Wikipedia talk:Non-free content

See also: Wikipedia talk:Non-free content criteria exemptions

Remove "disputed" tag from footnote?

NFCC#1 is misleading, and it is easy to read it and think cover art can be included in articles on musicians & authors. There was a lengthy RfC two months ago about how to clarify NFCI#1, and the conclusion was to add a footnote (FN #1). That was done. Editor Devil's Advocate disagreed, and tried to remove the footnote twice, but was reverted by Masem and Black Kite. Devils advocate then added a "disputed" tag to the footnote. The disputed tag should be a temporary situation, and this issue should be resolved one way or another. We can either

1. Remove the "disputed" tag if there is consensus to do so; or
2. Re-open the RfC and have an admin formally close it.

So, does anyone object to removing the "disputed" tag from the footnote? And, does anyone object to re-opening the RfC so it can be formally closed? --Noleander (talk) 15:11, 10 November 2012 (UTC)

We should have a formal closure of the RFC just to put it to rest. I believe that the dispute tag can thus be removed since the disagreement not a consensus view. --MASEM (t) 17:11, 10 November 2012 (UTC)

Okay. I'll post a request in the Requests for Closure. I've never seen an RfC get formally closed after it is in the archives, but maybe that will be acceptable to the closing admin. --Noleander (talk) 18:43, 10 November 2012 (UTC)

I posted a request for closure here. --Noleander (talk) 18:49, 10 November 2012 (UTC)

Sigh. The RfC was closed as "no consensus". I'm not sure that is right, so I've asked for a review of the decision here: Wikipedia:AN/RFC#Wikipedia_namespace. --Noleander (talk) 18:53, 18 November 2012 (UTC)

Okay, a 2nd person says there is no consensus, and recommends a new RfC. I'll do that. --Noleander (talk) 19:15, 18 November 2012 (UTC)

Bot for NFCC#10c enforcement

I (actually the idea for a bot is from another user, but I proposed something similar in the past) recently started discussing the possibility of a bot for NFCC#10c enforcement on some pages and would like to get some more input in order to prepare a proposal to be put on WP:VPR. See User:Toshio Yamaguchi/NFCC Bot VPR proposal draft where I started drafting some stuff, mainly how the algorithm for such a bot should work. Another user agreed that 10c enforcement should be automated. This proposal is specifically for a bot that targets non-free media that is first used in an article and later added to additional articles without a non-free use rationale. The bot should probably leave a message at the talk page of the article and the talk page of the original uploader. I already made a request at Wikipedia:Bot requests/Archive 50#Bot to tag articles and files listed at NFCR but (after another user pointed out that might be problem which should solved first) I realized that further discussion is needed. Also I wouldn't make this proposal if I thought we had enough people to deal with 10c violations (I already proposed such a bot in the past, please see Wikipedia:Village pump (idea lab)/Archive 6#NFCC enforcement for a previous discussion. That discussion was perhaps made at the wrong time and poorly thought out).

The goals of this discussion are to research:

• What were the problems with such bots in the past?
• What to do in cases where an article was moved but the rationale hasn't been updated? How is the bot supposed to handle such a case?
• What is the exact functionality expected from the bot?
• What is the principle algorithm to use? (differs from the previous bullet point in that here the exact order of the operations carried out by the bot should be specified)

The overall goal is to be able to prepare a proposal for WP:VPR, so that a community consensus can be reached and then (pending the outcome of the VPR discussion) to make an explicit request at WP:BOTREQ. I think the more precisely we specify the bots functionality, the more concrete a proposal at VPR we can make and the more likely it is for such a bot to actually be coded. -- Toshio Yamaguchi (tlk−ctb) 19:40, 13 November 2012 (UTC)

The last time we had a bot enforce #10c was from BetaCommandBot, and that was after the WMF issued their resolution on non-free content. The bot was required to go through all the non-frees we had and assure that (among a few other checks) #10c was met. This bot did have community acceptance, but what started happening at the end of the day is that users that had uploaded a lot of images but without necessarily meeting #10c, or where pages have been moved without a proper redirect trail, were getting spammed with BCBot messages on their talk page and they started to complain. It wasn't this that would cause problems with BetaCommand and his bot later and why we lost this check, but it was a stink raised about image checks. However, now, the pool of images that likely fail #10c is going to be much much smaller, and I doubt any single uploader will be spammed about #10c. So there really wasn't any major issues with the bot checks before, just that no one has stepped up to take on that role since. --MASEM (t) 23:35, 13 November 2012 (UTC)

• • Actually, the number is rather larger now. During the peak of all the fuss over Betacommand, the number was hovering around 8-9k. It's now around 12k. To the general idea; you might get consensus for such a bot, but the person who writes it and (worse) runs it is committing wikisuicide. The whole 10c thing is a ridiculously stupid fiasco anyway. The requirement needs to be removed from NFCC for good. The Foundation isn't backing it up, and editors who enforce it are routinely dragged over the coals. Further, enforcement is completely ineffective anyway, bot or no bot. Isn't that right, Masem? --Hammersoft (talk) 02:59, 14 November 2012 (UTC)
    • You may want to check that. That logo did get readded to two articles - with appropriate #10c rationales for each additional use, and none of the other stations have added it back yet. So yes, the system works regardless of your dismissals. --MASEM (t) 03:46, 14 November 2012 (UTC)
      • I'm aware of that. The point is, post facto, as it has routinely been happening. I've pointed out plenty of examples to you Masem. --Hammersoft (talk) 05:35, 14 November 2012 (UTC)
    • If the policy should be changed, then let the change occur first. As the policy stands, I support User:Toshio Yamaguchi working to get a bot in place to help assist with NFCC #10c violations. Also, where are you getting the 12k estimate from? --Odie5533 (talk) 03:54, 14 November 2012 (UTC)

• Regarding the User:Toshio Yamaguchi/NFCC Bot VPR proposal draft, I don't like the idea of automating posts to Wikipedia:Media copyright questions. I think the bot should either maintain its own list of 10c violations or tag the images as possible 10c violations for user review. I do support the idea of automated checking for NFCC 10c vioaltions. Also, Your pseudocode is hard to read. --Odie5533 (talk) 03:54, 14 November 2012 (UTC)
  • The idea is for the bot to make a post at MCQ only in case where a file page includes neither a free license tag, nor a non-free file copyright tag, which is a case I believe the bot could detect but could do nothing to resolve. Regarding the pseudocode, I will try to rewrite it as soon as I find the time so that it will be better to read. Also please note that the algorithm is far from being complete yet and might need further work. -- Toshio Yamaguchi (tlk−ctb) 08:57, 14 November 2012 (UTC)

• Point 1 is faulty. It's not enough if an image has a free licence. Photos of non-free 3D artworks are supposed to have two licence templates (a free one for the photographer's contribution and an unfree one for the artwork). It may be better to check whether the image appears in Category:All non-free media.
• Point 4.2 is also faulty. Files should only be tagged as orphaned if all transclusions are outside the main namespace. WP:NFCC#9 also says that non-free files can't be used on disambiguation pages, so I suppose that files can be tagged as orphaned if they only appear on disambiguation pages.
• WP:NFCC#10c requires you to indicate the page title, but maybe not as a link. However, there's also another problem: many people list e.g. "John Doe" as article title, although the article might be called "John Doe (politician)" or "John Doe (singer)" whereas the article "John Doe" is a disambiguation page. Strictly speaking, the fair use rationale is faulty, but it might be too picky. Also, many articles use non-standard fair use rationales instead of the usual templates. I guess that you can only check whether the page title is mentioned somewhere on the file information page, but not whether the page title is part of a fair use rationale. Still, if the page title doesn't indicate the title at all, then the file has either no fair use rationale or a faulty one.
• Also, it sounds like suicide to run a bot which performs this task, although it is definitely needed.
• The idea of a WP:MCQ notice sounds bad: the page would probably be spammed with lots of images. If the file doesn't contain any licence template at all, it could maybe be tagged with {{subst:nld}}, or placed in some maintenance category for people to check. --Stefan2 (talk) 13:58, 14 November 2012 (UTC)
  • BCBot would not require a link, just a reasonable regex match of the text on the file description page of either the article name or a redirect to that article (I believe BC including de-accented variations as well). That is, if the image was used on "John Smith (artist)", the bot would require "John Smith (artist)" though this didn't need to be wikilinked, while "John Smith" alone would not be sufficient. As long as the exact behavior is spelled out of what the name matching is, it shouldn't be a problem to be tagging #10c like this. -MASEM (t) 14:05, 14 November 2012 (UTC)
• With regards to the "John Doe" vs. "John Doe (artist)" problem; Betacommand was extensively (and sometimes brutally) attacked for not changing the target of the rationale from the former to the latter if the image was used on the latter and not the former but the rationale pointed to the former. Unless you have a plan in place to deal with the attacks that will happen as a result of a bot run like this, the failure of the bot will be epic in proportions. It will cause far, far more disruption than any notional disruption that is caused by not having a rationale. The reality is that not having a rationale doesn't cause a disruption. What it does cause is a failure of compliance with policy, which only a very few people bother to enforce. In exchange for 'fixing' this problem, we'll be causing thousands of articles to be stripped of images, thousands of talk pages spammed, and noticeboards to light up like a Christmas tree. Forgive me, but I fail to see the point, or any benefit worth the mass disruption this will cause. --Hammersoft (talk) 16:48, 14 November 2012 (UTC)
  • This will be an issue, but its one that I think we address when a BAG request is made, specifically pointing to what happened in the past and how to make sure that policing policy is not attacked. It might require a #10c maintenance board where images tagged can be double checked for obvious typos/broken moves w/o redirects, and so forth, though that should not function as a board to create rationales where ones don't exist. --MASEM (t) 17:18, 14 November 2012 (UTC)
    • Policing the policy has always been attacked. It's a given. It is also a given there is precious little support for the people who do the policing. Rarely is anyone admonished for even the most verbally brutal assualts against people who police NFCC. Regardless, this 'solution' fixes nothing. As I made an analogy before, it's continuing to pick up the cow patties in the field and never wondering how the cow patties got there in the first place. Case point; User:Pjoef, with more than five years experience and 38k+ edits to this project fails to see anything wrong with adding File:MMFF LOGO.jpg to dozens of articles and doing nothing about adding rationales. Until serious efforts are made to stop the cow patties being produced (most especially by very experienced editors), I will not support any effort to clean up the cow patties. Cleaning up the cow patties isn't effective. It does nothing to fix the problem. All it does is (very) temporarily remove a symptom. --Hammersoft (talk) 18:30, 14 November 2012 (UTC)
      • Did anyone inform him that he didn't add rationales when he added those images (which appears to be cut/paste operation?) Doesn't look like it, and that's part of the problem is that unless we have a bot to inform users of #10c problems, or change the Wikimedia software to immediately flag non-free image additions in some way, we simply can't stop users from adding non-free to begin with. That means its an education problem - since we don't have any requirements for editors to read any policy (much less NFC) before they edit, they can be completely oblivious to how to handle NFC. We can't do anything before that point. That's why a bot to do #10c is at least going to inform users of the problem and that will make them aware. Moaning and bitching that people are ignorant of NFC is not helping. --MASEM (t) 18:58, 14 November 2012 (UTC)
        • Masem, your tone in addressing me is becoming increasingly uncivil. You previously accused me of being "obstructionist" and a "petulant child". I disagree with your approach to NFCC enforcement. That doesn't constitute moaning or bitching or being a petulant child. I am asking you now for the second time to cease engaging in such uncivil dialogue. As to the points you raise, needlessly encumbered with insults; No you can't stop them from adding non-free content nor would we want them to. The point is that if after 38k edits and 5 years of experience an editor doesn't have any idea about the requirement for rationales, having a bot running around obliterating his work ISN'T going to help anyone. It WILL create acrimony. We've seen it before. There are other methods of handling this, without a bot. I've been trying for months now to get you to see this, without success. --Hammersoft (talk) 21:59, 14 November 2012 (UTC)

• • • • • • Your other methods all appear to involve getting rid of the NFC policy because you claim can't be enforced, or that no one including the WMF cares about its enforcement. Why have you not proposed this in an RFC? Instead, you're sitting and moping and taking valuable time from actual discussion about how to proceed forward. You're disrupting the work of others, as long as WP:NFC remains policy. --MASEM (t) 22:06, 14 November 2012 (UTC)
            • Masem, this isn't the first time you've claimed I am disrupting the work of others for voicing an opinion. Please stop. I am no more disruptive in voicing my opinion than you are in voicing yours. I've outlined methods before. I've done so again below. One option, yes, is to get rid of NFCC policy. That's not the only option. It's the one I am in favor of, that is all. Please stop insulting me, claiming I'm obstructionist, disruptive, etc. Comment on content, not on the contributor. Thank you. --Hammersoft (talk) 22:17, 14 November 2012 (UTC)

Reply to Stefan Regarding point 1: Is that a requirement? I checked some random images from Special:WhatLinksHere/Template:Non-free 3D art and found none that have a free as well as a non-free license template. Do you mind pointing me to an example? Also, is that a rule written down somewhere? I admit that I've never encountered such a case myself (at least not that I could remember).

Regarding point 2: I am not sure I understand your point here. 4.2 would return false only if there are no transclusions at all, otherwise 4.2 would return true. Do you mind clarifying what you mean, I don't seem to get it.

Regarding 5.1: Yes, indeed. Thanks for catching that. You are right that a wikilink is not actually required in the rationale, only the article title. So what is the bot supposed to do if a non-free image has a rationale without wikilink saying The Desert but the file is used in The Desert (film)? Maybe the bot could try to check whether there is a rationale that is similar in some characters or words to the title of the page the file is used on and not remove in that case.

Regarding the MCQ notice, your proposal of having the bot place such files in a special maintenance category sounds good.

Finally, regarding this being a suicide run, my opinion is that the only other option would be to trash the policy entirely, which of course would mean to delete all non-free media, which I believe isn't going to happen. Oh, and btw this being a suicide run seems to make it even more fitting for a bot, since contrary to a human editor it won't give a damn about what people say. -- Toshio Yamaguchi (tlk−ctb) 21:39, 14 November 2012 (UTC)

Point 1: See Special:WhatLinksHere/Template:Photo of art. A photo has two copyrights: the photographer's copyright and the sculptor's copyright. I believe that the idea is that the photographer's copyright usually fails WP:NFCC#1 and WP:NFCC#8. The article about a copyrighted sculpture is not specifically about the photo taken by Getty Images; any other photo of the same sculpture would serve the same purpose, and then you might as well obtain a free licence from the photographer. I think that this was discussed recently at WP:MCQ. You might find something useful in the archives to that page.

Point 4.2: Sorry, I misread the algorithm. --Stefan2 (talk) 22:03, 14 November 2012 (UTC)

It seems we should be fine if we simply regard any file like File:Jupiter.SouthStation.agr.jpg as non-free and ignore the free license tag. Also, the algorithm, as currently written would correctly detect this file as non-free, because what matters is whether a non-free copyright tag is present, not whether a free license tag is also there. -- Toshio Yamaguchi (tlk−ctb) 22:21, 14 November 2012 (UTC)

• But, the person who runs the bot will be the victim. It's still a suicide run. I want to point out that some time ago I did a study regarding the success of notifying the article talk page where non-free media was used but failing 10c in terms of whether the image was corrected with an added rationale or the image was removed. The notifying the article talk page via bot (which was done before, by the way), was an abject failure in this study. Rarely were the images fixed or removed. A parallel study where the images were removed for failing 10c showed considerably more success (most images remained off the articles). The idea that if we simply notify the people in question the problem will be fixed is flawed. With decreasing manpower all over the project, it was a problem and will become a larger problem.
• As to trashing the policy or not allowing non-free content at all; those are not the only options. One of the options is the status quo; 10c goes unenforced, the Foundation doesn't care, and only a very small handful of editors care (easily less than 20, probably less than 10). Another option is to develop better interface tools to handle the addition of non-free content into articles, such that compliance becomes remarkably easier to an editor. As is, an editor has no reason to have any knowledge of NFCC policy. You can be an editor here with years of experience, tens of thousands of edits, and not once have run into NFCC policy despite working with images on many occasions. Addressing this problem is another option, one that address the fact that cattle keep wandering into our field doing what cattle do.
• Personally, I think getting rid of the NFCC policy is the best route. That's not the corporate line though, and it's blasphemy to speak it. Reality; in all the years we've tried to keep a lid on NFCC use, no one has come up with a solution that works to minimize non-free content and keep us focused on our mission. I.e., no one can solve it. Strongly related to that is the Foundation simply doesn't care. They've been approached about this on quite a number of occasions, and refuse to even discuss it. The last time it even came up at one of their board meetings was five years ago now. They simply don't care. They make millions upon millions per year, and there's no motivation to upset that income stream. --Hammersoft (talk) 22:14, 14 November 2012 (UTC)
  • "the person who runs the bot will be the victim"
  • The victim of what? The victim of possible disagreement regarding the removal of a non-free image? If so, then what's the problem? Since I believe only an experienced editor and bot op would run such a bot that person should (hopefully) be able to deal with such a case in a meaningful manner. -- Toshio Yamaguchi (tlk−ctb) 22:30, 14 November 2012 (UTC)
    • People who have attempted to enforce 10c have come under an excessive amount of hate and derision. The people who levy this hate and derision are rarely, if ever, warned or blocked for their hate filled attacks. Usually, over time, a large number of threads erupt on various noticeboards, and the person who runs the bot or does the enforcement directly is found at fault. That's the short of it. Please see my comments further above; having a bot do this work solves nothing. It just (very) temporarily puts a bandage on it. --Hammersoft (talk) 22:36, 14 November 2012 (UTC)
    • The incidents that are most at the top of my mind are editors that complained to Beta and Hammer and others that were #10c enforcing in the past when the article name in the rationale was different from the article it was used on, expecting these #10c enforcers to "fix" the simple problems. Granted, some were simple typos, but I remember at least one case where there was a messed up move where a redirect wasn't left, and the editor complaining about the #10c notification complained that those enforcers should have gone back through the article history to correct it themselves. Granted, per BURDEN, the corrections should all be on those wanting to use NFC and thus they should fix the corrections, but these people were being chewed out for not doing "simple" fixes, though what was considered simple was broad and inconsistent. Thus, when getting approval for a new bot, it needs to be absolutely clear where the responsible and to what degree of responsible there has to be on the bot operator to fix "simple" problems (if any should be fixed this way). --MASEM (t) 22:38, 14 November 2012 (UTC)
      • Also consider this page [1] (note that the page was later courtesy blanked); this was before anything with Beta and ArbCom, at the near end of the image checks requires by the recent WMF resolution passage. --MASEM (t) 22:48, 14 November 2012 (UTC)
        • FWIW, technically NFCC#10c is not a Foundation requirement. [2]. Not all EDPs require it. Jheald (talk) 00:52, 15 November 2012 (UTC)
          • So let's ditch it. --Hammersoft (talk) 14:08, 15 November 2012 (UTC)
            • I would argue that while it is not set in stone, #10c or something like it is necessary. The WMF resolution requires, per #2 "Non-free content used under an EDP must be identified in a machine-readable format so that it can be easily identified by users of the site as well as re-users", and per #4 "They must be used only in the context of other freely licensed content." Thus, the need for the EDP rationale to spell out where the content is being used is necessary to have a machine-readable format that demonstrates that the image is only being used with free content (in our case, mainspace). Of course, that's one way to read it; I know its not explicit, but #10c lets us satisfy those in the easiest way possible. --MASEM (t) 14:19, 15 November 2012 (UTC)
              • I interpret this to mean that there must be a common text or template on all non-free files so that a computer can determine that the file is non-free, but that the machine doesn't have to be able to tell why it is used under the EDP. In our case, the resolution is satisfied if the file is a member of Category:All non-free media. --Stefan2 (talk) 14:28, 15 November 2012 (UTC)
                • I'm not expecting the machine to understand the rationale (in fact, that would be bad if it could). However, the Resolution does require a rationale for an image, so no, just being in the non-free media category isn't sufficient. Where we possibly differ is that we've asked for a separate rationale for each use of the image, which is clearly above and beyond the requirements by the Foundation. Arguably , one of the biggest points of contention in #10c is when users re-use a non-free over and over without adding separate rationales, even though each use, outside of the article, is for the same purported purpose. Barring other issues that are subjective (is this overuse per #3a?, is #8 met? etc.) are we being too anal retentive about a separate rationale when the use is effectively the same? #10c is by far the easiest requirement to met, and perhaps if we allowed for a single rationale to serve multiple articles, as long as they were all named, that might help downplay any complaints on the process. Of course, that said, if we go that route, we need to have different discussions when reuse is appropriate (case in point is the example Hammersoft gives of a logo repeated on 20+ instances of the same film festival) - not a problem for this #10c to even attempt to solve. --MASEM (t) 14:37, 15 November 2012 (UTC)
    • While Hammersoft does a brilliant job pretending Betacommand was a victim, he is right about one thing: If you go on a massive run of deleting images, you're going to have a pile of angry editors at your doorstep. Such a bot has to be mindful of not only the NFCC policy, but the fact that human beings are on the other side of the discussion. That, ultimately, was BC's failure. He not only had an overzealous attitude towards NFCC enforcement, but he behaved like an ass about it and dug his own grave as a result. There are, in my view, several ways that it could be done better this time:
      1. The truth of the matter is, our image policy is confusing as hell. People tend to be angry about potential deletions because they don't understand what the problem is. So when tagging, you need better and more polite messages. Messages that not only explain the problem, but also the solution. If there isn't one already, build a FAQ or workflow on how to do a proper FUR for various common image types - logos, album covers, movie posters, etc.
      2. Notification. Getting Betacommand to do this was like pulling teeth with tweezers, but having him put a message on article talk pages saved hundreds of images. My suggestion: Learn from the unsourced BLP fiasco and create lists relevant to Wikiprojects and leave notifications at those projects before you even begin tagging for deletion. I can assure you that several projects would follow through on the list.
      3. It shouldn't be hard to list images that don't have valid article targets due to moves to or from disambiguation titles. It may be impractical to go through the entire history of an article to see if it was moved to a wildly different title, but this much should be doable.
      4. There would ultimately have to be runs of tagging to delete images that nobody can or will apply a proper FUR to. I'd keep the number of images so tagged each day relatively small. There's no reason why this has to all be fixed tomorrow when taking some time will accomplish the same goal with less overall drama.

      Just some thoughts from someone who got stuck wasting a hell of a lot of time back in the day because of BC's methods. Resolute 23:45, 14 November 2012 (UTC)

      • I will thank you for not attributing thoughts to me that I did not have. I barely even mentioned Betacommand, and that is directly intentional. While Betacommand is one person who has enforced 10c in the past, he is _ONE_ person. I refer generally to all the people who have done extensive 10c enforcement. There's not a one I know of who has not come under attack repeatedly, and without anyone stepping forward to support their actions. Notifying, not notifying, being nice, not being nice, none of it has made a bit of difference. We are years down this road and despite the best efforts of everyone who has tried, we are in a worse position. We would be performing a serious disservice to ourselves to think this prior effort was somehow all Betacommand's fault, and we can somehow do a better job. Not to mention that such a stance completely ignores the efforts of many editors who tried and failed to fix this. There are lessons to be learned from how Betacommand handled 10c enforcement. But, they are very, veyr far from being the ONLY lessons. For that, I'll continue to return to the point that 'fixing' this problem does not involve picking up what the cattle are leaving in the field, but rather figuring out a great way to keep the cattle out of the field in the first place. Solve that, and there is no need for this bot. --Hammersoft (talk) 01:58, 15 November 2012 (UTC)
        • There are a lot of other policy-mandated aspects (WP:V, WP:NOR, BLP) where we cannot act before the deed is done - we can't force editors to read any policy, nor can the software catch these problems until after the edit button has been clicked. It is completely unreasonable to expect NFC to be "special" in any way. --MASEM (t) 02:36, 15 November 2012 (UTC)
          • It is possible to change the user interface to take into account changes where images have been added. That is one of the options. Special? Yes, it is special, and deservedly so if our mission means anything anymore. --Hammersoft (talk) 14:08, 15 November 2012 (UTC)
            • Which requires a change in the Mediawiki software, which likely means that the WMF has to support it. Given the indifference you've found with them on talking about it, I don't expect that to happen. However, the next best thing is a bot that patrols all such edits shortly after they are added and deals with problem images near-immediately (lets say, within 24hr of the edits), which can be set up irregardless of the Foundation's indifference. --MASEM (t) 14:19, 15 November 2012 (UTC)

I'm not sure how useful it is to send an automatic msg to the original uploader. He/she could possible help adding the rationale's or remove the images from the article. But it's not really that editor concern about the use of his uploads on other articles. Some editors could be swamped with msg's then. Garion96 (talk) 14:41, 15 November 2012 (UTC)

When a non-free is added to an article, the bot should check for whatever compliance it can that is objective (primarily #10c). If a #10c problem exists, the bot should notify the uploader (who supposedly created the initial rationale), the editor adding the image, the page the image was added on, and the file page. Yes, in the 2007-2008 image compliance run, people were swamped with such messages and boy, did they complain. I believe that the conclusion there was that BCBot was the only immediate option to reach compliance in time, but in the future care for messaging users (including opt-out mechanisms) would need to be in place. The reason to include the original uploader is that I've seen some editors take the original rational and change the article name to get their second use of the image appropriate, but now invalidating the original use. Yes, a new bot would have to check each existing use of the image when one new use is added due to something like this. --MASEM (t) 14:47, 15 November 2012 (UTC)

#10c would be fine if the rationale indicated the pages the image would be used on, not contained the name of the page the image would be used on. This, then, could not be enforced by a bot, leaving the question of bot enforcement moot. As has been pointed out many times, the requirement that the rationale contain the name of the page makes innocent actions cause images to be deleted when someone keeping track could have maintained the rationales.

1. E merges an article A into A' (foo), and creates a disambiguation page at A
2. A bot or a passing user notes that Image:A.jpg is used in A' (foo), but there is no rationale for that page. (A smarter bot might notice that there was an rationale pointing to a page on which the image was not used, but yet another bot might remove that.)
3. Image:A.jpg is deleted from A' (foo), per #10c.
4. Image:A.jpg is then deleted, per #7.

The regulars at article A' (foo) may have been informed, but probably not editor E, nor would E necessarily know what to do even if informed. — Arthur Rubin (talk) 15:01, 15 November 2012 (UTC)

Maybe better to read the article history & notify the user who added the image to the article instead? In many cases, this will be the original uploader, but not always. If you tag {{subst:nfurd}}, the file remains as-is for a week. Should the bot add a dated template and then wait for a week before removing the file from articles? --Stefan2 (talk) 15:06, 15 November 2012 (UTC)

This is why I don't think this bot can automatically remove images, because of situations like this. Tag them and put them into a maintenance category for a last human check before being removed from articles (with a chance to catch such errors)? Sure. But I don't think we can have a bot automatically remove such images. --MASEM (t) 15:12, 15 November 2012 (UTC)

• Which is effectively useless. We had notifications by themselves for a while at one point. It failed to provide any significant improvement in compliance. --Hammersoft (talk) 15:20, 15 November 2012 (UTC)

• With a BAG-approved task w/ wide notification that its going on, and messages on the uploader, the editor that added the image to the article (based on diffs), the article talk page where #10c isn't met, and the file page all within 24hr of adding, and a final human check before the removal enforcement, any editor that complains about images being removed via #10c simply have no legs to stand on. But this point (re: attacking the bot or bot operator) needs to be stressed in the BAG. Given that #10c is objective and one of the allowed exceptions of 3RR, we're basically covering our basis. Someone will still get upset, I'm sure, but really, that's basically likely complaining after your power gets turned off after getting numerous notices at your door that it was going to happen. --MASEM (t) 15:26, 15 November 2012 (UTC)
• As I said, learn from the uBLP mess. The zealots who wanted a scorched earth approach fought me over the idea of project-based lists. When they got implemented anyway, most projects tackled the issue with a will. Dashbot continued to run this task for over a year and a half, and most projects responded to new additions within days. You want a method to not only clear the backlog, but to pen new additions? Do the same thing. Given 10c lacks the same push that uBLPs had, I don't expect the results to be as dramatic. But give me a list of images used on Canada or Hockey articles, and I expect there would be more than a couple of volunteers to address the problem. I think you would find the same of most active projects. You will always have better success with central notification targeted at interested editors. Resolute 15:46, 15 November 2012 (UTC)

Here you go;

• File: Hockey india Logonewone.jpg
• File: Hockeysweden.svg
• File: International Hockey Federation Logo.svg
• File:Chinese Ice Hockey Federation Logo.png
• File:Hockey india Logonewone.jpg
• File:Hockeyrussia.PNG
• File:Hockeysweden.PNG
• File:IHAI.png
• File:Korea Ice Hockey Association.png
• File:Logo Chile Hockey Césped.png
• File:MichStars.PNG
• File:MLRH.jpg
• File:PIHA Logo 2.gif
• File:PIHA Logo.gif
• File:PIHA08.jpg
• File:Red Baron WCHA Final Five.png
• File:RIT Tigers Hockey logo.jpg
• File:Rusjr_logo.png
• File:Turkey Hockey Logo2.png
• File:UkrainianPHL.png
• File:UkrainianSupremeLeagueLOGO.png
• File:University of Edinburgh logo.svg
• File:WorldSeriesHockey-logo.jpg

Fixing these would rectify 0.23% of the 10c problems on the project. Enjoy. --Hammersoft (talk) 16:02, 15 November 2012 (UTC)

• Thanks. I'll address that list at some point today. I would add this reinforces my point. A giant list of 10,000 is pretty much unmanageable. A small list of 20-50 (or in some projects, more) is easier to deal with. The hockey list might only be .23% of the backlog, but multiply by 100 projects, and you start to put a big dent in it. If you can, place a list related to The Canada project at either my talk page or WT:CANADA, and I'll see if it can be dealt with as well. Resolute 16:19, 15 November 2012 (UTC)

• I go back to what I've been saying before. Fixing these is just a bandaid. It addresses the symptom of the problem without paying any heed to the problem itself. We can make these lists all day long, but a year down the road we'll be in the same place; an excessively long list of images not complying with 10c. Many editors have been hammering at the 10c issues for years, with no improvement in the situation. --Hammersoft (talk) 16:24, 15 November 2012 (UTC)

• My point all along has been continuously updated lists. Wikipedia:WikiProject Canada/Unreferenced BLPs, as an example, is already watched by several editors. If Dashbot or Toshio's proposed new bot were to implement a task that adds a continuously updated list of images failing 10c, then you stand a far greater chance of mitigating or preventing that list from growing. You're right that giving me a list of what needs work in these project today only solves today's issues. Give me a list that continuously updates, and I can do my part to solve tomorrow's as well. Resolute 16:30, 15 November 2012 (UTC)

• Coming up with a better bucket to bail the boat with fails to address the problem of why the boat is filling with water in the first place. --Hammersoft (talk) 17:14, 15 November 2012 (UTC)

• As long as WP maintains the goal of being an open wiki with minimal barriers for people to start editing, we are always going to be filling with water from new editors that take no effort to understand the processes required. As this is a fundamental goal of WP, it's a waste of time to consider how to stop that and instead figure out how best to education new users as fast as possible. --MASEM (t) 18:28, 15 November 2012 (UTC)
• That same defeatist attitude followed the uBLP opponents as well. "Can't win, don't try" was not the attitude that saw a 70,000 article backlog reduced to the present total of 487. Resolute 18:39, 15 November 2012 (UTC)

• <scratching my head> Perhaps I'm not articulating my position here very well. I keep trying to bring home the same points, and it's like I'm describing a tree and others are discussing a car. Maybe we're not speaking the same lingo or something. I'm at a loss here. I never espoused a "can't win, don't try" attitude. Ever. I also do not consider it a waste of time to identify and address root causes of this problem and seek the best ways to solve it. This is a fundamental principle in problem solving. Yet, I can't seem to convey this. There are multiple different ways in which the root cause(s) can be addressed. Yet, to a person, every one of you is absolutely sold on the idea of a gold plated bucket with extra widgets and an automatic digital logger of how many buckets have been bailed. There's ~10 other people in this thread, and not one of you is interested in root causes and addressing them. So, the problem must lie with me; I must be wrong because I think repeating the same thing over and over again and expecting a different result is insane. I give up. Go ahead and have your pretty little bucket. We'll be right back here a year, two years, three years from now with the same problem, and every one of you will still be dreaming of a day when we have an even better gold plated bucket, this time not operated by <insert person who did it wrong last time>, maybe this time with a Mr. Energy attachment that allows distributed computing world wide to deal with the problem. You can't fix Niagara Falls with a pipe wrench, people. --Hammersoft (talk) 18:53, 15 November 2012 (UTC)

• (Aside: Why do all these indentation codes have a "*" in them. It seems to violate WP:TALK, but I'm not going to reformat a controversial discussion in which I'm involved). This section is about #10c enforcement. NFC education is another topic, perhaps appropriate for this talk page, but not appropriate for this section. — Arthur Rubin (talk) 01:51, 18 November 2012 (UTC)

• NFC maintenance and education go hand in hand. We can't launch a #10c bot without being assured users knows what fulfilling #10c is. You can't discount that issue. --MASEM (t) 01:55, 18 November 2012 (UTC)

• @Masem: I seem to recall prior consensus that 10c enforcement is NOT exempt from 3RR. At best, it's a grey area. --Hammersoft (talk) 15:46, 15 November 2012 (UTC)

• For #10c, you either have or you don't have a rationale with the name of the article the NFC is used in. It is, per 3RR , "unquestionably" violating NFC, if it doesn't have it. Of course, we don't have to be jerks about it and to avoid that, we give warnings and enough times to get it fixed. But if after 4 messages and 7 days and a human check that the image is removed by lacking an objective requirement, people edit warring to keep it back in are violating 3RR. (The last time the NFC/3RR came up that I can recall was over currency images, which is not "unquestionably" a violation.) --MASEM (t) 16:09, 15 November 2012 (UTC)

• I'm not referring to the currency article (still extant) situation. From the mass rancor I've seen happen, I wouldn't in my wildest dreams violate even 2RR to enforce 10c. Not that I enforce 10c anymore. --Hammersoft (talk) 16:11, 15 November 2012 (UTC)

Going back to an earlier subthread: Notifying the user who added the image to the article was not (always) done by the bot. Notifying a user who renamed the article was not done by the bot (if it were done, the bot could change the name in the image rationale, and we wouldn't have lost images by rename.) If those were done, and the bot notified the last few editors who added the image or renamed the article (especially editors who renamed the article away from (an) article named in the rationale), I would have fewer objections to the bot.

I object to #10c because it's not at all necessary to track a challenged rationale, but only to support bots; and, further, because it was considered by at least one bot run by a now-banned editor to require a wikilink to the article, rather than just naming the article.

There are several ways an image can get into an article other than an editor specifically thinking it belongs there.

1. Rename; could be mitigated if the bot was properly written to track renames of the article the image is now in, as noted above
2. Split (if the image were properly in the original article, then it's likely appropriate in the split article, if relevant)
3. Merge (not as obvious, but an image might be appropriate in the merged article — however, see album covers in discographies, if the albums originally had separate articles)
4. Reverting "blanking" deletions or vandalism (in which the editor adding it might not be aware there might be a problem); could be mitigated if previous editors who added the image were also notified.

If a bot were to track all of these which can be tracked by a bot, and the notification provided a template for the notified editor to notify the appropriate editor if it's not a case where the bot can track the appropriate editor, the bot might be acceptable, as far as I'm concerned. — Arthur Rubin (talk) 17:54, 15 November 2012 (UTC)

First, BCBot never required a link. It looked for text in the rationale to match the article names or any redirects to those in the text. So this is a fallacy as #10c doesn't require a link (though its encouraged).

As for notification, the way that would be best to run the bot is to do it based on reviewing all edits that have modifications involving "file:" links so that the one that did the edit can be messaged, along with the uploader and affected talk page. (previously, it was based on running through images and looking for the lacking rationales, meaning that whomever added the image was left out of the loop) A bot can only do so much forensic work and that's why the bot can only tag and place images in a maintenance category to be reviewed by humans before deletion. --MASEM (t) 18:23, 15 November 2012 (UTC)

• I find it humorous how far down the tracks this proposal has already gone in terms of all the bells and whistles this bot must have in order to meet requirements. God help the contortionist code writer who has to come up with it. --Hammersoft (talk) 18:53, 15 November 2012 (UTC)
  • I don't necessarily see a problem with that. I don't think anyone has ever said that only the person actually doing the code writing has to come up with the entire functionality him-/herself. Also I am sure one could find ways to make it possible for several coders to work on the bot code (since nowadays we have things like Dropbox and other stuff). Also, since I do not expect the person coding the bot would run it without prior getting back to the community, the coder simply cannot be held responsible for it alone if the bot is running amok. The bot getting approved means other people (who hopefully would be able to understand the bot code) have seen it and approved it. In the end it shouldn't matter who codes and runs a bot at all. -- Toshio Yamaguchi (tlk−ctb) 20:25, 15 November 2012 (UTC)

• Comment Regarding files where the rationale could be simply fixed by changing the article name in the rationale, perhaps the bot could leave those cases (if they can be detected) as is and track them in some special maintenance category for humans to check. This would mean all those cases would get detected, but it would avoid the epic brouhaha a removal might cause entirely. -- Toshio Yamaguchi (tlk−ctb) 07:58, 20 November 2012 (UTC)
  • All images tagged by this bot need to end up in a human maintenance category; if it could tag those that are "close" name matches to a sub-cat of that, that would be great as that's an easier job to fix, but irregardless a human has to check before the images are removed from offending articles (and then if this makes them orphans, we let the existing processes take care of that). The bot cannot remove images itself, unless the mechanism used is to have image uses that have been reviewed by a human and tagged as failing #10c to have the bot run through and remove those - but again, this has to happen after the human check. --MASEM (t) 14:58, 20 November 2012 (UTC)
    • If this bot isn't supposed to remove cases lacking a rationale entirely, then I begin to wonder what exactly the purpose of this bot is. I thought we are discussing a bot that enforces 10c by removing image transclusions from pages where the file page does not contain a rationale and that we just have to make sure images where a rationale is present that closely matches the name of an article where the image is being used are NOT being removed. If that is not the purpose of this bot, then what exactly is it? -- Toshio Yamaguchi (tlk−ctb) 15:12, 20 November 2012 (UTC)
      • If you tag with {{subst:nfurd}}, you get a week to fix the problem. I suppose that the bot could do something similar: tag files with some template which says that the file seems to lack a FUR for article Xyz, and then place them in a dated maintenance category. After a week, a human checks all files in that category, possibly fixes any errors and then tells the bot to solve all remaining WP:NFCC#10c violations by removing images from articles. To avoid abuse, only trusted users should be allowed to tell the bot to remove images from articles. --Stefan2 (talk) 15:22, 20 November 2012 (UTC)
        • So the purpose of the bot we are discussing here would be to replace (or provide an alternative to) Wikipedia:Database reports/Non-free files missing a rationale and rationale missing? -- Toshio Yamaguchi (tlk−ctb) 15:32, 20 November 2012 (UTC)
          • It won't replace it, because we should assume that editors unintentionally create #10c violations all the time (reusing images, bad page moves, etc.) But the bot will formalize a process that initiates a 7 day (minimum) "fix or it will be removed" stipulation on images that fail #10c. The hope is that with the bot messaging the uploader, the file page, and the article where #10c isn't met, someone will fix it within that 7 days, minimizing the "work" needed by the human editors in reviewing the maintenance category; though editors may be able to fix a few more and only a few exceptional cases fall through the cracks. A bot helps to maintain this ongoing tasks, and with the notification of end users affected by it. --MASEM (t) 15:38, 20 November 2012 (UTC)
            • Okay, this pressure the bot would be generating seems to be exactly what we need without militarizing the processes, so I am supportive of implementing it that way. -- Toshio Yamaguchi (tlk−ctb) 21:28, 20 November 2012 (UTC)

RfC: Can cover art can be used within articles on authors/bands?

Is footnote #1 attached to WP:NFCI #1 a correct interpretation of WP policy? That footnote reads:

NFCI#1 relates to the use of cover art within articles whose main subject is the work associated with the cover. Within such articles, the cover art implicitly satisfies the "contextual significance" NFCC criterion (NFCC#8) by virtue of the marketing, branding, and identification information that the cover conveys. The same rationale does not usually apply when the work is described in other articles, such as articles about the author or musician; in such articles, the NFCC criteria typically require that the cover art itself be significantly discussed within the article

Background: This RfC is a follow-on to an RfC from 2 months ago, which asked whether or not NFCI#1 permitted cover art to appear in articles on authors/bands. That RfC was a bit convoluted and the closer of the RfC closed as "no consensus". --Noleander (talk) 19:29, 18 November 2012 (UTC)

Yes - NFCI#1 is misleading: it clearly implies that copyrighted cover art can be used in within articles on authors/bands (provided that those articles critically discuss the music/book). However, the policy enforced for many years throughout all WP articles is the opposite: cover art cannot be so used (the rare exception is if the cover art is especially iconic in the author/band's history). The prior RfC demonstrated that the widespread practice in WP is to prohibit such image usage. There are a few editors that believe that cover art should be permitted in those situations. Personally, I don't care one way or another. But whatever the rule is, NFCI#1 should be improved so editors don't have to guess. --Noleander (talk) 19:29, 18 November 2012 (UTC)

Yes The footnote is correct because the intent is for cover art to only have immediate allowance on articles on the published works they represent; it does not imply this is the only time cover art images can be used but that other uses generally require a strong reason they meet NFCC (eg like a cover that has become an iconic representation for a band backed by sourced commentary). --MASEM (t) 19:59, 18 November 2012 (UTC)

Yes - Masem hits the nail on the head here. Album art's acceptability outside the main (immediate) subject requires a strong reason for NFCC compliance. Der Wohltemperierte Fuchs^(talk) 20:50, 18 November 2012 (UTC)

Request for any 'no' !voters - If you believe that non-iconic cover art is permitted in articles about authors/bands, could you please identify specific WP articles that include such cover art? Also, please include links to RfCs or talk page discussions where the community decided that such usage was okay. Thanks. --Noleander (talk) 16:56, 19 November 2012 (UTC)

Even though I voted yes, I will point to Wikipedia:Featured article candidates/Pink Floyd/archive2 (which I think started the previous discussion) where selective inclusion of specific album covers was discussed at length. Note that per my reasoning above, each of the 3 covers actually used in Pink Floyd have reasonable in-prose explanations for inclusion, but it's not outright allowance to use the covers without other commentary about them as the individual album articles have. I think a few other band articles that are FAs (like Rush and the Beatles) are also examples brought up in that. --MASEM (t) 17:01, 19 November 2012 (UTC)

I agree. Those covers are very famous, and played an important role in the bands' history. Those articles about the bands specifically discuss the album cover (not just the album's music). The whole point of footnote #1 is to help editors distinguish those iconic covers from the run-of-the-mill covers. NFCI#1 (without the footnote) does not address that distinction at all. --Noleander (talk) 17:15, 19 November 2012 (UTC)

Yes The footnote correctly elaborates on the context required to use cover art on articles for which the subject is not the same subject of the artwork. I do not see anything wrong with the footnote as it is shown, but I would be interested to hear from anyone that does believe it is incorrect. --Odie5533 (talk) 15:47, 20 November 2012 (UTC)

In the prior RfC on a similar question there were two editors that felt that cover art should be permitted within articles on authors/bands, for all covers, not just iconic covers, provided that the album/book was discussed in the author/band article. I cannot speak for the other editors, but my recollection is that one of their arguments was the following: There is an article on a band B, and there is a section on that article on an album A. There is no article on album A. That section in B contains the cover art of album A. The cover art of album A is not especially iconic in relation to the band's history. Then, a new subarticle "A" devoted to the album is split-off from the band's article B. That new article A inclues the cover art, per NFCI#1. Query: should that split-off event cause the cover art to be removed from article B? Those two editors felt that removing the cover art from B was not rational, because the existence (or non-existence) of article A should not impact the content of article B. --Noleander (talk) 16:24, 20 November 2012 (UTC)

Non-U.S. works ineligible for copyright in the U.S.

I believe that the below quote from this policy

“ Anything published in other countries and copyrighted there, is copyright in the United States. ”

is misinterpretted in relations to ineligible non-U.S. images, like ineligible logos. This matter has been discussed in village pump and WP:MCQ. How shall we reword it? --George Ho (talk) 20:54, 28 November 2012 (UTC)