Wikipedia:Village pump (idea lab)
Policy | Technical | Proposals | Idea lab | WMF | Miscellaneous |
Before creating a new section, please note:
- Discussions of technical issues belong at Village pump (technical).
- Discussions of policy belong at Village pump (policy).
- If you're ready to make a concrete proposal and determine whether it has consensus, go to the Village pump (proposals). Proposals worked out here can be brought there.
Before commenting, note:
- This page is not for consensus polling. Stalwart "Oppose" and "Support" comments generally have no place here. Instead, discuss ideas and suggest variations on them.
- Wondering whether someone already had this idea? Search the archives below, and look through Wikipedia:Perennial proposals.
3-segment wikilinks
- For use in making wikilinks, particularly in languages with case-endings and similar, try allowing 3-segment wikilinks:-
- [[X]] displays and links to X , as now.
- [[X]]Y links to X and displays XY , as now.
- [[X|Y]] links to X and displays Y , as now.
- [[X|Y]]Z links to X and displays YZ , as now.
- And idea: let [[X|Y|Z]] link to XY and display XZ . For example, "he [[procrastinat|ion|ed]]" would be more compact than "he [[procrastination|procrastinated]]".
- And if so, [[X|Y|]] would link to XY and display X .
- [[X]] displays and links to X , as now.
- That could be interesting. --167.58.26.73 (talk) 19:55, 3 August 2017 (UTC)
- I don't think this is that much needed on the English wikipedia, but it could be massively helpful for other language versions. I think this should be proposed at the next Wishlit Survey (the last one was in 2016), with input sought from the other language projects. – Uanfala 11:10, 6 August 2017 (UTC)
- Commenting exclusively on English as it's the only language I'm fluent in, my concern here is that while "[[procrastinat|ion|ed]]" is marginally more compact than the current "[[procrastination|procrastinated]]", it's significantly less readable, and requires the reader to parse the string in their head. Reading it for the article that will be linked to (procrastination) is still fairly trivial, but the text that will be displayed in the article ("procrastinated") is far less clear and obvious. Whereas the current form, while perhaps not perfect, explicitly displays both of those pieces of text.
- The ability to append characters to article links (e.g. [[link]]ing) is a nice shorthand because it's still easily readable as both the article name (inside the brackets) and the link text (ignore the brackets). The expanded form "[[link|linking]]" would contain two adjacent instances of the same string (in bold), so it's convenient to "reduce" them for clarity and conciseness. But by chopping up the non-redundant bits of the text, and reducing strings that aren't adjacent, isn't that conciseness coming at the expense of both clarity and convenience? -- FeRD_NYC (talk) 03:47, 19 August 2017 (UTC)
ooh yeah. it looks like great feature. wow. I'm going to want it.
Text searching
Currently the searching templates {{look from|xxxxx}} and {{intitle|xxxxx}} seem to look only for letters and numbers and ignore punctuation etc. This can be a nuisance, because, currently there is a policy to remove the comma when a page name ends in ", Sr." or ", Jr.", and to help to do this work, there is no easy way to search for these two character sequences, but someone must look at every page name by eye, and requests to remove these "senior and junior commas" come in endlessly in dribs and drabs instead of someone being able to quickly call search for ", Sr." and ", Jr." and get the job complete and done and over. Please provide an option for Template:look from and Template:intitle to search for all characters, not only for letters and numbers. Anthony Appleyard (talk) 05:34, 2 August 2017 (UTC)
- The latter is phab:T156510. I'm not sure how you mean to use the former for your use case, but it doesn't seem reasonable off the cuff. --Izno (talk) 12:31, 2 August 2017 (UTC)
- @Izno: As regards Template:intitle , what is the likely progress with https://phabricator.wikimedia.org/T156510 ? Ability to search for characters other than letters and numbers would be useful. Anthony Appleyard (talk) 05:17, 3 August 2017 (UTC)
- I do not know. That is something to ask on the phabricator task. :D --Izno (talk) 13:37, 3 August 2017 (UTC)
- There is a beta search tool on wmflabs that uses regexp to search article titles. Unfortunately, you're on your own for constructing efficient regexp searches. Also, the tool can be very slow and can return a LOT of results if you put in a too generic a regexp. For example, this search for ", Sr." took quite a while to run, but it looks like what you describe. (and I'll refrain from commenting on whether it is a worthwhile use of time to "fix" such trivial formatting differences). older ≠ wiser 14:11, 3 August 2017 (UTC)
- Having regular expressions in page search seems a bit excessive. With a one-character sign ('.') and an any string sign ('*'), that should be enough. --167.58.26.73 (talk) 19:58, 3 August 2017 (UTC)
- There is a beta search tool on wmflabs that uses regexp to search article titles. Unfortunately, you're on your own for constructing efficient regexp searches. Also, the tool can be very slow and can return a LOT of results if you put in a too generic a regexp. For example, this search for ", Sr." took quite a while to run, but it looks like what you describe. (and I'll refrain from commenting on whether it is a worthwhile use of time to "fix" such trivial formatting differences). older ≠ wiser 14:11, 3 August 2017 (UTC)
- I do not know. That is something to ask on the phabricator task. :D --Izno (talk) 13:37, 3 August 2017 (UTC)
- @Izno: As regards Template:intitle , what is the likely progress with https://phabricator.wikimedia.org/T156510 ? Ability to search for characters other than letters and numbers would be useful. Anthony Appleyard (talk) 05:17, 3 August 2017 (UTC)
- Are you just looking for https://en.wikipedia.org/w/index.php?search=insource%3A%2F%5C%2C+Sr%5C.%2F&title=Special:Search&profile=default&fulltext=1 or for something else? Whatamidoing (WMF) (talk) 22:37, 11 August 2017 (UTC)
- @Whatamidoing (WMF), Izno, and Bkonrad: That search is finding ", Sr." indeed, except that it is searching in the texts of pages, and what is needed here is an option for search for ", Sr." and ", Jr." only in page names. Anthony Appleyard (talk) 22:48, 11 August 2017 (UTC)
- @Anthony Appleyard:, the search I gave above only searches article titles. older ≠ wiser 23:00, 11 August 2017 (UTC)
- You'll have to use Bkonrad's link to tools.wmflabs.org, then. It's https://tools.wmflabs.org/grep/index.php?lang=en&project=wikipedia&namespace=0&pattern=%5C%2C+Sr%5C.%24&limit=on Whatamidoing (WMF) (talk) 23:09, 11 August 2017 (UTC)
- @Anthony Appleyard:, the search I gave above only searches article titles. older ≠ wiser 23:00, 11 August 2017 (UTC)
Tool that suggests images to add to articles without an image (using Wikidata item's image)
Many Wikipedia articles have no image despite the associated Wikidata image having a P18 (image) property.
For instance, the article Thug Behram has no image, despite the Wikidata item having a perfectly usable image. Wikidata has more images because in non-English speaking countries (ie most of the world) the local language article is usually the best illustrated, and images make it to Wikidata via infobox harvesting or WDFIST.
Idea: How about having a tool that would suggest images for articles?
Use case: I am presented with an English Wikipedia article on the left, and an image (together with its title/description/categories/discussion) on the right. If I find the image fitting, I press a button and the image is automagically added to the article. Ideally the image is added to the article infobox if there is one with a recognized image property.
Images make Wikipedia articles much more attractive and informative, so I believe that would help a lot. Such a tool could then be ported to other Wikipedias too. Is there maybe such a tool, and I have not managed finding it? Cheers! Syced (talk) 03:28, 9 August 2017 (UTC)
- Sounds like a great idea. — InsertCleverPhraseHere 04:18, 9 August 2017 (UTC)
- There is an opportunity to go in multiple directions here. It isn't just Wikidata that has images. Many articles have an image from commons in one language but not in several other language versions. Lots of species have a category on Commmons but no images in articles, or only in one language. I do a bit of categorisation on Commons and I sometimes add images to species articles when I create a new species category on commons. There are big opportunities for apps that make it easy to add images, and if you speak the language to add a caption. We have done proof of concept editathons as long ago as 2011 in London inviting donors in and showing them how to add images, even quite hesitant people got very confident when you explain that a computer can suggest a dozen images for an article, but we still need a human to choose the image. Unfortunately we lost access to the donor data when the fundraising as centralised in San Francisco. ϢereSpielChequers 09:50, 9 August 2017 (UTC)
- Indeed, the tool could be extended once the low-hanging fruits are harvested. The low-hanging fruits are Wikidata, because if Wikidata has an image then you can be 100% sure this image is a good addition to the article. On the opposite, pictures from the Commons category can be a bit random, and many are only remotely related to the topic, for instance most pictures in the category about Jonan Island are pictures of airplanes rather than pictures of the island. I am not saying Commons categories are bad, I am saying Wikidata is probably easier to start with. Thanks for the idea, and thanks for adding images to Wikipedia articles! :-) Syced (talk) 14:44, 9 August 2017 (UTC)
- VE already does this. If you open Thug Behram in VE, and choose Insert->Media, it will suggest the image. I believe it simply searches based on the article title, and perhaps a more sophisticated search would be useful, but as it stands I've found this feature very helpful. Mike Christie (talk - contribs - library) 10:51, 9 August 2017 (UTC)
- Interesting! Unfortunately 1) that takes many clicks 2) Image metadata is not available so it is hard to tell whether the picture is actually appropriate or not 3) Most importantly, going through all Wikipedia articles without pictures and opening VE for each of them hoping that some pictures will be suggested would be very inefficient. The VE approach is great for people who know what article they want to add images to. Thanks for your input! :-) Syced (talk) 14:44, 9 August 2017 (UTC)
- What metadata do you want? Right now, if you click a plausible looking image, then it gives you size, license, author, etc. data on the next screen. What else would be helpful? (Please {{ping}} me.) Whatamidoing (WMF) (talk) 22:41, 11 August 2017 (UTC)
- Whatamidoing (WMF): When I add an image, I look at the image's categories, as they often give hints about what is going on in the picture. For instance, if there is a "Chess World Cup 1987" category then I can write a better caption like "Tom Beh at the Chess World Cup 1987". If the image has a category "Joe Csze" then I will double-check who is really pictured. I also look at the discussion page in the rare cases where it is not a red link, because any information there is often useful, for instance an anonymous saying "This is not Tom Beh this is his brother" is a valuable tip. Also, a really neat thing would be to see all articles the image is used in, together with the caption for each (most writers can understand several languages so that would help a lot when writing the caption). Thanks! :-) Syced (talk) 05:48, 15 August 2017 (UTC)
- When you click on an image in the media dialog, it lists some machine-readable information (e.g., license, upload date, image size, file type, etc.) Then it has a link labeled "More information", which takes you to the Commons page. I think that you would probably find that the most useful approach, especially if you want to read the talk pages at Commons. Whatamidoing (WMF) (talk) 17:21, 15 August 2017 (UTC)
- Whatamidoing (WMF): When I add an image, I look at the image's categories, as they often give hints about what is going on in the picture. For instance, if there is a "Chess World Cup 1987" category then I can write a better caption like "Tom Beh at the Chess World Cup 1987". If the image has a category "Joe Csze" then I will double-check who is really pictured. I also look at the discussion page in the rare cases where it is not a red link, because any information there is often useful, for instance an anonymous saying "This is not Tom Beh this is his brother" is a valuable tip. Also, a really neat thing would be to see all articles the image is used in, together with the caption for each (most writers can understand several languages so that would help a lot when writing the caption). Thanks! :-) Syced (talk) 05:48, 15 August 2017 (UTC)
- What metadata do you want? Right now, if you click a plausible looking image, then it gives you size, license, author, etc. data on the next screen. What else would be helpful? (Please {{ping}} me.) Whatamidoing (WMF) (talk) 22:41, 11 August 2017 (UTC)
- Interesting! Unfortunately 1) that takes many clicks 2) Image metadata is not available so it is hard to tell whether the picture is actually appropriate or not 3) Most importantly, going through all Wikipedia articles without pictures and opening VE for each of them hoping that some pictures will be suggested would be very inefficient. The VE approach is great for people who know what article they want to add images to. Thanks for your input! :-) Syced (talk) 14:44, 9 August 2017 (UTC)
Interesting idea. Building a naive tool do to this is trivial. The problem is dealing with all the drama that comes with it and the mangling that may happen in complicated cases.
There also aren't that many apparently. A bot could probably add all images within a day or two, less than a week for sure. The problem of course is that it has false positives because it relies on mw:Extension:PageImages, and some images may come with all the drama that naturally follows any automated task of that nature. Any semi-automated tool doing this on a continuous basis would fail miserably without a lot of work because of the unstructured nature of wikitext, the inconsistent way in which templates add images to pages, and a lot of edge cases.
It might be easier to make a bot that suggests such images on the talk page, and then editors can easily add (or not) them as needed. This would also reduce the need to randomly come across an article that needs file. Currently though, it might be better to create a bot that suggests images that are used in linked interwiki pages, maybe if 3 wikis or more are using the same image it then adds it to the talk page (and / or wikidata).
See also: https://tools.wmflabs.org/wikidata-todo/wp_no_image/enwiki.html , https://phabricator.wikimedia.org/T54464, https://phabricator.wikimedia.org/T53031 . 197.218.80.151 (talk) 21:14, 9 August 2017 (UTC)
- 197.218.80.151: The wp_no_image page is great, it could be used as a source. I don't think a bot add images automatically, because only a human can write captions. Adding to infoboxes might be an exception, as some infoboxes do not require a caption. "Lot of edge cases"? I would put the image just before the first line of text that does not start with "{", if you know articles where that would not work please send me the wikilinks, thanks! Syced (talk) 06:43, 10 August 2017 (UTC)
- It is possible to generate simple captions automatically, after all the data stored in wikidata is structured.
- There are so many edge cases that are easy to find, you've already mentioned the most common, many articles have template (e.g. notices or warnings) in the first line and no image at all elsewhere, along with other issues:
Templates in first section but without images:
Images in another section
- Might deliberately have no images due to some guidelines (see comment in one of the infoboxes)
- The funny thing about "wikidata no image" is that in some cases even if the image is added it might continue showing up there because of how the pageimage extension works. The images need to fulfill certain criteria (e.g. certain resolution) before they are labelled as "page images". Just to be clear, this is a good idea, it just needs a good implementation that won't cause more problems than it solves ... 08:54, 10 August 2017 (UTC) — Preceding unsigned comment added by 197.218.88.131 (talk)
- 197.218.88.131: No problem with "Templates in first section but without images", I can automatically find a line to put the images in such cases. No problem with "Images in another section" as these articles will not appear in the list of articles with no images. No problem with "The_Jungle_Book" and similar, as Wikidata is much more stricter on copyright than wp.en: For instance, Wikidata does not contain any "fair use" image. Syced (talk) 02:25, 14 August 2017 (UTC)
With no provenance suggesting that it might have been drawn from life, the Thug Behram image linked from Wikidata blatantly fails WP:PORTRAIT. It should not be used as is. My fear is that making bots or automatic tools suggest the addition of such images means making the rest of us spend too much more of our time policing the articles they're added to and undoing the damage. —David Eppstein (talk) 07:40, 13 August 2017 (UTC)
- David Eppstein: You raise an important point, Wikidata items may contain a mistaken image. Google Scholar does not have many papers about Thug Behram, and even less about its portraits. How about limiting automatic import to Wikidata images that have a reference? By the way, what do you think a reference for an image should look like? Do you have good examples that I could study? Thanks! Syced (talk) 02:25, 14 August 2017 (UTC)
- In this case it's not mistaken (I have good faith that the artist really did intend to depict Behram) so much as valueless (if we don't know the context for the image then we can't put it into context in an article and we certainly can't use it without context as if it were an actual photographic depiction of the subject). In contrast, looking for other people from around the same time period found File:JOHN ADAIR colour corrected.jpg, a painting for which we know the artist, date, and that it was painted from life rather than from the imagination. That's the sort of information I'd look for when determining whether a painting or drawing should be used. But I got all that from its page on commons; on Wikidata it is just an image with no references. I conclude from this that the metadata on images that we're placing into Wikidata is still too sparse to be useful for this task. And because Wikidata only seems to allow "references" on image links, not any other kind of metadata, it will remain too sparse until the schema changes. —David Eppstein (talk) 04:02, 14 August 2017 (UTC)
- In the idea proposal, I wrote that the tool should show the image's "title/description/categories/discussion" to the user. That's because I believe this information enables the user to decide whether an image is usable, and know what to write in the caption. Syced (talk) 05:29, 15 August 2017 (UTC)
- At the very least there needs to be a way of flagging an image as unsuitable (making any such tool automatically skip it) so that people don't keep trying to add it over and over and over and over and over and... —David Eppstein (talk) 06:28, 15 August 2017 (UTC)
- Indeed, the tool should include that flagging feature, similar to WDFIST. Syced (talk) 07:46, 16 August 2017 (UTC)
- At the very least there needs to be a way of flagging an image as unsuitable (making any such tool automatically skip it) so that people don't keep trying to add it over and over and over and over and over and... —David Eppstein (talk) 06:28, 15 August 2017 (UTC)
- In the idea proposal, I wrote that the tool should show the image's "title/description/categories/discussion" to the user. That's because I believe this information enables the user to decide whether an image is usable, and know what to write in the caption. Syced (talk) 05:29, 15 August 2017 (UTC)
- In this case it's not mistaken (I have good faith that the artist really did intend to depict Behram) so much as valueless (if we don't know the context for the image then we can't put it into context in an article and we certainly can't use it without context as if it were an actual photographic depiction of the subject). In contrast, looking for other people from around the same time period found File:JOHN ADAIR colour corrected.jpg, a painting for which we know the artist, date, and that it was painted from life rather than from the imagination. That's the sort of information I'd look for when determining whether a painting or drawing should be used. But I got all that from its page on commons; on Wikidata it is just an image with no references. I conclude from this that the metadata on images that we're placing into Wikidata is still too sparse to be useful for this task. And because Wikidata only seems to allow "references" on image links, not any other kind of metadata, it will remain too sparse until the schema changes. —David Eppstein (talk) 04:02, 14 August 2017 (UTC)
There's an image adding section at Wikidata:The Game. --167.58.55.39 (talk) 15:05, 18 August 2017 (UTC)
Upgrading Authority control
I guess the time has come to upgrade {{authority control}}
to a core component of wikipedia. As it already uses wikidata, it would be a good step to show authority control bar above categories on pages by default, if it has values and get rid of the manually added templates in all the articles. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 07:24, 13 August 2017 (UTC)
- Authority control is not always applicable to every topic. Also strictly speaking these are controlled topic identifiers (e.g., we also have LCSH in there and strictly speaking that is not authority control) that are most often in the form of external links (I know some editors have confused them with references). If your intention is to make this some sort of core component (which I am not sure we are ready for), methinks it might be more interesting to put these vertically in the navigation sidebar, e.g., under the interwiki and interlanguage links. 50.53.1.33 (talk) 01:42, 17 August 2017 (UTC)
- I am proposing it as a core component only. The locations you have proposed seem good too. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 15:26, 18 August 2017 (UTC)
Wikiproject Members
Currently members have to be added to Wikiproject directory manually. Is there a way to auto-populate the list based on certain criteria like User pages with Userbox WikiProject, etc.? Also to know more about the level of user, his permissions and contribution history links can be shown. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 08:09, 13 August 2017 (UTC)
- Userboxes usually add users to a category, which can be used to list participants who added it to their user page. A bot task could produce a more detailed or prettier list, but I'm not sure if a currently active bot does this. —PaleoNeonate – 23:44, 13 August 2017 (UTC)
- You might be interested in Wikipedia:WikiProject Directory/All.
- Mind the gap between "happens to work on these articles" and "actually wants to work as part of a group to work on these articles". Only some editors want to be part of a WikiProject (a WP:WikiProject is the people, not the subject area). WhatamIdoing (talk) 23:58, 13 August 2017 (UTC)
- That's quite nice, thanks for the link. —PaleoNeonate – 01:54, 14 August 2017 (UTC)
- @WhatamIdoing:That is indeed great, thanks! Is it possible show these results in Wikiproject page? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 02:34, 14 August 2017 (UTC)
- You could transclude a subpage, e.g., Wikipedia:WikiProject Directory/Description/WikiProject Mythology (just pretend the page is a normal template, wrap it in curly braces, and the contents will appear on another page), but it's not very elegant, especially if you have a long list (WikiProject History's page, for example).
- User:Harej is probably the best person to ask about this. WhatamIdoing (talk) 15:33, 14 August 2017 (UTC)
Wikiproject articles by size
A list of articles by size and quality comparison would be good, just as we have importance and quality comparison. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 07:14, 14 August 2017 (UTC)
Rather nice idea, I like it. Vorbee (talk) 07:59, 16 August 2017 (UTC)
Improving Merge procedure
Article Merger procedure is pretty slow and cumbersome, and most editors choose to stay away from the same. Category merger and template merger procedures have improved over time, but article Merger hasn't progressed as much as it was required to. Few of the major factors include lack of standardisation, non-effective templates, complex process and no bots involved for any of the tasks. I would like to invite suggestion to improve upon the same. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 02:54, 17 August 2017 (UTC)
- On the contrary: The process is quite easy. You turn on Twinkle, nominate the pages for deletion, write a brief note, and determine if there's any opposition.
- The problem is getting someone to actually merge the pages. This is easy, but it requires someone to care enough to actually copy content out of one page, paste it into the other page, and then copyedit the results. If you've got maybe 100 edits and know how to create a redirect, then you already have all of the skills. You just need the time and interest in doing important but slightly boring work. WhatamIdoing (talk) 20:10, 17 August 2017 (UTC)
Or, Capankajsmilyo, you can use the RfC process instead. But, like WhatamIdoing said, merging is a lot of work with needed skills. --George Ho (talk) 20:27, 17 August 2017 (UTC)
- Some areas of Wikipedia require you to do one before nominating one, eg GA and DYK. This is a QPQ. Perhaps nominating for merge should be the same. Nominators have to do a merge before nominating another page for a merge. Graeme Bartlett (talk) 22:07, 17 August 2017 (UTC)
- How about the proposal given at Template talk:Merge#Adding criteria to template itself for a start? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 15:24, 18 August 2017 (UTC)
Unsourced info
Would it be a good idea to involve bots in the process of removing unsourced info from Wikipedia? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 02:56, 17 August 2017 (UTC)
- @Capankajsmilyo: I dont think so. First of all, it will be very difficult to explain "what unsourced content is" to a bot. How do we define it? "a citation after every fullstop"? or "a citation after every 15 words"?
I have seen content in which an entire paragraph is covered by only one source, as well as content where just one sentence requires 4-5 sources. - How do we tell the bot to handle the paragraph mentioned above? He might keep only the last two sentences of the para, and remove everything preceding to it as "unsourced". If we need a bot to this task, the bot would require artificial intelligence. I couldnt find the essay/policy where I read it, but I think a bot with artificial intelligence is not recommended on wiki projects. —usernamekiran(talk) 13:31, 17 August 2017 (UTC)
- @Usernamekiran: ClueBot NG is an artificial intelligence as we use the term when applied to bots. --Izno (talk) 13:42, 17 August 2017 (UTC)
- @Izno: Yes. I think it was related to Clue family where I read about it. There is a cap/upper limit for the IQ of bots. —usernamekiran(talk) 13:54, 17 August 2017 (UTC)
- To start with such a bot can start cleaning those articles and sections which don't have even a single source / cite. I am sure this simple logic can be fed into a bot. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 15:21, 18 August 2017 (UTC)
- Why do you think that is a good idea? --Izno (talk) 15:37, 18 August 2017 (UTC)
- To support WP:V and reliability of info on Wikipedia I guess... -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 16:00, 18 August 2017 (UTC)
- Even that isn't as simple a problem as one might think. A naive approach would be to consider any article without any <ref> tags as uncited, but that is vulnerable to both false positives (articles which use {{sfn}}, articles which use plaintext references in parentheses) and false negatives (explanatory notes in ref tags which do not cite any articles). A limited approach could be to work on category:All articles lacking sources, but even that category contains articles which have external links to sources supporting claims in the article, even if they are not explicitly used as references (Special:RandomInCategory/All articles lacking sources gave me BoyBand (film) in four tries, for instance). Even if we could target this bot precisely, however, there's still the issue that we would be losing a significant amount of potentially valuable content (over 200k articles are tagged as having no references at all) without any oversight. Even if 99% of that is crap which is better off being TNT'd, that's still 2,000 articles which have value which we are losing. Caeciliusinhorto (talk) 17:02, 18 August 2017 (UTC)
- A possible solution to preserving that information could be to move that info to talk page for active editors to consider. Then they can decide whether to source it properly and restore or to let it die. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 17:23, 18 August 2017 (UTC)
- Why do you think that is a good idea? --Izno (talk) 15:37, 18 August 2017 (UTC)
- To start with such a bot can start cleaning those articles and sections which don't have even a single source / cite. I am sure this simple logic can be fed into a bot. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 15:21, 18 August 2017 (UTC)
- @Izno: Yes. I think it was related to Clue family where I read about it. There is a cap/upper limit for the IQ of bots. —usernamekiran(talk) 13:54, 17 August 2017 (UTC)
- @Usernamekiran: ClueBot NG is an artificial intelligence as we use the term when applied to bots. --Izno (talk) 13:42, 17 August 2017 (UTC)
About offices
First of all, does every conceivable type of office really need its very own Infobox? For example, Template:Infobox Political post, Template:Infobox Bishopric, and Template:Infobox Monarchy are all arguably merge-able into Template:Infobox official post. Political posts, bishoprics, and even monarchies are all examples of offices, right?
Also, the Article Acting president is less than a stub and could easily be a Section of another Article, or maybe 2 other Articles. I could see a sort of fork-merge where it is "merged" into both Acting (law) and--hear me out on this--Regent. There are countries that don't fit neatly into a single category. The United Arab Emirates, although it consists of monarchies at the local level, actually has a President and not a King at the national level. (Although the Presidency is customarily held by the Emir of Abu Dhabi, strictly speaking the Council could elect someone else as President if it saw fit.) North Korea behaves exactly like a hereditary monarchy where the actual means of succession is concerned, but all 3 of its rulers have still insisted on calling themselves "President" and not "King." Here on Wikipedia we insist on calling Vatican City a "monarchy," but that is not what the Roman Catholic Church teaches! The Church's teachings are very clear that the Pope is not a King, but rather, a servant of Jesus Christ the King. (And this is coming from a Catholic.)
It is, however, quite cut-and-dry as to when an office is technically vacant and has a placeholder or substitute to tie things over. My point is this: The concept of a placeholder or substitute for a vacant office is much clearer than the exact borderline between a monarchy and a presidential republic. The qualifier "acting" and the title "regent" are 2 names for the same concept, albeit usually the application of that concept in different forms of government.
I know not which one of these 2 office- and officeholder-related issues to bring up first, which is why (for now) I brought them both up in the Idea Lab area of the Village Pump. The Mysterious El Willstro (talk) 03:08, 17 August 2017 (UTC)
A doubt regarding generalised notability
While reviewing new pages, I am often seeing articles of the pattern "List of programs broadcast by <TV channel>". I have even nominated an article for deletion before. Even though wikipedia has list articles, I am not sure about encyclopaedic nature of the lists of TV shows, as such a list is ever on-going, most of the lists are but this list would be increasing a lot as there are obviously many shows in a given month on a TV channel. Even though such a list would be "harmless", I believe it would not be an encyclopaedic one. (I cant think of any reason that would make such lists encyclopaedic.)
I am not sure what's going on with me. Maybe, with time, my expectations for standards of the term "encyclopaedic" have gone very high. But talking about these lists, I am not sure if a person checks such lists to find a show. If a particular show is notable enough for enwiki, the article obviously explains on which TV channel(s) it was/is broadcast. If a show doesnt have an article but is mentioned somewhere in another article, it can be mentioned where the show ran. Not exactly, but it is sort of "existence is not notability" here.
Do we have to create lists of everything that exists? These lists of TV shows are exactly same to "List of United States Armed Forces employees". A common argument to defend that list is "it is list of shows broadcast by a notable TV channel, it should be kept". A same argument can be made: "it is a list of employees of a notable armed force". Do we include only the subjects in the list who have their own article (TV shows, and employees of armed forces)?
What I am trying to say here is, I think the lists are not encyclopaedic, and we should come up with a guideline regarding the lists. Kindly let me know what you think. —usernamekiran(talk) 13:06, 17 August 2017 (UTC)
- WP:LISTN exists; I suspect that the argument for notability for many/most channels's lists can be made based on that current guidance. Where these pages might fall afoul of current PAG is WP:NOTDIR. --Izno (talk) 13:24, 17 August 2017 (UTC)
- @Izno: Apologies for not mentioning the WP:LISTN in original post. After the failed nomination I did go through it carefully, along with many other policies. Cant/shouldnt we make appropriate changes to this criteria, being a little precise about list of TV shows without being too hard/strict? —usernamekiran(talk) 13:42, 17 August 2017 (UTC)
- @Usernamekiran and Izno: WP:LISTN is a terribly weak guideline. It lists one accepted notability criteria for lists, implying that other implicit criteria exist. For complex "List of X of Y" lists (which probably most lists are, including "List of programs broadcast by <TV channel>"), it simply notes that there is no consensus. In practice this is interpreted that such lists are always notable, which is terrible and the opposite of what notability guidelines are supposed to do. We should arrive at a consensus regarding this and make the implicit assumptions in LISTN explicit. – Finnusertop (talk ⋅ contribs) 16:25, 17 August 2017 (UTC)
- List articles, such as list of programs that have been aired on a network, are frequently seen as split-off material from the main article on the network itself (in this case); if we didn't have WP:SIZE issues, that content would be part of the article on the network. Assuming we're talking only the programs that the national broadcaster airs , and not all the other programming its affiliates have, this can be historically relevant information if it is presented appropriately (eg by decades at a high level). However, wit SIZE, we often delegate such information to a separate page, and this is where and why LISTN has to be vague because there is no consensus that such splits have to have separate notability. LISTN prefers if you can, but if you try to make it stronger, there can be other valid lists that, as a whole, don't have notability as a list but are valid list in en.wiki's eyes, such as the numerous "List of people from (place)". I've offered the idea long ago that while these don't have inherited notability, which such lists are otherwise natural parts of coverage of a larger topic and just split off due to SIZE issues, that's a reasonable allowance for such lists. --MASEM (t) 16:46, 17 August 2017 (UTC)
- @Usernamekiran and Izno: WP:LISTN is a terribly weak guideline. It lists one accepted notability criteria for lists, implying that other implicit criteria exist. For complex "List of X of Y" lists (which probably most lists are, including "List of programs broadcast by <TV channel>"), it simply notes that there is no consensus. In practice this is interpreted that such lists are always notable, which is terrible and the opposite of what notability guidelines are supposed to do. We should arrive at a consensus regarding this and make the implicit assumptions in LISTN explicit. – Finnusertop (talk ⋅ contribs) 16:25, 17 August 2017 (UTC)
- @Izno: Apologies for not mentioning the WP:LISTN in original post. After the failed nomination I did go through it carefully, along with many other policies. Cant/shouldnt we make appropriate changes to this criteria, being a little precise about list of TV shows without being too hard/strict? —usernamekiran(talk) 13:42, 17 August 2017 (UTC)
External links to religious texts (e.g. Bible)
I noticed that quite a few articles contain biblequotes that have external links. Why don't we replace them with a link to Wikibooks-hosted version of the bible? I assume its public domain by now. (((The Quixotic Potato))) (talk) 21:01, 18 August 2017 (UTC)
- such as wikisource:Translation:Genesis#Chapter_1_.E2.80.94_Originally ? — xaosflux Talk 21:09, 18 August 2017 (UTC)
- @Xaosflux: Exactly. I don't see the point of linking to an external site when we (can) host these texts ourselves. (((The Quixotic Potato))) (talk) 21:36, 18 August 2017 (UTC)
- I think it's a good idea, although the choice of translation(s) should be made, the external links are likely to remain when refering to a specific translation that is not locally hosted. —PaleoNeonate – 00:12, 19 August 2017 (UTC)
- @PaleoNeonate: I am not an expert (I am an ignostic theological non-cognivist who knows next to nothing about copyright law) but I think that for example the King James Version can be hosted on a Wikimedia project, copyright-wise, and it seems to be widely used in the areas where most Wikipedians come from. https://en.wikisource.org/wiki/Bible_(King_James). Of course there may be a minority of links to translations that cannot be hosted on a Wikimedia project, but I think that that will be a small minority. (((The Quixotic Potato))) (talk) 00:41, 19 August 2017 (UTC)
- Some of these are incorporated already, see {{Bibleverse}} — xaosflux Talk 02:27, 19 August 2017 (UTC)
- We should assume that when the article links to a certain translation, it is done on purpose. Namely, the translation is the exact source that was consulted and that verifies the article content. For quotations this is painstakingly obvious: you can't quote one translation and add a citation to something entirely else. Why that particular translation is cited or quoted in the first place is a different question (for instance, KJV is not considered a particularly good translation and that is why modern academic scholarship doesn't reference it). We obviously cite the best, most reliable sources, not those that we can hoard to Wikisource. – Finnusertop (talk ⋅ contribs) 10:40, 19 August 2017 (UTC)
- Agreed, we should only do this when it is the same translation. (((The Quixotic Potato))) (talk) 19:34, 19 August 2017 (UTC)
- We should assume that when the article links to a certain translation, it is done on purpose. Namely, the translation is the exact source that was consulted and that verifies the article content. For quotations this is painstakingly obvious: you can't quote one translation and add a citation to something entirely else. Why that particular translation is cited or quoted in the first place is a different question (for instance, KJV is not considered a particularly good translation and that is why modern academic scholarship doesn't reference it). We obviously cite the best, most reliable sources, not those that we can hoard to Wikisource. – Finnusertop (talk ⋅ contribs) 10:40, 19 August 2017 (UTC)
Talk header
Would it be a good idea to add talk headers to all article talk pages? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 09:34, 19 August 2017 (UTC)
- The template documentation currently reads:
"In accordance with Wikipedia:Talk page guidelines, this template should not be added to otherwise empty talk pages"
; I myself for a short while was adding it to talk pages as part of routine citation cleanup but have stopped when an editor pointed me to this. Some (like the recent deletion nominator) have suggested to replace the talk header by automatic messages which would be another possibility... —PaleoNeonate – 17:14, 19 August 2017 (UTC)
Removal of subject's name
It is not uncommon for someone to write into Wikimedia (OTRS) expressing concern that a search engine search of their name produces some hits in Wikipedia. We have well-established procedures in the case of two situations:
- Existing article If there is an existing article about them, we explained that we do not simply delete upon request, but we show them how they can request deletion (which they typically cannot do on their own) and offer to nominate at AfD if they need help. We encourage them to express their views in the AfD as some editors feel that in the case of a borderline case the subject's views ought to be considered.
- Afd In some cases, an article about them has been deleted but the AfD discussion has some commentary about them that, while it might not qualify as an attack, is nevertheless not very positive. In those cases we may use {{Xfd-privacy}} which hides the discussion from immediate view but is easily accessible for knowledgeable editors. It will also be found in a search of their name. (As an aside, we are discussing whether this ought to be done automatically upon request or whether it should be more selective but will sort that out.)
- Other mentions In other cases, the name may be mentioned in pages such as this archive page or this archive . It is is clear to me how this should be handled which is why I am posting this, to get community feedback. On the one hand, these archive pages may be viewed as necessary for editors to keep track of various things, and the fact that it mentions that an article about John or Jane dough has been deleted is a fact and we shouldn't go out of our way to hide it. On the other hand, some people are extremely concerned about privacy, and perhaps we shouldn't tell them simply that it is too bad, we insist on retaining the information in an easily accessible way. While no indexing might make it invisible to search engines, I think we ought to be careful about overuse of no indexing, and it may turn out that no indexing makes it difficult for editors to find (some still use internet search engines rather than the internal search function.)
I'm currently looking at a real live request (ticket:2017080410013898) where a person's name is mentioned in the page I linked and they have requested that we remove it. How does the community think we should respond?--S Philbrick(Talk) 19:05, 19 August 2017 (UTC)
- AFD is currently NOINDEXed, so they are getting to those pages some other way. --Izno (talk) 20:39, 19 August 2017 (UTC)
- ...assuming that the page is somewhat recent, and that the page wasn't mirrored to another site, and that the web search engine they use respects the NOINDEXing request, and... WhatamIdoing (talk) 01:30, 21 August 2017 (UTC)
- No: AFD is categorically NOINDEXd and has been since 2006. My comment is true regardless of your other caveats--those are all other ways of reaching the page in question, as all of the major search engines respect NOINDEX. --Izno (talk) 02:55, 21 August 2017 (UTC)
- I wonder what the effect would be if you set up a redirect to an AFD page, or if you transclude the AFD page onto another one (e.g., User:Example/AFD's I'm watching).
- I don't believe that it's reasonable to assume that NOINDEXing the XFD pages is a complete solution. WhatamIdoing (talk) 19:23, 21 August 2017 (UTC)
- Did someone suggest otherwise? --Izno (talk) 19:42, 21 August 2017 (UTC)
- No: AFD is categorically NOINDEXd and has been since 2006. My comment is true regardless of your other caveats--those are all other ways of reaching the page in question, as all of the major search engines respect NOINDEX. --Izno (talk) 02:55, 21 August 2017 (UTC)
- ...assuming that the page is somewhat recent, and that the page wasn't mirrored to another site, and that the web search engine they use respects the NOINDEXing request, and... WhatamIdoing (talk) 01:30, 21 August 2017 (UTC)
I had forgotten that AfD was no-indexed. I can think of one plausible possibility. Many people right into complain that an article about them is deleted, and and we commonly point them to the AFD page. Even though it is no index, obviously they can reach it if they have the actual URL. I thought we were getting requests for removal of information as a result of a Google search but it is possible they are seeing the page because we told them about it; I'll have to watch these careful in the future to find out if they are true search engine hits.
However, whether it is indexed or not or whether there are clever or alternative ways to get around it, misses the main point of the request — given our interest in protecting the privacy of people at some level, should we accede to requests that a mention of a name in the list of deleted pages should qualify for removal. Our interest in protecting the privacy of individuals is in support of removal while our interest in monitoring our own workflows is in opposition to the request. Is there a way to accomplish both?--S Philbrick(Talk) 01:04, 22 August 2017 (UTC)
- In this sort of case I cannot see a reason to try to hide the information. Just having a name listed for deletion does not prove it was the same as the person requesting anyway. The fact that there was such a name on Wikipedia in the past is not a secret. However if a name of a page is outing, for example Jone Doe born 16 November 2007 Denver Sands phone 555-1234-123 then perhaps a bit of log hiding or oversighting can take place. Graeme Bartlett (talk) 13:22, 22 August 2017 (UTC)
Watchlist
I suggest a slight modification to user watchlist pages. Instead of just showing pages updated since the last view with a green dot, the modification proposed is to show the number of edits since the last view - say, a number following the dot or a number within a larger dot. The utility of the proposed would facilitate whether the user chooses the dif option (to see just one change) or the hist option (to track all changes since the last view). Regards Cinderella157 (talk) 08:17, 20 August 2017 (UTC)
- I agree that it would be nice to see a pending changes count. I myself generally click on the history by default, because multiple edits may have occurred. However, with the multiple option, it is possible to see a certain number of recent edits. I don't use this feature because I find it more difficult to follow efficiently. For discussions I more often click on the little arrow before the edit summary, although oftentimes this is missing depending on the editor in use or if the user left a summary. What I find the most difficult is tracking new messages in huge threads where editors also comment anywhere instead of at the bottom. In this case I generally resort to using the diff (or the history if I've not tracked it recently enough). I also enable the bold option for unchecked edits because the green bullet is very small; making this bullet larger may be a good idea. A larger edit summary arrow would also be welcome as it is difficult to quickly click. I tried using custom css to make it larger but I couldn't target it without also affecting other unwanted elements. —PaleoNeonate – 10:18, 20 August 2017 (UTC)