Wikipedia talk:Non-free content criteria compliance

A start

This proposal is too much for any single person or small group to develop and maintain. Please help out by editing the proposal and discussing here. Thanks. Carcharoth (talk) 14:17, 17 February 2008 (UTC)

Statistical analysis

This section is for discussion of the Wikipedia:Non-free content criteria compliance#Weekly uploads and deletions and bot taggings section and its graphs. Criticisms and suggestions for improvements welcomed, along with more data please! Carcharoth (talk) 01:26, 18 February 2008 (UTC)

• Image:Wikipedia Uploads.png is relevant here (from User:Dragons flight/Log analysis). My graphs are effectively focusing on the far left of that graph, and extending it. The running total for the net increase over the year is unique to my analysis, I think, though this graph is easily repeated with other sets of data. Carcharoth (talk) 01:45, 18 February 2008 (UTC)

Everything looks spot on to me. The only comments I have is that the big deletion spike in week 48 was indeed caused by MiszaBot deleting 20,000 orphans, and regarding free images - STBotI and ImageTaggingBot catch the freshly uploaded ones without any sources, which are then deleted by MiszaBot in a week. I'm satisfied with the amount of emphasis provided on newly uploaded images claiming to be free.

I fear though that all this hard work of yours will go to waste, as the bots' compliance tasks will last only for another month, and I sense a general apathy among both sides of the fence of the NFCC debates. The editors who stay away from images are mostly too fed up to change the policy, and the image enforcers have far too much momentum to care. east.718 at 01:59, February 18, 2008

East has a point, the relative failure of WP:TODAY shows that most users don't want to involve themselves in this rather tedious process. Since continuing compliance will be an issue for the new images uploaded each day, I really think the Upload interface for non-free image licenses needs to be changed. Mainly it needs to replace the generic sumary box, with mandatory form fields that map to the generic FUR template. If a user can't provide the source, description, and intended use, then it shouldn't let them upload it. And yes, I realize they may mess up the exact article name, but we have automated scripts that will fix a broken fur for 1 article use. MBisanz ^talk 02:28, 18 February 2008 (UTC)

What exactly do you mean that the bots will only last for another month? I know there's some sort of deadline, but do we block uploads then? What happens to make these bots not necessary? --uǝʌǝsʎʇɹnoɟʇs(st47) 13:06, 18 February 2008 (UTC)

What they mean is that much of the furore created by BetacommandBot is the tagging of images from pre-2006 (I think). These are legacy images. It was agreed that a different, more conciliatory, approach was needed there, and so Betacommand held off until 2008. The other plans came to nothing (for a variety of reasons) and so Betacommand started tagging the legacy images. That created the furore over the past few weeks. Anyway, once it is only new images being tagged, there should be less of a problem. But I still think changes should be made in case there is ever a need to bring all our image into line with some future change in the compliance requirements. ie. Don't let Betacommand or his bot near any future compliance enforcement until things are much better organised. Carcharoth (talk) 14:22, 18 February 2008 (UTC)

I've done an analysis of image deletion reasons - User:Hut 8.5/image deletion stats. I can add new search strings easily. Hut 8.5 11:57, 19 February 2008 (UTC)

Thanks! I've added that and two new pages (well, only discovered recently) to the page. User:BetacommandBot/Non-Free Template Useage and User:BetacommandBot/Free Template Useage. Carcharoth (talk) 16:08, 19 February 2008 (UTC)

Trends

I'm noticing that image uploaded per month, declined by about 40% over 2007. I know the usual reason for fewer articles is the no-IP creations, but that was 2006. Any other ideas why fewer images are being contributed? This might dovetail with dragon flight's analysis. MBisanz ^talk 18:07, 18 February 2008 (UTC)

People are realising they should really try and find free images instead of uploading non-free ones? Hopefully. Given the amount of upheaval with non-free images, I don't think image trends can be reliably linked to overall Wikipedia trends, but if you can make a definite link, by all means try. Carcharoth (talk) 16:09, 19 February 2008 (UTC)

Handling rationales post-March 23

Obviously there's some good ideas for making bots work better for rationales here, but we probably need to decide if there's a better way to track rationales after March 23, 2008.

It would be great if we had a data store that for each image, and each article used in that image, both human editors and bots can checkmark a image-page use pair as having a compliant rationale. Then, regularly, a bot (not necessarily BCB) would run through every image and:

1. If the image is used on a page and had a rationale per datastore, and to the bot the rationale still appears good, the bot does nothing.
2. If the image had a rationale for the page, but the rationale appears broken, it tags it and drops it into one special category bin for human review "broken rationale review".
3. If the image is newly used on a page, but has a rationale, tag it and drop it into a "rationale review" (noting the page, if possible, of the new rationale)
4. If the image is newly used on a page, but lacks a rationale, tag it and drop it into "image lacking at least one rationale, but otherwise still valid" (image is not threated to be deleted, just the use on the offending page)

There's likely other cases I don't have listed here, but the key thing is that an existing image will either have nothing done into it (beyond maybe a last date checked) and any other image will be put into exactly one category above for human correction. The timing of how this bot operations probably means this would need to be done once a month for all images, but assuming BCBot finishes its job soon, the volume of images that fall into the categories above should be very small to make it a managable workload for human editors to review.

New images are processed near-immediately after upload by a bot; ones with good rationales are placed into "new image requiring rationale review", bad ones are tagged similarly. If the count of 500-1000 images a day that are being uploaded is correct, then we likely need to have a good dedicated base to this aspect here.

The only problem is that datastore: there's no easy MediaWiki mechanism to create such without spamming a lot more pages, and then this makes it prone to malicious users tricking the system. (But of course, all parts of WP are like that, so maybe we deal with it). --MASEM 17:02, 19 February 2008 (UTC)

Two comments. One, many bots have 'datastores', also called databases - the toolserver allows each user to have their own.

Second, new images are already processed near-immediately, and I haven't seen any reasons to change that except that we want rationales to be reviewed by people. Perhaps we can ask the devs to consider something like flagged revisions/patrolled edits for the image namespace? --uǝʌǝsʎʇɹnoɟʇs(st47) 19:18, 19 February 2008 (UTC)

I think that image patrol is a good idea. The bots tagging the new image uploads do a good and necessary job, but it would be good to have images checked by people also. Bláthnaid 12:03, 24 February 2008 (UTC)

Most defenently, bots are nessesary but they can basicaly only check if one of a few required components are missing from an image page. The only way to check if a provided rationale is even remotely valid is for a human to review it. For example as long as the article name is provided Beta's bot for example will not see any problem with an image who's purpose of use is stated to be: "4u89a34ioøjfd" or "because I like it" or whatever. So there is no way we can rely on bots to verify that an image complies with the policy, they are usefull for clearing out low hanging fruit (images with no rationale altogeter etc), but what we rely need is people, and lots of them, to do image patroling (not not just for non-free images either, there are tonnes of images claimed to be free licensed that are clearly not) --Sherool (talk) 12:36, 24 February 2008 (UTC)

Yes. That's why the page says "HUMAN + BOT" under 10c. :-) You are right about the need to keep all images under scrutiny, but I've avoided free images because all the free image checks have to be carried out by humans (apart from the magic 'source' bots that reduce the flow of new images). Oh, and the number of free images on Wikipedia is far greater than the number of non-free images, though that could be tackled by increasing the flow towards Commons. Carcharoth (talk) 15:34, 26 February 2008 (UTC)

"It is estimated that the number of images uploaded with a non-free tag accounts for about a quarter of the uploads."

I must confess that I haven't been able to digest all of this page, but I encountered that estimate and wanted to know how it was made. Was it based on a sample of uploads in real time, or some other method? --Anticipation of a New Lover's Arrival, The 00:13, 27 February 2008 (UTC)

Hmm. Can't remember. Try clicking special:random/image 100 times and see what you get. You could also collect data from special:log/upload. I think it might have been based on an offhand comment by Betacommand that there are "about 500 non-free image uploads per day" (or something), and I approximated the number of total uploads a day as 2000 (from the data on the main page), and hence arrived at the wildly inacccurate 25% figure. So yes, please do try and find more data that can be reliably interpreted. :-) Carcharoth (talk) 00:42, 27 February 2008 (UTC)

Yes, the wording is ambiguous, I suppose. If you click on special:randomimage you're favoring retained images (which is skewed by all the different factors influencing early deletion), whereas if you sample in realtime you're looking directly at the uploaders' behavior. It would be nice to get some better data, and I suspect the raw rate is probably closer to 50% non-free if you count images that aren't tagged at all (which I'm inclined to do, because we can't treat an image as free unless we know something about it). --Anticipation of a New Lover's Arrival, The 00:48, 27 February 2008 (UTC)

I love the random image link! I clicked and ended up at Image:NY Euphoria celebrates @ LB III.jpg. :-) But the wierd thing is, that is precisely the sort of image BetacommandBot should have tagged. Can't see why it hasn't done so yet. Carcharoth (talk) 00:57, 27 February 2008 (UTC)

According to Betacommand, some images might have been missed previously, and s/he is doing smaller runs to catch the missed ones. Bláthnaid 09:34, 27 February 2008 (UTC)

Question: Is there any way of knowing how much of the free images uploaded to Commons are for use in Wikipedia? A lot of Wikipedia editors upload their free images onto Commons instead of Wikipedia (I do). Those images should be taken into account when calculating the amount of free vs non-free images used on Wikipedia. Apologies if this has been answered elsewhere, but WP:AN/B is so long that I haven't had time to read it all yet. Bláthnaid 09:56, 27 February 2008 (UTC)

For our purposes (controlling non-free uploads to English Wikipedia) it doesn't matter how many free images are also uploaded, or to which of the two projects that can host them. The estimate was, I think, simply a way of relating the rate of non-free image upload to the raw upload graphs.

Although of course (and I'd forgotten) free images should always go to commons in preference to any single-language project. --Anticipation of a New Lover's Arrival, The 10:11, 27 February 2008 (UTC)

That's what I was thinking. Ideally, all image uploads to the English Wikipedia should be of non-free images :p. It would be nice to know how many free images from Commons are used on the English Wikipedia, since these images are a victory for the free-content mission. Bláthnaid 10:19, 27 February 2008 (UTC)

Sample

For what it's worth here's what I got from a (tiny) realtime sample:

• Image:Habenaria lacera 300x400.jpg

  no copyright tag at all, described only as "self-made", already tagged by a patroller

• Image:Sentimiento.jpg

  no information at all, obviously commercial, already tagged by a patroller

• Image:Terrorbird.jpeg

  tagged as non-free album cover, fair-use case written

• Image:Oceansat1LACoast.JPG

  tagged as "freely available for noncommercial use" by uploader and alread scheduled for deletion

• Image:Nemesisprime-alternators.jpg

  tagged "Non-free 3D art", with fair-use rationale

• Image:Mail Call Aboard Du Pont.jpg

  tagged as "pd-self"

• Image:J dilla Donuts.jpg

  tagged "Somewebsite", and dated tag "No license needing editor assistance"

• Image:Cooperclipsal.JPG

  tagged "pd-self"

• Image:Desceliers 1550 map detail.jpg

  deleted before I could examine it. Probably by the uploader.

--Anticipation of a New Lover's Arrival, The 00:38, 27 February 2008 (UTC)

Stats from upload log

I ran a script parsing the upload log (which only has data for the past week, including incomplete data for today and for the 21st, so those aren't included) and got these results. Note that non-free images were determined by what has a non-free license template on it and free images are those that have a PD, CC, or GFDL license template on it. So, I wouldn't say these numbers are spot-on, but they are a good starting point (given the information we have to play with).

	Uploaded Images
Date	Non-free	Free	Patrolled and Tagged	Tagged for Deletion	Deleted	Total
26 February 2008	576	485	191	276	138	1872
25 February 2008	536	562	215	167	173	1918
24 February 2008	600	558	242	186	185	2011
23 February 2008	531	590	172	167	184	1860
22 February 2008	525	480	203	131	204	1754

I would say the 25% estimate is not wildly inaccurate. - AWeenieMan (talk) 06:17, 27 February 2008 (UTC)

I don't understand those figures. If for instance on 26 February 576 images were non-free, 485 free and 138 deleted, that makes 1199 images. The total you give is 1872. What of the remaining 673 images? Are they simply untagged images awaiting tagging or deletion?

In any case, it seems more accurate to summarise the above data as "of images that are tagged, more than half have non-free tags." --Anticipation of a New Lover's Arrival, The 06:25, 27 February 2008 (UTC)

I added a column for images with a speedy or pseudo-speedy deletion template on them (of course this classification and classification based on license tag are not mutually exclusive). I am going to dump the unclassified images to a file and see what's going on, but a quick look shows many images with either no template or just an {{Information}} template and no license. Of course the other option is that I have missed a template that doesn't follow the standard license template rules. - AWeenieMan (talk) 06:42, 27 February 2008 (UTC)

A typo might mean the tagged column is off, give me a minute to rerun. All fixed now. - AWeenieMan (talk) 06:46, 27 February 2008 (UTC)

Thanks. I'm still seeing over 50% of tagged images as non-free, and we can probably assume that many if not most of the untagged ones are just some picture somebody grabbed from a website. --Anticipation of a New Lover's Arrival, The 07:03, 27 February 2008 (UTC)

(undent) I would say that is an accurate summation. I have updated the stats to take into account some of the things I saw on all the unclassified images. Practically all of these images fall into one of two categories. The first being images that were patrolled and tagged with some kind of deletion warning tag (but not a real deletion tag) such as {{imagevio}} or {{No license needing editor assistance}}. The other are images that include no template at all or just an {{Information}} template (usually incomplete). - AWeenieMan (talk) 07:21, 27 February 2008 (UTC)

BetacommandBot and other bots

I've changed some of the wording which seemed to focus (without explanation) on BetacommandBot. Some explanation about these bots should be included: at least the criteria used for tagging, and whether the criteria have changed over time. --Anticipation of a New Lover's Arrival, The 00:24, 27 February 2008 (UTC)

Yes. If you have the time to do the research and find the diffs, please do. This is very much a work in progress, and I'd love more people to join in and help. Carcharoth (talk) 00:48, 27 February 2008 (UTC)

"We cannot tell which of our images are compliant or not"

I don't understand what is meant by this:

The problem we have is the following: "By March 23, 2008, all existing files under an unacceptable license as per the above must either be accepted under an EDP, or shall be deleted." - the way the EDP is worded on en-Wikipedia, we cannot tell which of our images are compliant or not

I've read the EDP and it seems pretty uncompromising. Can someone give examples of media about which we cannot say that either it qualifies under the exceptions or it does not? Surely if we lack the required data to make an exception, the media cannot qualify. --Anticipation of a New Lover's Arrival, The 02:14, 27 February 2008 (UTC)

The main problems are that some of the criteria are subjective (there is no easy yes/no "data"), and there is disagreement over how to interpret some of the criteria. The one for which disagreement has been most heated, in my opinion, in NFCC #8 (significance for the article). An extreme interpretation of this sees most non-free images go, including album covers, book covers, logos, and pictures of people (unless the article discusses the person's appearance in some detail). A less strict interpretation of NFCC#8 sees most of the aforementioned examples kept. NFCC #1, #2 and #3 have also seen heated debate in the past. When disagreements arise, debate takes place in places like the image talk page, an 'Images for discussion' page (IfD), or at a policy talk page (such as Wikipedia talk:Non-free content - WT:NFC). In other words, the subjectivity of the wording (unavoidable) means that case-by-case discussions are inevitable. Would you like some examples of debates? Carcharoth (talk) 06:32, 27 February 2008 (UTC)

The year set aside for debates is nearing an end.

The E in "EDP" means "exception" and reading the resolution I see:

Such EDPs must be minimal. Their use, with limited exception, should be to illustrate historically significant events, to include identifying protected works such as logos, or to complement (within narrow limits) articles about copyrighted contemporary works.

"Minimal" and "within narrow limits" are pretty unequivocal. Moreover the non-free stuff comprises for the most part fripperies such as album covers and television screen shots. These are of little (I almost said "no") value to the free encyclopedia. There are defensible exceptions--I have often cited certain album covers, the likes of, Led Zeppelin IV, Close to the Edge and Brain Salad Surgery as good examples of significant album cover artwork that merits coverage in the article (the Led Zeppelin symbols, Yes's Roger Dean artwork, and the ELP's H.R. Giger artwork) but those cases can easily be argued at deletion review.

In any case we've had our twelve months and obviously a few of the problems pertaining a year ago have been addressed by the uploaders, but most have not. We can delete them unless we have a clear idea to whom the original copyright belongs, reasonable confidence of the provenance of the copy--that will account for most of the non-free media we have. They can be replaced, if necessary, by properly sourced copies with good fair-use cases. --Anticipation of a New Lover's Arrival, The 06:49, 27 February 2008 (UTC)

I agree, actually, but there will be a lot of resentment that this wasn't clearly stated. Like it or not, people will have got the impression that the BetacommandBot and 10c business meant that the images were "OK". Obviously I don't agree with that. BetacommandBot was only ever reviewing a small subset of non-compliant images, and I've said for ages that the fuss about it was only putting off the day when humans would need to properly review the use of each non-free image. Which is why I wrote this document, to get bots to help humans with the size of the task. Allow humans to tick reviewed images, and bots and categories to show which images haven't been independently reviewed by a human other than the uploader or user. Carcharoth (talk) 07:10, 27 February 2008 (UTC)

Yes, thanks for a good job of organised thinking about that. I must confess that, having nothing to do with images from one year to the next, I had forgotten all about the March, 2008 deadline. --Anticipation of a New Lover's Arrival, The 07:12, 27 February 2008 (UTC)

Compliance priorities

The compliance criteria are subdivided by this document as follows:

• 1 - WP:NFCC#1 - replaceable (must be replaced if possible)
  • HUMAN (case-by-case discussion)
• 2 - WP:NFCC#2 - commercial opportunities (our use mustn't compete with commercial use)
  • HUMAN (case-by-case discussion)
• 3
  • 3a - WP:NFCC#3a - minimal use (low resolutions and reduce non-free use per article and in Wikipedia as a whole)
    • HUMAN + BOT (algorithm and generally agreed principles)
  • 3b - WP:NFCC#3b - minimal extent of use (don't overuse individual non-free images)
    • HUMAN + BOT (algorithm and generally agreed principles)
• 4 - WP:NFCC#4 - previous publication (must have been previously published)
  • HUMAN (simple check of the details provided)
• 5 - WP:NFCC#5 - content (generally encyclopedic)
  • HUMAN (simple check against examples of non-encyclopedic content)
• 6 - WP:NFCC#6 - image policy (complies with Wikipedia:Image use policy)
  • HUMAN (simple check against examples of non-NFC violations of image policy)
• 7 - WP:NFCC#7 - not orphaned (all non-free images used in at least one article)
  • BOT (purely mechanical check - humans needed to guard against vandalism)
• 8 - WP:NFCC#8 - significance (non-free images must contribute significantly to the article)
  • HUMAN (case-by-case discussion)
• 9 - WP:NFCC#9 - location (with limited exceptions, non-free images only used in article namespace)
  • BOT (purely mechanical check)
• 10
  • 10a - WP:NFCC#10a - source (attribute the original source and copyright owner and intermediate sources)
    • HUMAN (simple check of the details provided)
  • 10b - WP:NFCC#10b - copyright tag (use one of the copyright tags suggested when uploading)
    • BOT (purely mechanical check against approved list of copyright tags)
  • 10c - WP:NFCC#10c - rationale (explain the reason for using the non-free image and name the article it is used in)
    • HUMAN + BOT (algorithm detects existence of rationale and article name and overuse of images outside articles mentioned in rationales - humans check whether the rationale or overuse is valid or not, remembering that some of the components of a good rationale directly or indirectly address other aspect of the NFCC, specifically (eg. in Template:Non-free use rationale) source [#10a], description [#5 and #6], portion [#2 and #3a], resolution, [#2 and #3a], purpose of use [#5 and #8] and replaceability [#1])

This makes the problem look much bigger than it actually is. The vast majority of our non-compliance problems, after removing media tagged but without any semblance of a fair-use case, will be those for whom the copyright owner is not known (10a). This will be most existing non-free media. Once those have been deleted, then we can look at the remaining media and consider how big the task is. --Anticipation of a New Lover's Arrival, The 02:43, 27 February 2008 (UTC)

"without any semblance of a fair-use case" as a criteria for speedy deletion will be bitterly disputed, I fear. I do agree that 10a is the largest problem, and the page should be edited to give some idea of the scale of the problem for each criteria. But this doesn't mean that the images don't need to pass the other criteria as well. You are right, though, in that the most efficient approach would be to go through the non-free images and remove those without sources (though it is quite fun finding the sources, and in many cases it is obvious). Please note the difference between copyright owner, proximate source and ultimate source. The proximate source will be "the website I got the image from" or "the book I scanned it from". The ultimate source will usually (not always) be the copyright owner, and in some cases is harder to identify. In many cases though, the ultimate source (the copyright owner) can be identified as "the source" without needing to say which website or book it came from. Though some argue that "source" is needed for verification as well as copyright considerations (ie. making sure it is not a fake image). We also need to distinguish between source "not provided" and "not known". Some images (usually older ones) are genuinely of unknown source, or have an uncertain copyright status, but enough is known to make a reasonable fair use case, or even, sometimes, to argue that the image is now public domain (ie. no copyright renewal was performed, or the photographer is unknown or not names). In many countries, not naming the photographer on publication means that an image falls into the public domain relatively early. I'll try and dig up a few examples of images and templates on that kind of thing. Carcharoth (talk) 06:43, 27 February 2008 (UTC)

I think the important thing to recognise is that this is only a huge job if we decide to let it be. The instructions we have from the Foundation are clear, and require mass deletion. The EDP is only there to safeguard works important to the encyclopedia, which are very few and far between. We could delete almost every television and film screen shot without lessening the value of the encyclopedia by one single farthing. We shouldn't set ourselves up with he task of trying to save every single non-free item. They don't belong to us and we have no intrinsic right to use them. --Anticipation of a New Lover's Arrival, The 06:57, 27 February 2008 (UTC)

Yes, but please don't let historical images get caught up in that. Many have no named copyright holder and are incredibly unlikely to even have existing claims from descendents of whoever took the photo. In many cases, we simply don't know who took the photo, and it is likely that no-one will ever know or care. These are just about the safest examples of fair use (in terms of anyone being around to actually dispute the fair use) that you can think of. Carcharoth (talk) 07:39, 27 February 2008 (UTC)

Yep, historically important stuff must be kept. Now I think of it, I'd be surprised if we don't have a tag for that. --Anticipation of a New Lover's Arrival, The 07:56, 27 February 2008 (UTC)

Template:Non-free historic image. Carcharoth (talk) 08:02, 27 February 2008 (UTC)

Appropriate mechanism

The resolution is worded strongly, and should be taken in the spirit in which it is intended. This makes our choice of mechanism easy.

1. Non-free media may be deleted by any administrator who cannot positively confirm, from the media's wiki page, that the media complies with all provisions of the Exemption Doctrine Policy.

2. At deletion review, the media may only be undeleted if there is consensus that compliance has been achieved.

This is simple and allows for the mass deletions we will have to perform in order to comply with the Foundation's resolution. It also minimises the work and places the onus where it belongs: on the person who uploads an image or wants it undeleted. --Anticipation of a New Lover's Arrival, The 03:15, 27 February 2008 (UTC)

Taking this to WP:CSD. --Anticipation of a New Lover's Arrival, The 04:30, 27 February 2008 (UTC)

I'll come over there, then, but I hope you will find time to reply to what I wrote above in the other sections. Adminsistrators will disagree on the interpretation of the NFCC. Taking this to deletion review will simply overwhelm DRV. The correct place to have such debates, in my opinion, is IfD. There are those who specialise in images and copyright of images, who can often demonstrate that an image is public domain. And one of the most sunjective criteria, which I forgot to mention, is whether a rationale is "valid". ie. Are the reasons provided good enough? "The rationale is presented in clear, plain language, and is relevant to each use." - that is, again, easily interpreted strictly, or less so. It is effectively NFCC#8 placed inside NFCC#10c. In my opinion, it should only apply to whether there is a logical link between the rationale and the article. ie. If the rationale actually refers to the article. If there is disagreement over how relevant the rationale is, that is really an NFCC#8 argument, and should go to IfD or another discussion forum. Carcharoth (talk) 06:51, 27 February 2008 (UTC)

We can't adopt the IfD approach. There are thousands upon thousands of non-free images to be deleted, and the uploaders have had their year's grace. --Anticipation of a New Lover's Arrival, The 06:59, 27 February 2008 (UTC)

Actually, if you took part in the WT:NFC debates, you would have known that part of the reason for developing the rationale and copyright tags the way they were, was to allow easily identifiable categories of images to be debated as a whole, and to see where the bulk of the images are. have a look at User:BetacommandBot/Non-Free Template Useage. Instead of arguing at deletion review (or IfD) over each image, it would be more productive to have large debates over classes of images (eg. album covers, magazine covers, book covers, logos), and decide where the line is drawn in those cases. The ones I'm most interested in are in Category:Non-free historic images and those marked with Template:Non-free unsure, all of which need individual discussion. Many of the images are either public domain (or will be soon), or have clear educational use in articles on historical topics. Also, I've yet to see people address the issue of cases where research has yielded good evidence of provenance and a determination of when the image becomes public domain, but then this all gets deleted. We should be keeping that research and flagging the image for undeletion when it falls into the public domain. This probing of public domain issues directly helps the free content mission. It is often said, with some justification, that 99% of the media produced before a certain date (can't remember the exact date, maybe 1940?) and currently considered copyrighted is actually public domain - due mainly to non-renewal of copyrights. Those who care about the free content mission should be actively pursuing this, as well as stemming the tide of the uploading of contemporary, popular culture, images (ie. album covers and screenshots), but because it is easier to do the latter, the former gets neglected. What I object to is people lumping the latter (older, uncertain, probably PD images) in with the former (clearly identified, modern imagery). Your approach would do nothing to solve that problem. Carcharoth (talk) 07:24, 27 February 2008 (UTC)

Here is an example of well-defined historical photos: Category:Denver Public Library images, and Category:Otto Perry images and Category:Robert Richardson images. See Image:Ashcroft hotel.jpg for one of those tagged by BetacommandBot. I would hope that someone would get round to fixing that eventually, but how would you ensure (in your proposed system) that images don't get tagged, reviewed and deleted by those who misunderstand what those categories are about? A clear distinction can (and should, in my opinion) be drawn between older and newer images. Carcharoth (talk) 07:33, 27 February 2008 (UTC)

A problem with fixing the historic images is that it can be difficult to find the source and/or copyright holder for the images. I came across some such images on User talk:Hayford Peirce yesterday. It's a real shame to lose those. That retired editor has also uploaded a lot of images of mid-20th century books. Logos, book (especially first edition) and album covers can also be old images, but I don't know of a way to make a distinction between them and, say, an album released last week. Bláthnaid 09:46, 27 February 2008 (UTC)

Retired? Are you sure? Sometimes a Google search can provide image and source information, though you have to be sure they didn't get the image from us. Carcharoth (talk) 10:46, 27 February 2008 (UTC)

Oops! How did I miss the 2008 edits? Excuse: I haven't had my morning coffee yet :p I'll leave a message to him about the historic images, and hopefully they can be saved. I was able to fix the book covers, because he said that he scanned them and gave the copyright information. Bláthnaid 10:53, 27 February 2008 (UTC)

Previous Publication - Criteria 4

Virtually all magazine covers, book covers and CD/DVD covers have been published. The cover that was not published would most likely be notable on it own. We need to allow the existing covers by default or automatically mark them. We don't need to start another wiki-wide commotion over a technical detail. -- SWTPC6800 (talk) 04:41, 10 March 2008 (UTC)

Well, no, but there should be an entry in the image datastore that says "previous publication not yet explicitly noted". And we should be requiring people to say where something has been previously published. It is especially important to try and find out the date of first publication, because (in many cases of images of pseudonymous or unknown authorship) this date often determines when the image falls out of copyright. For example, when deciding what book cover to use on an article about a book, it often makes sense to select the cover of the first edition because that is (a) notable and (b) most likely the first image about the book to fall out of copyright and into the public domain. Carcharoth (talk) 01:00, 12 March 2008 (UTC)

New bot in BRFA for image reduction

There is a bot request for approval at Wikipedia:Bots/Requests for approval/ImageResizeBot‎. This is relevant to non-free image compilence, so I thought I'd spam here. Comment there :). —— Eagle101>^{Need help?} 19:06, 17 March 2008 (UTC)

Tag conflict

So, the great cross-tabbing is done and there is a list of 700 images at User:Betacommand/Sandbox 3 that show images classed as both free AND non-free. Obviously, an image can only be one of the two, so if editors could go through and correct the images, striking them out on the master list. Thanks. MBisanz ^talk 02:57, 18 April 2008 (UTC)

Unambiguous rationales

We all know that criterion 10c causes a considerable amount of angst. Part of it, I'd say, is that it doesn't follow common sense. People select a copyright tag that has an extensive justification of why they can use the image, and then they are asked for a "rationale" for "each article".

For the purpose of this discussion, I'm going to capitalize "Rationale" to evoke legalese when I'm referring to the technical NFCC definition, and leave it uncapitalized when I'm writing from the perspective of a user who's unfamiliar with NFCC.

So someone who's not well-versed in NFCC theory might upload an image, check over what they've done, and say "Yep, I've added one rationale for why I'm using it on one article". Then somebody or somebot comes along and notifies them that they haven't written a Rationale for each article. Confusion and accusations ensue. The problem, of course, is that copyright tags apparently contain rationales for why the image is being used under fair use, but these rationales are not generally Rationales because they don't meet the technical definition.

It's a very common confusion, and I think we should be doing what we can to fix it. The constructive way I see to fix it is to automatically add Rationales for images have only been used on one article, and have a proper and sufficiently informative copyright tag, but have never had a Rationale. Maybe drop the uploader a note saying that we've added a Rationale for them and explaining to them how to do it next time.

This is probably about the point where someone taking a hard-line stance on enforcement runs in screaming "BOTS CAN'T WRITE RATIONALES". This is the assumption I am questioning. There are many, many situations where the copyright tag gives us enough information to write a Rationale automatically. The objection I've heard is that a bot can't tell what article it was supposed to be for. But just because certain cases of a problem are unsolvable doesn't mean the problem is always unsolvable. In the situation where an image has only ever been used on one article, there is only one article it can be for.

So, what are people's thoughts? Can bots write Rationales in this unambiguous case? rspeer / ɹəədsɹ 18:14, 20 April 2008 (UTC)

If people want to mass write rationales with AWB, I can live with that. However, I personally have headaches looking through and finding movie posters and other promotional items tagged with {{logo fur}} from I can't tell if it was a bot or just laziness. It just becomes more and more complicated to both find and to fix problems once large rationales are put in place. I've always wondered why not have the bot dump a list of the images that are most concerning and let people tag them for deletion themselves. Let people review it on the front end. I'm working through a subset of the problem at Category:Images lacking a description here and can list about 50 or so a day (while doing tons of other stuff). If his runs are producing smaller amounts, getting groups of people to attack them and fix those problems would be better. Having numerous people be the listers should also alleviate some of the flooding that one person would deal with. It is only about 1 in 100 uploaders who really drive you up the wall. -- Ricky81682 (talk) 09:56, 22 April 2008 (UTC)

The image upload form now gives you a rationale template to fill in. When the uploader understands template syntax, this works as a way to add the rationale. I'd be willing to believe that there are uploaders who misunderstand what the tag is for, but then fill in the template the correct way for that incorrect tag. There will always be people mis-tagging images, so I'd consider this a separate problem. rspeer / ɹəədsɹ 19:15, 24 April 2008 (UTC)

A template

I'm going to begin work on a machine readable template to show that an image complies with WP:NFCC, just to get this back up to speed. Suggestions? ViperSnake151 18:14, 13 June 2008 (UTC)

Great idea! Sorry I missed this. Any updates? Carcharoth (talk) 13:59, 10 July 2008 (UTC)

Templates - older versus newer

The labelling of the templates - older and newer - was the wrong way round. The {{Non-free use rationale logo}} was the first of its kind, and created by Wikidemon on 8 July 2007‎. Prior to this templates didn't contain any information already built-in. The {{Album rationale}} was created by me on 3 September 2007, and was the first of this particular style. PhilKnight (talk) 17:46, 3 July 2012 (UTC)