite anything. Those creators were not there last time I checked (I think). I will check by hand this batch of 85 uploads and modify my upload code to skip if page is present in potential future uploads. Thanks for letting me know. --JarektBot (talk) 12:14, 9 March 2012 (UTC)
I created these templates after you asked for help on Commons talk:Walters Art Museum :) Thanks for the explanation. Also thanks for the bot; it's very tedious to create these templates by hand. Jastrow (Λέγετε) 13:41, 9 March 2012 (UTC)
Bot is only uploading what I assembled in a spreadsheet. The data collection is still mostly manual process. --Jarekt (talk) 13:46, 9 March 2012 (UTC)

bot problem

Just noticed a problem with the bot. Check out this diff. For some reason the bot converted a question mark in the provenance into a whole new parameter. Kaldari (talk) 04:10, 25 March 2012 (UTC)

I am look into this. No need to stop the bot by a block - this part is done as AWB and any message stops the bot. --Jarekt (talk) 04:14, 25 March 2012 (UTC)
Ah, that's good to know. I'll unblock it. Kaldari (talk) 04:17, 25 March 2012 (UTC)
I found it. It is this regex "(date\s*=\s*)([^\(]*)\(([^\)]*)\)(\s*\n)" -> "$1$2$4\n|period = $3" designed to pull occasional period parameters tucked on the end of the some dates and put them as a separate parameter "period" so they can be piped through {{Period}} template. It can be fixed by adding "\n" into second and 3rd group "(date\s*=\s*)([^\n\(]*)\(([^\n\)]*)\)(\s*\n)" so it wont match any multi-line parameters. But for now I will disable it and look at other regex to see if there is a potential for multiline matches. I will write something to identify all the files where such incorrect substitution happen. I think I just have to find cases where "period" parameter is not after "date". --Jarekt (talk) 04:33, 25 March 2012 (UTC)
  Fixed This should be all fixed now.--Jarekt (talk) 19:46, 27 March 2012 (UTC)

Walters

Just one minor thing I forgot to mention. In the dimensions field, Framed {{size| could be replaced with {{with frame}}{{int:colon}} {{size| --Zolo (talk) 10:54, 27 March 2012 (UTC)

I will add it. --Jarekt (talk) 10:56, 27 March 2012 (UTC)
Thanks but it seems that it did not work [2]. --Zolo (talk) 13:26, 28 March 2012 (UTC)
I added the replacement rule to the rule set, but I guess this one did not meet the rule. I will check why. --Jarekt (talk) 13:35, 28 March 2012 (UTC)

Creator:Adam-Wolfgang Töpffer

 
Creator:Adam-Wolfgang Töpffer has been listed at Commons:Deletion requests so that the community can discuss whether it should be kept or not. We would appreciate it if you could go to voice your opinion about this at its entry.

If you created this creator, please note that the fact that it has been proposed for deletion does not necessarily mean that we do not value your kind contribution. It simply means that one person believes that there is some specific problem with it.

Please remember to respond to and – if appropriate – contradict the arguments supporting deletion. Arguments which focus on the nominator will not affect the result of the nomination. Thank you!

Leyo 23:07, 11 April 2012 (UTC)

Category:William_McGregor_Paxton

 

William McGregor Paxton has been listed at Commons:Categories for discussion so that the community can discuss ways in which it should be changed. We would appreciate it if you could go to voice your opinion about this at its entry.

If you created this category, please note that the fact that it has been proposed for discussion does not necessarily mean that we do not value your kind contribution. It simply means that one person believes that there is some specific problem with it. If the category is up for deletion because it has been superseded, consider the notion that although the category may be deleted, your hard work (which we all greatly appreciate) lives on in the new category.

In all cases, please do not take the category discussion personally. It is never intended as such. Thank you!


Handcuffed (talk) 03:27, 21 May 2012 (UTC)

type:landmark

Hi. What’s the point of removing it? --AVRS (talk) 09:36, 5 August 2012 (UTC)

+1. -- Smial (talk) 11:12, 5 August 2012 (UTC)
+2. Removing it is contrary to the documentation en:WP:GEO#type:T which says type:landmark should be used for buildings etc!! "buildings (including ...), caves, cemeteries, cultural landmarks, geologic faults, headlands, intersections, mines, ranches, roads, structures (including ...), tourist attractions, valleys, and other points of interest" Rwendland (talk) 12:06, 5 August 2012 (UTC)
+3. Please stop this immediately. --Haselburg-müller (talk) 12:38, 5 August 2012 (UTC)
+4. Whats the point? Why do you think buildings like this are no landmarks? --TMg 15:04, 5 August 2012 (UTC)
+5. At the risk of piling on: I don't get why type:landmark is being removed either, when it is valid at best and ignored at worse. (And it is still somewhat harder to determine what to re-add it to than it is to remove it.) Is there a bot task page for this? --Closeapple (talk) 18:00, 5 August 2012 (UTC)
This run is a part of fixing problematic {{Location}} templates requested by the people using the information provided by attribute parameter. See Commons:Bots/Work_requests#Remove_type:landmark_from_Template:Location for more details. --Jarekt (talk) 21:27, 5 August 2012 (UTC)
Es bleibt völlig unverständlich und ich finde keine weiterführende Diskussion, in der die Notwendigkeit dieser massiven Änderung verdeutlicht würde. -- Smial (talk) 10:52, 6 August 2012 (UTC)
I don't get it. You are deleting valid and important information from hundreds, if not thousands (?) of image description pages! Please stop this! Fix the location template if it is broken but do not simply remove valid informations for a reason nobody cares about! --TMg 22:09, 6 August 2012 (UTC)
TMg, the only purpose for attribute parameter is to pass information to the GeoHack server about what to do with the coordinates. That information is not used by {{Location}} template in any way so there is noting to fix with the template. People who run GeoHack, which is used on all Wikipedias and Commons asked for help cleaning up small percentage of files which use incorrect attribution values, not recognized by the GeoHack. They rely on community of each project to perform the cleanup since they are not directly involved in all those projects. "Type:landmark" was a setting added by default by the template some years back; however, the documentation specifically says that it is not supposed to be used otherwise. --JarektBot (talk) 02:40, 7 August 2012 (UTC)
I am not sure about use with “camera”, but WIWOSM uses the type in Wikipedia to choose the icon. --AVRS (talk) 07:41, 7 August 2012 (UTC)
On Commons we do not parse the attributes and none of them is used for anything, other than being passes to GeoHack. --Jarekt (talk) 11:28, 7 August 2012 (UTC)
With that argument you could delete the attributes from the {{Object location}} templates as well. If the issue for the removal is that it overrides type:camera, it’s better to fix the template or GeoHack. If the issue is that the coordinates are of a camera and not of a landmark, maybe the type is still useful, as a page can have the camera location but no object location specified. --AVRS (talk) 08:13, 8 August 2012 (UTC)
@JarektBot: I know this is not a template parameter but a geohack attribute. Does not make a difference. You are deleting valid information. What made it invalid? And why? Why is this not a landmark any more? It's not true that "landmark" is the default now. "Landmark" is not in the Location template. There is a confusing "camera" type in the template, whatever that means. As said neither the type of ma images nor the type of my motifs is "camera". I do pictures of "landmarks". Remove the "camera" garbage if there is a conflict with an existing attribute. Also since you removed thousands (?) of types nobody knows any more what the type of these images is. Is it still a landmark or something else? A river? Does this change the icon of all these images to a question mark? Introduce a new attribute like class:camera or whatever and add it to the Location template, I don't care, but restore everything you deleted. Thanks. --TMg 02:48, 8 August 2012 (UTC)

Wrong Creator

Is there any way to keep the bot from identifying a Christopher Wren who lived in Pennsylvania around 1900 as w:Christopher Wren of London in the 1600s? See diff. Thanks, Ruhrfisch (talk) 11:19, 26 September 2012 (UTC)

Disambiguation is always an issue. I added "active ..." to the files author field. That should fix the problem. --Jarekt (talk) 12:03, 26 September 2012 (UTC)
Ruhrfisch, I just noticed that my bot changed it a few times. Sorry about it. I was adding creators to files in Category:Author matching Creator template, Creator template not used and needed to restart the process a few times. --Jarekt (talk) 12:10, 26 September 2012 (UTC)
No worries - thanks for the work you and your bot do here! Ruhrfisch (talk) 15:33, 26 September 2012 (UTC)

Josef Engelhart Foto bei Erhiehungssystem

Das ist nicht der Josef Engelhart auf Wikipedia, sondern ein anderer - nicht ändern. This is not the Josef Engelhart on Wikipedia, but another - do not change File:T Erziehung Winter 01.jpg Josef Engelhart aus Veitshöchheim bei Würzburg. --Bauer Karl (talk) 14:31, 26 September 2012 (UTC)

I added clarification to this and several other files by the same photographer. --JarektBot (talk) 14:44, 26 September 2012 (UTC)

Category:Non-empty category redirects

Hello Jarekt,
First of all: thanks a lot for your excellent bot.
I wanted to warn you that I did some change that could bother your bot:

With Foroa, we are trying to test (see here) the bot actions on Category:Non-empty category redirects .
Do you have any idea what problem my 4 modifications could cause ?
Best regards Liné1 (talk) 09:51, 15 October 2012 (UTC)

--Jarekt (talk) 13:22, 25 October 2012 (UTC)

en:Thomas Birch (historian) vs. en:Thomas Birch (painter) & creator template

Hi Jarekt,

your bot has made an little mistake by mixing the painter and the historian named Thomas Birch: Look at this automatic fixing of the creator template. The Book A complete collection of the historical, political and miscellaneous works of Milton, edited by Thomas Birch, the historian, was published in 1738 (!). The painter Thomas Birch was born in 1779 (!). I think this could be a possibilty to fix the bot. Greetings from Germany --Laibwächter (talk) 07:24, 26 October 2012 (UTC)

Rewriting WP iwls to avoid redirects - good or bad?

I'm concerned about changes like this and [3]

This replaced the previous interwiki link on the Commons category of Category:Girder forks from [[en:Girder fork]] to [[en:Motorcycle fork#Girder]]

At [[en:Girder fork]] we find:

#REDIRECT [[Motorcycle fork#Girder]]

{{R to section}}
{{R with possibilities}}

[[Category:Motorcycle suspension technology]]
[[Category:Motorcycle technology]]

This is a valid redirect. In particular it's a redirect tagged and categorized as "to section" and "with possibilities". en:Girder fork is a notable topic that warrants an article. As yet, no-one has had time to write it (motorcycle suspension is oddly poorly covered at en:WP). So in the meantime, we point the redirect at a paragraph within a much broader article at en:Motorcycle fork#Girder. We would hope though, possibly soon, that en:Girder fork, will become the full article that it clearly ought to.

Redirects are useful. They are not a bad thing. MediaWiki has good support for redirects. However the Wiki community often has some strange aversion to them, seeing them as a target to be hunted down and removed. There is some trivial benefit here about efficiency (yet that is never considered when embedding WP:SCARE BLOCK CAPS links into arguments for removing them "to avoid server load"). There is certainly a benefit to removing double redirects.

The virtue of redirects though is one that's long familiar to programmers, even if not to wiki editors. This virtue is that of abstraction. A redirect to [[:en:Girder fork]] goes to :en:WP and allows WP to select the best current content available for that topic according to WP's understanding of the content at the time. A link to [[en:Motorcycle fork#Girder]] though will always go to that same small section, even if a full article later appears under its own name. (Note too that such a change can't be 'bot reversed in the future). As any programmer knows, an external API should be decoupled from internal implementation details.

I would like to see JarketBot's behaviour here changed so that it doesn't replace redirects like this, i.e. to not replace those redirects representing a useful abstraction, not merely a technical fix-up (spellings etc.). This could be done by identifying redirects that are categorized as either (not and) those with possibilities, or those to sections (by either categorization or the use of a fragment identifier in the target). Andy Dingley (talk) 11:54, 23 January 2012 (UTC)

User:JarektBot is running the standard mw:Manual:Pywikipediabot/interwiki.py code which is the same code used on all the other wikipedias and other wikimedia projects. The code has many tweakable parameters including "-noredirect means that if a redirect page is found, the redirect is not followed, as is the normal behaviour, but the page is skipped." I could turn that option on; however I assume that this will skip all the redirect even more frequent cases where the article was moved and the redirect points to the updated title. --Jarekt (talk) 13:14, 23 January 2012 (UTC)
Following redirects isn't a problem, but no 'bot should modify a page unless we know that such changes are of benefit and also (following the principle of the Hippocratic Oath) that "they should do no harm" - i.e. in a case when some redirects can reasonably be removed, but others would best be left, then the 'bot should modify none of them.
I would assume that a 'bot can be constructed that can distinguish these two cases though. Surely any 'bot framework can return the initial link address, the final target address and hopefully also the annotated content of the redirect itself? That's enough to identify and exclude redirects to sections (robustly, in all cases), and redirects with possibilities (where these have been manually tagged).
Even if determination is difficult or impossible, I would argue that replacing redirects should only be done in cases where it's a demonstrable improvements - i.e. if it can't be proved to be useful, then it shouldn't be changed. Redirects just aren't a problem that need urgent fixing. Andy Dingley (talk) 13:32, 23 January 2012 (UTC)
All this might be great suggestions, but I did not write that bot and do not know enough about it to be able to modify this enormous code. See mw:Manual:Pywikipediabot/Development#How_to_report_a_bug to report your suggestions. What I can do is modify one of the tweakable parameters of that software and what I can gather from the brief description "noredirect" option might allow me to skip commons pages which points to redirected wikipedia pages. This probably will have a side-effect that the bot will only add new links but will not correct no longer valid links. But that might be preferable option. --Jarekt (talk) 14:59, 23 January 2012 (UTC)
I saw that behaviour of "redirect re-solving" at other iw bots, too - yes. I think it is wrong and I frequently revert such bot changes. IWs to redirs are there for a purpose (article may be written in future at the place of the redir). Please skip such iw pages if it is not possible to turn it of. Cheers --Saibo (Δ) 15:15, 23 January 2012 (UTC)
I will work in the evening on testing the bot with the new setting, to make sure that it does what I think it does. --Jarekt (talk) 15:18, 23 January 2012 (UTC)
Thank you! You may also agree that it is "job" of the bot owners to report problems / feature requests upstream to the programmers of the bot framework. It is not efficient if non-bot users need to dig in the framework and its talk pages. :-) If you had written the bot's framework on your own you would also not ask non-bot users to fix bugs. --Saibo (Δ) 15:38, 23 January 2012 (UTC)
I would agree with you in general and in many cases I did report problems / feature requests upstream (so far for AWB framework). However, in this case I am rather ignorant about redirects as used on Wikipedias and do not feel like I would be able to intelligently argue for software change. I personally (and manually) only add interwiki links to existing pages with exact matches. I would also not add interwikis to the article subsections. I do not think I encountered before interwikis to "articles which may be written in future", and I do not have opinion if I like that concept or not. --Jarekt (talk) 16:09, 23 January 2012 (UTC)
Even if the redir may not be replaced by an article the section title in articles changes every now and then (and by far not always {{Anchor}} is added with the old title as the incoming links are not known). In case the iw goes to a redir the iw continues to link to the correct section (after the section link was fixed in the redir). In case the iw goes to the section directly potentially iws in ~50 wikis need to be changed (manually - since that isn't done by bots). --Saibo (Δ) 16:35, 23 January 2012 (UTC)
Agreed. Sum-it-up and en:wiki redirects include frequently links to sections; there is no problem with that as they are maintained at a global level. When we stick to that level (and do not follow the redirects), we can profit from the global system maintenance. I don't know however what happens with the other interwiki links. For example Category:Schoonaarde contains a redirect and a reference to a section in an article, while if one dares to use sum-it-up, it gets all mixed up with the parent article. Curious to see what the bot update will do. --Foroa (talk) 17:26, 23 January 2012 (UTC)

  Fixed I am running now with "-noredirect" option. I tested it first on Category:Girder forks and it skipped that page. --Jarekt (talk) 12:09, 24 January 2012 (UTC)

Thanks! Andy Dingley (talk) 14:37, 24 January 2012 (UTC)
Unfortunatelly there is not much I can do about it. That is the the same set of interwiki links as used on all the wikipedias (see for example de:Oldham-Kupplung). The algorithm of mw:Manual:Pywikipediabot/interwiki.py standardizes the look of interwikis so they are consistent among all the projects. If at some point en:Oldham coupler is written than the next time me or someone else runs mw:Manual:Pywikipediabot/interwiki.py the link will be fixed. --Jarekt (talk) 15:57, 27 January 2012 (UTC)
Still happening: [6] [7] [8] [9] Andy Dingley (talk) 11:46, 2 December 2012 (UTC)
Yes, I run the code about once a year, and yes the code is still following the same standards as used on all the other wikis and is eliminating unnecessary redirects whenever found. This is how it is done on all the wikipedias. I am sorry it does not match your outlook, but the easier way to overcome this problem is to write the articles where the redirects are. --Jarekt (talk) 22:20, 2 December 2012 (UTC)
"Screw you puny human, do more work and save my 'bot the trouble" is not an acceptable result!
This is a good 'bot, doing useful tasks, but that attitude makes it just a whisker from getting 'bot permission revoked for it.
There is one position that says that this sort of change is just a bad idea anyway: redirects are Wikipedia's problem, not Commons' and it's a useful abstraction for WP to use them as a method of hiding its internals from outside sites, including Commons.
With more subtlety, then a 'bot might be excused in changing iwl targets if they point to a "clearly inappropriate" target. However that's pretty narrow. There are plenty of cases, like these, were it's demonstrably better to rely on the redirect. Even in others, it's a very marginal advantage that could be gained by any such change (Mediawiki redirects are robust and work fine as they are). In such a situation, it's the 'bot's burden to prove that changing such a link is an improvement, otherwise leave well alone. Andy Dingley (talk) 01:01, 5 December 2012 (UTC)

It says above that this problem has been fixed, but it hasn't.
Last week, JarektBot changed a link in 'Category:Circus high wire acts' so that instead of pointing to 'en:Circus high wire acts' it pointed to 'en:Tightrope walking'. This isn't the same thing. 'Category:Tightrope walking' exists, and that is where 'en:Tightrope walking' is correctly referenced. Furthermore, since 'Category:Circus high wire acts' is a sub-category of 'Category:Tightrope walking', what this robot did was doubly wrong: as well as removing a (potentially) useful link, it created in effect a new over-categorisation problem.
Please can something be done to prevent this kind of error happening in the future.
Thank you.
GrahamN (talk) 17:55, 8 December 2012 (UTC)


Rewriting interwiki links to redirects

(This is a continuation of the previous topic, relocated from Commons:Village_pump#Malfunctioning_JarektBot_and_rewriting_interwiki_links_to_redirects)

<The following comments were originally made at the Village Pump>

An old issue, raised at User_talk:JarektBot#Rewriting_WP_iwls_to_avoid_redirects_-_good_or_bad.3F some time ago. However there's no progress, it's still broken and response from the 'bot operator has been "dismissive", to say the least.

Should 'bots change iwls like this at all? Should JarektBot be doing it like this? If JarektBot is malfunctioning and the operator refuses to fix it, should JarektBot be running at all? Does anyone have the time and 'bot skills to assist in fixing this technical issue? Andy Dingley (talk) 18:43, 8 December 2012 (UTC)

Seems to be standard for interwiki bots. --  Docu  at 18:45, 8 December 2012 (UTC)
Without meaning to cause offence, I wonder whether a person who admits to being "rather ignorant about redirects as used on Wikipedias" should be operating a robot that alters links to redirects on Wikipedias? GrahamN (talk) 19:04, 8 December 2012 (UTC)
I am rather ignorant about many nuances of how different wikipedias do things, that is why I use standard interwiki.py code. The same one as used by all the other projects, so our interwikis are treated the same way as anybody else. The edit that angered GrahamN was this. In that case Category:Circus_high_wire_acts had interwiki link (aka interlanguage link to en:Circus_high_wire_acts which redirects to en:Tightrope walking, so when you are at Category:Circus_high_wire_acts and click on the interwiki link "English" you go to en:Tightrope walking. After the bot edit if you click on interwiki link "English", you go to the same place, so the objection is mostly about the aesthetics of the wikicode. If "Circus_high_wire_acts" and "Tightrope walking" are not the same concepts, than they should not be linked by the interwiki links (with or without redirects). The bot did not create those connections, only removed superfluous link in the middle, and if you believe that the two concepts are not close enough to have a link, than you do not have to be shooting the messenger: all you have to do is to remove the offending links. Restoring previous version will only work until the bot is run again, by me or someone else. If there are some pages for which we have consensus that we need to have interwiki links that do not follow the standards used by all the other projects, we should have a list of them and I will try to either skip those or undo the changes myself. Finally, all the issues with interwiki bots will be irrelevant in a year or two, when all the interwikilinks are kept in a single place on en:Wikidata.--Jarekt (talk) 05:13, 9 December 2012 (UTC)
Redirects are sometimes used for "sub-topics or other topics which are described or listed within a wider article." (See w:WP:Redirect.) So interwiki links to redirects sometimes make more sense than either a direct link to the redirect's target or a link to a section within that article, and this situation is a good example. "Tightrope walking" and "Circus high wire acts" are not the same concept - the former is broader - which is why our Category:Circus high wire acts is a subcategory of Category:Tightrope walking. It would probably be more accurate to do without an en interwiki link from Category:Circus high wire acts altogether, but that would also be less useful to readers. And yes, this will all be overtaken by Wikidata in due course. --Avenue (talk) 08:19, 9 December 2012 (UTC)
This will happen more and more as people prepare for subpages that are still to be written. Maybe a parameter such as 3|leave redirect would help for such exceptional cases ? --Foroa (talk) 08:51, 9 December 2012 (UTC)
When I run the code a year ago and heard first concerns about it I did try some code options that supposed to help with redirects, but I was informed that there were still redirect issues. I can try them again, when I resume at some point. I am running something else at the moment but I was planning to resume When I am done with other tasks ( A, B & C took over a week this year). --Jarekt (talk) 01:35, 10 December 2012 (UTC)
My objection to the behaviour of JarektBot has nothing to do with "the aesthetics of the wikicode", as Jarekt puts it. It is purely practical. Redirects are not "superfluous links in the middle" as Jarekt calls them. For a variety of reasons, some more subtle than others, they are a very useful feature of Wikipedia. But if you bypass them they become useless.
I'm not sure how Jarekt gained the impression that I had been "angered". Some things do anger me, but his hasn't been one of them. All that happened was, I noticed that a robot had made an error and, as any good citizen would, I went to its talk page and left a message about it. I don't decry the useful work that Jarekt carries out in areas within his competence. It is very laudable. Thank you for it, Jarekt. And I haven't called for JarektBot to be shut down - just for it to stop making these errors. The best way of achieving this is for it to stop changing links to Wikipedia redirects, an area in which Jarekt admits to being, and clearly is, somewhat out of his depth.
No matter how well designed a piece of equipment is, if its operator doesn't properly understand the task it is performing, then things are very likely to go wrong. Recent history is littered with real world examples of people making this mistake, some of them actually tragic. GrahamN (talk) 23:37, 10 December 2012 (UTC)
Before talking about disallowing this sort of bot run, could we please try to quantify the contested behaviour? So far, two such edits have been pointed out. I am afraid that, though unfortunate, this is nowhere near enough to incline me to shut down this otherwise perfectly needed service that Jarekt provides with his bot. Jean-Fred (talk) 11:47, 9 December 2012 (UTC)
The problematic behaviour that is being objected to is that JarektBot is routinely editing links to Wikipedia redirects so that they point to the article instead of the redirect. These changes can (and do) cause problems, but as far as I can see they provide no benefit to the user. I am not aware that anybody has called for JarektBot to be shut down - just for it to stop doing that particular thing with redirects. You don't need a quantitative study of JarektBot's activities to know that it is happening. GrahamN (talk) 23:37, 10 December 2012 (UTC)
Read talk:JarektBot. I've fixed half a dozen of these in the last week (see my Contribs) and noted a few there. Over the Summer, there were half a dozen redirects broken, just of those that involved the same WP page (a big page with many sections, where each section justifies a Commons cat and a potential separate WP article).
JarektBot (and other bots) should just leave all redirects alone, without even needing to made judgement calls about them. This can't be hard. Andy Dingley (talk) 11:36, 11 December 2012 (UTC)
The problem is that there is a single enormous code for interwiki maintenance with 2500 lines of code, which is generally agreed to be safe and correct. I am quite reluctant to change that code. There is an "-noredirect" option that should result in "bot not following redirects nor category redirects", whatever that means. I run the bot with that option last year but Andy Dingley reported to me that that did not solve the problem. I can try again using that option, but there is a good chance that it is still not going to help. I think that the best solution would be not to use interwiki links, if wikipedias do not have full stand-alone articles to link to. We can always use links in the description pages to specific articles or subsections of articles, like en:Help:Interwiki linking. --Jarekt (talk) 19:04, 11 December 2012 (UTC)

I'm no technical wizard, and the descriptions of the effects of many of the command-line arguments listed at [10] are rather opaque to me. But the description of "-noredirect" is quite clear: it simply says "do not follow redirects nor category redirects". That is exactly what we are asking for, isn't it? If JarektBot still follows redirects when that argument is applied, then there is a bug in the code. I suggest that Jarekt should do a test, to make sure.
Jarekt, when you get a moment, if you don't mind, please could you point JarektBot back at Category:Circus high wire acts (which I have reverted to its previous state) and run it with the "noredirect" argument. If it does the same thing again, that will prove that there is a bug in the code, which should be reported and fixed.
GrahamN (talk) 02:54, 12 December 2012 (UTC)

I am running some other code for last week and half, (see Special:Contributions/JarektBot) and I am about half way done (done with A-G out or A-Z) range. So it might be a week or two. The test you are proposing is exactly what I have done last year and the code did not touch the test page (see here); however Andy Dingley informed me that despite that option the bot was was still removing superfluous redirects, so this year when I run A-C I did not use that option. --Jarekt (talk) 03:16, 12 December 2012 (UTC)

I see. The place where Andy Dingley reported that JarektBot had followed a redirect even though noredirect was on, was Category:Oldham couplings. A very useful experiment, then, would be to test the code on both Category:Circus high wire acts and Category:Oldham couplings. If you could do this first without noredirect, then revert the changes that JarektBot made, and immediately run it again on both categories with noredirect, then we'd have a clean picture of what is actually going on. If you are happy to do that when it is convenient to you then I think we'd all be grateful. Thank you.
(Incidentally, doesn't this whole discussion really belong at User_Talk:JarektBot, rather than here at the Village Pump? Would anybody object to it being moved there?)
GrahamN (talk) 03:46, 12 December 2012 (UTC)

Yes my talk page would be a better place to keep this discussion. --Jarekt (talk) 20:44, 14 December 2012 (UTC)

<End of comments relocated from the Village Pump>


Creator template

Why does your bot think, this picture was created by this en:Andreas Schlüter (1664–1714)? I doubt that such edits make sense, especially if they are done by a bot. Will you re-check all edits of your bot, if they make any sense? – Felix Reimann (talk) 14:10, 5 November 2012 (UTC)

Commons:Database reports/Unbelievable life spans

Several of the entries there were caused by your bot. Please have a look. --Leyo 13:33, 30 November 2012 (UTC)

I will thank you. --Jarekt (talk) 13:43, 30 November 2012 (UTC)

17th-century photo

Hi, Can you please provide evidence for your statement that this photo was created before 1694? -- Asclepias (talk) 22:54, 4 December 2012 (UTC)

Raised here too: Template_talk:Creator#Purpose_of_this_template.3F
Should the template have been used at all in this case, or used with different parameters? Andy Dingley (talk) 00:43, 5 December 2012 (UTC)
This is a disambiguation problem when multiple people use the same name and there is no disambiguation information in the author field. I will not run the task of adding creator templates until I find a way to prevent this from happening. --JarektBot (talk) 14:10, 6 December 2012 (UTC)
Return to the user page of "JarektBot/2012".