A problem with basic license process
- first posted to the Village Pump
We have a problem with image licensing that has gotten slowly worse over time.
- 1a) We make it hard for established users to upload photos they didn't take themselves, unless they lie about the provenance of the photo.
- 1b) We make it easy for anyone to upload any photo as long as they lie about its provenance. regardless of whether we have any reason to trust that user or not.
- 2) We allow Flickr-washing: accepting bot-added photos from any random account on a third-party photosharing site if the user slapped a free-license tag onto their account, regardless of whether we have a way of contacting that user or not.
- 3a) We make it easy to delete in-use images, for all sorts of technicalities, if an active Commonsist decides that the original uploader should do more to prove the original copyright release. (with only a 1 week required wait for a response - if the uploader is hard to track down, the image gets deleted)
- 3b) We make it hard to request a takedown of out-of-use images, by the creator of subject of the image, even if the source of the original upload is hard to track down or verify. (if details are hard to track down, the takedown request is denied)
Current practice tends to err on the side of frustrating active users, would-be contributors, and the creators or subjects of photos who are offended that we are keeping their images without their permission. There is no single direction to this set of practices: allowing Flickr-washing makes it harder to trust the free license of a randomly-selected photo. denying takedown requests by default tries to maximize the number of files included in our "Ark of Free Media". deleting uploads by long-time users who are uploading on behalf of someone else tries to reduce the number of files in our collection, while trying to make it easier to trust the free license of a randomly-selected photo.
It might be more accurate to say we have something like "bronze", "silver", and "gold" levels of assurance that a file is freely copyrighted.
- "The image was scraped from a random external photosharing account, or uploaded by a newbie user with no real identity, claiming it's available under a free license. there is no special reason to feel confident the uploader was the creator, and no way at all to get in touch with the original uploader."
- "A well-known user uploaded the image, in consultation with the creator/photographer. There is an explanation in the description/metadata about the license it was released under. The uploader got permission from or was asked to upload by the creator, after explaining free licenses; they may have been sitting next to one another when uploading. However it may be hard or awkward to get back in touch with them in the future. (it is generally hard for them to make an account and upload themselves; too many hoops to jump through.)"
- "A known user claiming to be the creator uploaded the image, or it was taken from an external photosharing account or website of a known person who maintains that account for long stretches of time."
We currently allow Bronze and Gold media, but delete Silver media. Even though, as far as I can tell, Silver media are more likely to actually be intended for distribution under a free license.
We delete Bronze or Silver media on request by any active Commons user, but refuse to takedown Bronze media on request of the creator/subject unless they can prove their identity.
- In my experience, what you'd term as "silver" uploads are really quite rare. Most "I have permission" uploads are not so likely to be intended for distribution under a free license, but are actually just intended for use on Wikipedia. Most people, including otherwise "well-known" users, simply don't know how to ask for a licensing permission in a way that explains to the would-be licensor what it is we actually need from them. In many cases, they even ask the wrong person (e.g. the subject of a photo rather than its copyright holder). Users who do know enough to identify the real copyright holder, to explain the licensing to the copyright holder, and to document their permission also typically know our procedures well enough to send that documentation in via e-mail. —LX (talk, contribs) 11:07, 17 February 2013 (UTC)
It should be known, that this discussion is as a result of File:Paul Myners.jpg and File:JCA3 edited-1.jpg, and subsequently these notes which were left for JW. With File:Paul Myners.jpg, the EXIF data clearly states:
“LICENSE AGREEMENT This Photograph taken by VisualMedia is supplied with an indefinite license for editorial purposes, but excluding advertorials or competitions. The license also covers internal communications requirements such as newsletters and non-commercial website use. Usage for external marketing / advertising purposes will attract additional fees that need to be negotiated dependent on requirements.
For further information please contact VisualMedia on +44 (0)20 7613 2555.Note to Press: These images are supplied free of charge for editorial usage. Mandatory credit: VisMedia”
It clearly states that it is only for non-commercial usage, and only for website usage. This is exactly why COM:OTRS was brought about. That others refuse to see the bigger picture that has been clearly explained on both talk pages and via OTRS is the problem as I, and undoubtedly others, see it. russavia (talk) 19:38, 17 February 2013 (UTC)
- Thank you, SJ: a lot of good points here. I think your proposal is a good one, if we remove the "silver" stage which is hardly ever useful (as noted by LX): too much work and clutter to replace the rather simple current practice by introducing a backdoor/special process. Refining the definitions of the two levels a bit would be needed, of course.
- It's probably not worth throwing such a classification on users, but it could be used internally as a guideline, to remember ourselves to be consistent, instead of just pretending that bronze is the same as gold, or that gold is diamonds: perfection can't be achieved; even when we have super-formal contracts as with Bundesarchiv, which would be "gold", errors are possible; we must assume there is none, but someone may complain and force us to revisit the statements. --Nemo 23:31, 17 February 2013 (UTC)
- The # of colors is arbitrary, and should probably change based on what is common practice. But I think in general a better tagging and quarantine scheme is needed rather than deletion for all uploads that involve an established user who is in touch with the media creator / copyright holder. (An "established" user doesn't mean "someone we like" but "someone with time sunk into their user account, which limits spoofing".) In those cases, we are likely closer to getting a proper free license than we are for the average anonymously-uploaded image with minimal metadata and no verifiable source.
- Deleting those images runs counter to the idea of building the best long-term free media repository and creates perverse incentives not to share accurate source information.
- At any rate, I chose colors based on what I think is the real risk to a reuser for getting sued for copyright infringement if they take a Commons image and distribute it. There is a risk for all our images - which is why reusers like Britannica still carry out (c)-clearance for Commons photos. But I think the risk looks like [Bronze] > [Silver] > [Gold] > [Platinum] where Platinum is a traditional copyright-clearance deal such as with the Bundesarchiv: an established relicensing organization, not just an established user. --SJ+ 11:52, 25 February 2013 (UTC)
Comment So what's your solution ? OTRS agent should have less strict policy ? Commons community should accept permissions even if OTRS ticket said it's not valid ? For sure we should pay more attention to uploads from flickr (however we really try to delete licence laundry files). --PierreSelim (talk) 09:46, 25 February 2013 (UTC)
- Some suggestions:
- Be more strict with Bronze images - if challenged, they should be deleted.
- Define a gray area for media whose license isn't 100% cleared, rather than pretending this doesn't already exist. (It does, but is hidden). Have a tag-and-quarantine solution rather than a "30-day ultimatum before deletion".
- Don't delete files to make a point. If it seems likely that an image can be cleared, help clear it; don't rely on arbitrary policy to delete at the first opportunity.
- Share the burden of clearance with the uploader, for images that are in use or have lasting value. Put the burden of "getting out of the gray area" on the uploader, but the burden of deletion on the deletion proponent should be "demonstrating no net value".
- Make the license-clearing process less of a pain, and less of a personal challenge.
- Give people credit and thanks for adding detailed source information - even when this slightly reduces certainty in the license.
- Make it easier to find and track and restore [one's own] deleted images.
- I am a bit with Sj here. There are cases where our current policy either discourage good contributors, or encourage them to lie to upload pictures. Example: If I upload pictures from a close friend of mine or a family member, and then credit him/her, I'd be required to provide a permission. If I lie and say that the pictures are mine, no question... We need to find solutions for this kind of cases without puting to much burden on the uploader. Yann (talk) 14:15, 25 February 2013 (UTC)
A longer list of suggestions:
- Be more strict with Bronze images - if challenged, including by creator or subject (or people reasonably claiming to be same - burden shouldn't be on the challenger to jump through hoops) they should be deleted. Exceptions made for active and irreplacable use, in which case the burden is on us to justify keeping it and confirming we have the right to do so.
- Define a gray area in the Commons repository rather than asking users to pretend it doesn't exist. (It does, but is hidden.) Have a tag-and-quarantine solution that does not involve deletion for Silver and Gold images that are challenged. View this as a long-term process to resolve license uncertainty, not a short-term debate about whether or not to delete and reject an upload [which rolls back the first 5 steps of the 10-step clearance process!].
- One tag should indicates a low level of uncertainty in the license, while allowing its use. [Including all Bronze images and probably some of the others. Right now this is invisible to reusers.] Another tag should indicate a high level of uncertainty, that needs to be resolved before the image can be used. In both cases, until there is a specific negative claim that says we should not keep an image, we should have it in the repository.
- Don't delete files to make a point. If it seems that a file is likely to be available under a free license if the rights-holder is asked in the right way, don't delete simply because noone has asked [or gotten written confirmation that they asked].
- Once an image is deleted, it is unlikely to return. If there is a reasonable chance it can be made clearly free, deletion should be a last resort, not a "standard 30-day ultimatum". A lot of personal image donations fall under this category: the intent was clearly to share, the source would likely send an email confirmation if asked the right way, but the OTRS process is too awkward / the uploader not precise enough, and the image gets deleted.
- This really is a problem, year after year. It usually has nothing to do with Jimbo; it just happened to involve him in the example that inspired me to write this.
- Put the burden on "getting out of the gray area" on the uploader, but put the burden of deletion on the deletion proponent.
- Many deletion discussions boil down to "burden on the uploader - if they're not here to defend their uploads, they go". If anyone finds a complaint, they can challenge the uploader to fix it, and delete the image if it isn't fixed. Even if it is clear that anyone /could/ fix it. Sometimes it is enough to go look at the original uploader's edit summary, and simply place that summary in the file's structured data, to end an aggressive deletion proposal.
- If a trusted uploader states or has stated that a file is available under a free license, and another user challenges that claim, deletion should require demonstrating that it is not under a free license. Keep active and potential use more clearly in mind as one of the factors in deciding how much time to give to facilitating license clearance.
- Make the license-clearing process less of a pain, and less of a personal challenge.
- Send a simpler, positive message to uploaders, with one-click ways to resolve each step in the clearance process. "Thank you for sharing this image. We can't yet recommend that it be included in articles or made available to reusers because 2 steps are missing from the license-clearance process. Here's how to add those steps."
- Consider the time invested in the projects by uploaders and community members. The more time they have invested, the more time we should spend helping make their uploads work out. Give people credit and thanks for adding detailed source information - even when those new details highlight the fact that one of the license-clearance steps needs to be resolved.
- Make it easier to track deleted images.
- Right now there's no obvious way for me to find or track my own deleted images. They no longer show up in my contributions; only admins get an easy link to that list. And there's no easy way to restore an image post-deletion if an obstacle is overcome.
- What we should do: publicly archive all images that seem likely to be clearable. If there are 10 steps to clearance, note which steps have been done. If someone comes by and fills in the unfilled steps, then the image can finally be cleared - no drama needed.
- What we currently do: a confusing combined discussion/debate about merits focused on a yes/no deletion vote. With a similarly confusing appeal process - ask the admin, post to undeletion requests, be available for Q&A. Almost noone ever figures out how to do this.
I first made this comment in response to an OTRS thread where Jimbo was attempting to send in notice of copyright release for two photos. This is the Nth time I have seen a long-time user bitten by this relatively broken process. But it reminded me that not only is the process broken, but we have gotten used to it and are not trying to make it better. Those two cases described below; feel free to add others, including any that turned out to be 'eventual problems'.
Both of these files are ones where [someone / we] can undoubtedly get full copyright clearance. The question is: whose obligation is it to get that, over what timeframe? And should we delete the files while clearance is transitioning from "maybe" to "clear"? [as Nemo notes, nothing is ever 100% certain; this is really moving from 80% to 99% clear.]
- Commons:Deletion_requests/File:JCA3_edited-1.jpg - a private photo donated to the uploader by the subject's family. Likely donated under "whatever license you need" like most private photos. I've done this a bunch of times - you send someone a template email, they fill it out and send it back to you or to OTRS. The more complex you make the process for them the easier it is for them to lose interest. Sometimes they aren't good with email and you have to really struggle to get a reply from them; but in person they are happy to give whatever approval is necessary.
- Possible fix: an Android app with a one-click form that both submits to the right OTRS address and copies the creator's email address (so both parties have a copy).
- Commons:Deletion_requests/File:Paul_Myners.jpg - a publicity photo donated to the uploader by the subject. Likely donated under the "I hired this firm, I can use the photo as I wish" theory. While this is not guaranteed to be true, it usually is in practice. The client asks their publicist to confirm they can release it under a free license, publicist confirms, client passes on that email. The photo was better than any alternative photo of Myners - though we have one that is low quality. [this is also generally a sure sign that you can get license release from an official photographer]. So deletion was not because we couldn't get release of this photo - of course we can; anyone could by writing to the agency, which was even noted in the deletion request.
- How does the uploader feel about such a process? Probably a bit bullied (ok, not in this case; but in general. especially after trying to work with OTRS). How does the subject feel about it? Probably like their donation was ignored. Neither outcome is desirable.
- Possible fix: add a queue/backlog of images waiting for agency/creator confirmation of free license. Interested people can then specialize in efficiently clearing that queue [to counteract those who specialize in efficiently listing things for deletion ;)]