Commons talk:Project scope/Update 2013/Must be freely licensed or public domain/Precautionary principle

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search
COMMONS SCOPE REVIEW
  • Click on the 'Project page' tab, above to see the current policy/guideline wording that is under discussion on this page.
  • To make a specific proposal, please start a new subsection and use the code below to put it in its own box. You can sign underneath the resultant box, but for technical reasons you can't use "~~~~" within it. Please number your proposal for ease of reference.
{{divbox|amber|Proposal number and title|Introduction
*text
*more text}}
Scope Review 2013 links:

Discuss stage 2 of this review

Translation

Background

Links to current rules

Discussion: Introductory Scope wording

Discussion: Files

Discussion: Pages, galleries and categories

Discussion: Areas of particular concern

Discussion: Identifiable people

Other proposals

Some suggestions for discussion on this page
  • Is this too lax? Or is it too tight?

Proposal 1[edit]

Proposal 1: Keep the existing policy untouched
  • Please discuss the above proposal here

Proposal 2 - Define significant[edit]

2 - Define significant
I would propose adding the following:
In this policy the doubt is considered to be significant when there exists some evidence that the media maybe unfree even if that evidence is inconclusive. Examples of such can be: past copyvio uploads, given source points to the search engine, contradictory information (such as "This is a photo of myself at the age 3 and my own work"), etc. In some cases lack of metadata and filename may also be such evidence. However, lack of evidence to prove the claim of own work when there is no evidence to suggest otherwise does not constitute significant doubt.

I see a lot of people appealing to this policy and even quoting from it, but they completely miss the term "significant". It needs to be defined to make it more clear what it means.Sinnamon Girl (talk) 04:10, 22 June 2013 (UTC)[reply]

I'd be happy with something along those lines assuming we could agree some detailed wording. Am I right in thinking that the intention would be to discourage users from nominating a file for deletion on the basis of no evidence at all? The current wording might imply that the existence of some evidence (no matter how trivial) would mean there is by definition significant doubt, which probably was not the intention. --MichaelMaggs (talk) 11:33, 24 June 2013 (UTC)[reply]
That was my intention. And my proposal should be read as the general idea, rather than the final wording. The files where significant doubt about the copyright exists should still be deleted, but not the files where the new uploader has contributed {{Own}} file. Sinnamon Girl (talk) 12:18, 24 June 2013 (UTC)[reply]

Fuck NO! As I've argued before, we should delete or keep based on a preponderance of the evidence. If someone doesn't want something on commons with this change, all they have to do is make a claim that, even if it's ridiculous, creates the tiniest shred of doubt that the work isn't PD, and we must delete (or violate policy). Even a ridiculous claim counts as "some evidence" I've seen files deleted based on nothing more than rather farfetched claims that were considered 'significant' just the same. Alternative below. --Elvey (talk) 02:04, 3 July 2013 (UTC)[reply]

  • oppose. Not keen on loading this policy point with lots of examples. Also it is contradictory because "no metadata" is lack of evidence: evidence that we routinely (but not universally) expect in a user-generated image. So the whole definition that significant doubt requires evidence is flawed. The further problem with this proposal is that should such guidance be needed, it should be part of guidance on deletion reviews. This is a policy page on scope should should be setting principles, not getting dirty with the details. Colin (talk) 07:52, 4 July 2013 (UTC)[reply]
    • I am open to rewording of the proposal so that it defines the principle only. The reason why I wrote it in the way that I did, was because I simply didn't know how to do otherwise. But reading policy pages I find that examples or often useful, so maybe we can state the principle first and then do some sort of indented "For example" section. Sinnamon Girl (talk) 12:17, 4 July 2013 (UTC)[reply]
      • But this is a policy page defining that, in "principle", we take a "precautionary" approach to possible licensing/copyright problems. There's no need to go into detail on the policy page. The actual way someone might examine and present evidence for making a deletion review comment belongs on the deletion guidelines, not here. And it involves making a judgement of the "available" evidence and the "unavailable" evidence factored together. Colin (talk) 13:45, 4 July 2013 (UTC)[reply]
        • Maybe technically you are correct, and then we should remove "Also, arguments that amount to 'we can get away with it', such as the following, run counter to Commons' aims 1)...2)...3)...4)...5)...", but that would completely gut the policy. I would like to make the policy more clear rather than saying "may the closing admin be god, and the whole outcome be dependent on who closes the discussion". Since it's the term "principle" that is the issue, we can rename this policy into "Precautionary Rule". Sinnamon Girl (talk) 02:21, 5 July 2013 (UTC)[reply]
          • There's already an "Evidence" section above the "Precautionary principle" section and we have Commons:Deletion policy too. Please stop doing this "if we do that then XYZ also need to be done / are likely to be done / is inevitable" argument. Let's just stick with one thing at a time. It keeps the discussion focussed. The current text might be improved but here we're discussing your proposed change. The term "principle" isn't an issue at all and is fantastic for policy pages. "Precautionary Rule" sucks. Colin (talk) 07:28, 5 July 2013 (UTC)[reply]

aracters. ChrisHodgesUK (talk) 14:48, 23 July 2013 (UTC)[reply]


This is a topic of considerable community interest/importance. Please see Commons talk:Project scope/Update 2013/Stage 2 to discuss how we should proceed from here. --MichaelMaggs (talk) 17:52, 23 July 2013 (UTC)[reply]

Proposal 3 - Define significant appropriately[edit]

2 - Define significant appropriately
I would propose adding the following:
In this policy, doubt is considered to be significant when a preponderance of the evidence indicates that the media is unfree even if that evidence is inconclusive. Examples of such can be: past W:copyfraud, contradictory information, metadata. Lack of evidence to prove authorship or copyright ownership when there is no evidence to suggest otherwise does not constitute significant doubt. The quality of the source, including the reputation of the source (both of the file and any evidence for or against its status as a free work) should be weighed.

Files should NOT be deleted for licensing reasons if "to the best of our knowledge" they are free. Files do NOT to be deleted if there's but a small chance that they're not free. So our policy should state explicitly that files should NOT be deleted for licensing reasons if "to the best of our knowledge" they are free. This strikes an appropriate balance among the goals of the project (including managing liability risk, being ethical, doing important work, and using limited resources efficiently). Significant is an extremely vague term (much like "material", as in "material adverse effect on earnings") that doesn't really have a generally agreed upon definition and doesn't really belong in policy as is. It's exactly that vagueness that causes much of the common, and problematic friction on commons. So, I do think a policy change is needed. The "significant doubt" term needs to be removed, defined, or replaced with something with some precision to it, e.g. "where the balance of the evidence indicates that the work is likely (as in 50%+ odds) non-free, or there is equally strong but conflicting evidence, it should be deleted." The quality of the source, including the reputation of the source (both of the file and any evidence for or against its status as a free work) should be weighed. --Elvey (talk) 02:04, 3 July 2013 (UTC)[reply]

  • A user's only contribution is a 300x200px image of a sporting event, clearly taken from the sidelines. There is no camera metadata. A reverse image search using Tineye and Google finds nothing. The uploader has licensed the image as PD-self. Does this count as "significant doubt" for the purposes of deleting the image? --Carnildo (talk) 22:03, 3 July 2013 (UTC)[reply]

Reply to Carnildo: Consider what changes with this policy! Here's what changes:

Under current policy, the answer is most definitely: There is no answer! So under current policy, if there's a deletion discussion of this hypothetical image, the closer is completely free to decide that there isn't "significant doubt" for the purposes of deleting, or that there is.
Under the proposed policy, on the other hand, the evidence must be weighed: Do we have evidence that the image is free? Yes - the uploader's say so and reverse image search results. (And if this wasn't hypothetical, we'd have info on uploader reputation.) Do we have evidence that the image is non-free? Yes, info you provided - the low (web) resolution and lack of camera metadata, when combined with the PD-self tag are evidence thereof. Without further info, the uploader's say-so has little weight (we find many users often upload free work and many often upload non-free work), so the evidence that the image is something someone found somewhere is controlling - we know that this combination of factors are ones we find on non-free work more often (by a factor of well over 2:1, I'd bet) than we find on free work, so we can say with confidence that a preponderance of the evidence indicates that the media is unfree, but the evidence is inconclusive. Therefore, if there's a deletion discussion of this hypothetical image, the closer is not free to decide; the work would need to be be deleted under the proposed policy. --Elvey (talk) 07:19, 9 July 2013 (UTC)[reply]
    • I agree with this proposal in spirit and in a perfect world this is how it should work. We would be able to trust people unless there is some reason not to. Unfortunately we have seen this. Sinnamon Girl (talk) 03:16, 4 July 2013 (UTC)[reply]
      • Clarifying question: Your point is what? What is that link (to a whole bunch of images of different statuses) supposed to be evidence of? You need to connect it to a point about what policy does or should say about trusting people when there is no reason not to for it to be meaningful.--Elvey (talk) 07:19, 9 July 2013 (UTC)[reply]
Comment: NO, 84User! THIS POLICY needs to change because admins are using THIS POLICY to justify decisions that are often indistinguishable from "I'm deleting whatever I feel like deleting!" or "I'm keeping whatever I feel like keeping!" Therefore, THIS POLICY must change so that can't keep happening. --unsigned
  • Yes 84User, I have deleted that file because as you correctly established, it was an illegal copyright violation. Thankyou for your help, please feel free to nominate more images that you think are of similar dubiousness. --99of9 (talk) 09:17, 9 July 2013 (UTC)[reply]
Comment: That makes no sense; blank metadata fields are evidence, Colin! Do you dispute that admins can use THIS POLICY, as long as it remains unchanged, to justify deleting or keeping anything, as they feel like it, as long as the evidence isn't conclusive, one way or the other? That's why THIS POLICY must change; we can't leave it unchanged and just provide guidance elsewhere if we want that to chagne. --Elvey (talk) 07:41, 9 July 2013 (UTC)[reply]
Comment: This !vote seems internally contradictory. 99of9, did you really mean to say it's way too high a standard? Because a lower standard of evidence than preponderance would require deleting fewer works than having preponderance as the standard would demand. You really meant to say it should be much harder to delete works? Per the current Precautionary Principle: The onus in fact appears to be on the uploader to show sufficient evidence to convince every closer that there's no "significant doubt" that the image is free; otherwise, a closer is not wrong to delete.--Elvey (talk) 07:19, 9 July 2013 (UTC)[reply]
I mean preponderance (>significant) of evidence-of-unfree (doubt) is too much to require for deletion. I do not want to make it harder to delete possible copyvios (especially from drive by accounts that never reply to our request for more evidence). The onus is on the uploader to show that they are not copyvios. If they cannot do their job, the file should be deleted. --99of9 (talk) 09:06, 9 July 2013 (UTC)[reply]
ARRGH! Significant is an extremely vague term (much like "material", as in "material adverse effect on earnings") that doesn't really have a generally agreed upon definition and doesn't really belong in policy as is. That vagueness that causes much of the common, and problematic friction on commons.


  • Comment I agree with the spirit of both of the above proposals, if not the wording. Images are regularly deleted even though they are a) uploaded by a trusted user who clearly understands free licensing, b) seem likely to indeed be free, but c) are missing some piece of common metadata. [unless the creator notices a talkpage message within a few days].
Two pieces of language that I like here are "to the best of our knowledge" and "The quality of the source, including the reputation of the source (both of the file and any evidence for or against its status as a free work) should be weighed." Perhaps there is a way to fit those phrases into a revised proposal? --SJ+ 03:27, 9 July 2013 (UTC)[reply]
Reply to SJ: As I see it, two terms I use: "a preponderance of the evidence indicates that" (which I use in the actual proposal language) and "to the best of our knowledge" (which you like) are interchangeable; both hinge on whether the odds are above or below 50%. OTOH, "significant doubt" is completely subjective (as it stands currently, undefined); it doesn't hinge on anything remotely objective; that's the problem with it. Currently, admins regularly delete images that to the best of our knowledge are free, justifying the deletion as based on "significant doubt", even where there's consensus that the evidence that creates "significant doubt" is clearly weaker than the evidence that the works are free. --Elvey (talk) 07:19, 9 July 2013 (UTC)[reply]
  •  Oppose The burden of proof should be on those wanting us to keep hosting a file, as COM:EVID says. Absolute proof is often too hard to provide, but requiring deletion if there is any significant doubt about licensing/PD status is a reasonable standard. A mere "preponderance" of evidence is too weak a standard for keeping a file. --Avenue (talk) 12:35, 11 July 2013 (UTC)[reply]
This is insanity. As used in current policy, "significant doubt" isn't well-defined. Significant doubt isn't a higher burden than preponderance. It isn't a lower burden either. It isn't a defined burden of proof at all! It can be and admins regularly do interpret it to mean both more and less strict than preponderance. If "significant doubt" is well-defined, what is that well-defined definition?--Elvey (talk) 07:48, 13 July 2013 (UTC)[reply]
I don't claim every admin would interpret "significant doubt" in exactly the same way, although I do think your concern about varying interpretations is overblown. And my main point is quite simple and sane - requiring a "preponderance of evidence" that media are unfree before they can be deleted seems a much less stringent criterion than deletion upon "significant doubt" being raised, and IMO is too low a standard. Adopting it would put reusers at unnecessary risk. --Avenue (talk) 20:43, 15 July 2013 (UTC)[reply]

This is a topic of considerable community interest/importance. Please see Commons talk:Project scope/Update 2013/Stage 2 to discuss how we should proceed from here. --MichaelMaggs (talk) 17:52, 23 July 2013 (UTC)[reply]

Retrieved from "https://commons.wikimedia.org/w/index.php?title=Commons_talk:Project_scope/Update_2013/Must_be_freely_licensed_or_public_domain/Precautionary_principle&oldid=110932893"