Wikipedia:Featured picture candidates/Criteria RFC

From Wikipedia, the free encyclopedia
Jump to: navigation, search

OK, I was reading over the stuff [on WT:FPC] about sniping and declining quality of reviews. Someone said up there that they would like to see reviews more of the form "oppose - blown highlights (1a), too small (1c), subject cut off (1e)". I agree. Time to revive one of the things I drafted up before I went on wikibreak - a general overhaul of the criteria. Hopefully this proposal will increase the quality and consistency of discussion and remove some of the more emotional responses. It should make the criteria easier to understand, especially relating to photo quality.

The proposal[edit]

Change 1[edit]

This supersedes criteria 1 and 2.

1. meets professional image quality standards. In particular, at full resolution:

a) Optical and sensor quality: The image should not suffer from avoidable distortion or tilt, especially in architectual photos and scans. Blown highlights, crushed blacks [2], vignetting and noise/film grain should be minimised. There is no significant chromatic aberration and dust spots should be removed.
b) Editing and post-processing: Images are edited in a competent manner which does not degrade image quality. The image must not have visible compression, posterization or oversharpening artifacts or other signs of inappropriate/incompetent post processing. Any manipulation that causes the subject to be misrepresented is unacceptable.
c) Resolution. The image allows quality print reproduction. Still images must be at least 1000px on a side, though leniency is given for animations and videos. Panoramas need to be substantially larger than 1000 pixels in the longer dimension in order for sufficient details to be seen.
d) Exposure, lighting and tonality: The picture must have an accurate exposure, encyclopedic white balance and appropriate lighting. The image should have good contrast. Portraits should not exhibit the red-eye effect nor should they be taken with strong fill flash (typified by strong shadows and harsh highlights).
e) Composition and focus: The image has good composition. The subject must not be cut off or obscured without good reason. All important aspects of the image should be in focus, while those unimportant should be deemphasized (see depth of field). There are no distracting elements. Consideration is given to macro photography where the depth of field is often constrained by physical limitations and so the entire subject may not be fully in focus. Diagrams and maps are professionally and aesthetically laid out.
f) Special cases.
i) Panoramas should have no stitching errors. The frames must be consistent in exposure, focus and lighting.
ii) Historical images should be of professional quality corresponding to the available technology of the era. Digital restoration is encouraged, but is not required. Reproductions should not exhibit moire and restorations should be faithful to the original work.
iii) Diagrams, animations and maps. Vector graphics should employ web-safe fonts and are valid SVG. Animations are smooth and at an appropriate speed.

I think that covers most flaws. Let me know if I've left any commonly recurring ones out.

Change 2[edit]

This also incorporates suggestions some time ago to address reluctancy of GLAMs to donate materials due to lax attribution and whatnot.

This supersedes criteria 4, 7 and 8.

5. is accompanied by an image description.

a) Image description. Image description pages should have a description of the subject and any metadata in English. The main subject is properly identified, including Latin and technical names where applicable. For historical images, the creator(s), year(s) of creation, title of the work, medium, place of housing and (physical) size are provided, if available. Geotagging of images is encouraged.
b) Documentation of manipulation and restoration. Any digitally restored image or edit of a nominated image [3] should have a summary of the edits performed on the image description page, with links to the original/unrestored version(s).
c) Licensing. The image description page indicates the image is freely licensed and provides information to verify its copyright status. Fair use images are excluded and copyright violations will be deleted. The image is not watermarked.
d) Caption The image has an informative, succinct and complete caption that provides a context-specific summary of the description page. It properly identifies the subject and states the most relevant metadata (date, location, etc.)

Change 3[edit]

Proposed by Kaldari.

This supersedes criteria 4 and is incompatible with 5c above.

4. Is a Free Cultural Work.

a) The image must either belong in the public domain or be licensed under a free license appropriate to the media.
b) The image must not be covered by any legal restrictions or limitations which would impede reuse of the image.
c) The image should be free of copyright tags or watermarks.
d) Fair use images are not allowed.


  1. 3, 5, 6 will be renumbered to 3, 2, 4 respectively.
  2. Quantitatively <0.02%
  3. I should note that some of our photographic nominations are already extensively edited when they are uploaded.
  4. Existing footnotes will probably be kept.

It should be assumed that all of the applicable criteria must be met for the promotion of the image.

Some of these are negotiable in the context of discussing them here, especially those with respect to the image description. I wouldn't consider "no visible JPEG artifacts" to be negotiable, though. MER-C 07:32, 17 November 2009 (UTC)

  • 19 November: 1f, g, h => 1f i, ii, iii; 5a => 5c, 5b => 5a, 5c => 5b; revise caption; restorations faithful to original work
  • 20 November: 5c => 4; 5d => 5c
  • 23 November: No watermarks, keep my original proposal separate.


  • Leniency should be given on resolution for highly important images where no larger resolution could reasonably be expected to be found or is impossible to get larger resolution versions. Also leniency should be allowed for subjects that due to rules about cameras (recently came up with pictures inside Islamic mosques) prevent high quality DSLR images from being obtained, potentially indefinitely. — raeky (talk | edits) 09:03, 17 November 2009 (UTC)
    • I haven't been promoting FPCs for a while. When was the last time an image under the size limit was passed? I don't remember passing anything under the size limit that isn't an animation, SVG or video during the two or so years I've had a virtual monopoly on the closing process (if there was, I would have complained). If there are exceptions, they are very infrequent as to virtually not be worth talking about. MER-C 12:20, 17 November 2009 (UTC)
The bit about flexibility regarding unique/irreplaceable images applies to the whole of 1, so said pictures are permissible (but the exceptions are infrequent). The existing footnote will be kept. MER-C 08:27, 19 November 2009 (UTC)

Some initial thoughts:

  1. I like the collapsible compatibility table idea and the new Image Description criterion. Plenty remains to thrash out in general (most obvious I'll mention below) but great ideas.
  2. Criterion #1 should end at (e), with the remaining points under a new Criterion #2 for exceptions, non-photographic stuff and special cases.
  3. Point 1(g) on historical images is short of the mark in almost every respect; "professional quality" is misleading and of doubtful use, moiré is only one of many potential problems with scanned stuff, and there are currently no criteria at all for exactly how we encourage people to digitally edit scanned/historical works.
  4. The Descriptions Page stuff (#5) should be outlined in the exact same order it appears: Description, Source, Author, Permission. They all relate to important requirements that have been pulled up in recent noms.
  5. Not convinced #7 should be shoehorned into the new #5, as captions are a quite separate and specific issue that even need a footnote at present.

Those are my first impressions, anyway. I'd also suggest that the actual wording and style shouldn't be debated until we get agreement on basic content, to save complicating things. mikaultalk 20:33, 17 November 2009 (UTC)

I second that, wording for later. My concerns are about wording, not content, so I'll save them for now. Nezzadar [SPEAK] 04:03, 19 November 2009 (UTC)
  1. That bit comes from Durova's suggestion when I drafted this. I'll ask her for feedback.
  2. Done.
  3. I've always thought the caption was a context-specific summary of the stuff on the image page. I've revised the caption one so that it says everything in the current caption criterion. The existing footnote will be kept. MER-C 08:27, 19 November 2009 (UTC)

Here's wishing that a bit more give and take had occurred before moving to formal RfC. We had an RfC on this process less than half a year ago; I don't see the need for another. It's more a matter of reviewers being familiar with the existing criteria and using them. Durova366 18:51, 19 November 2009 (UTC)

I think having more specific criteria is helpful so that we don't continually argue over what are appropriate reasons to support or oppose images. Personally, I support all of the changes listed above. Kaldari (talk) 19:00, 20 November 2009 (UTC)
It can only be helpful to the extent that people pay attention. Without putting any particular editor out on front street, we've had problems from several directions of editors posting reviews which are clearly not compliant with existing criteria. It won't do much good to change the criteria when reviewers post without reading them or doing other basic checks. Durova366 05:55, 21 November 2009 (UTC)
I remember that problem... :) I can think of two plausible groups of clueless reviewers - those that haven't read the criteria and those who have but are uncertain as to their implications with respect to the various image forms. (I remember someone saying the current criteria are mainly suited to digital photography ... not sure if it was you)
We can't do anything here about the former, though a friendly note on the talk page will help; in which case the new criteria will help explain what's wrong with their votes (maybe a template message "Thanks for reviewing/nominating File:Example.png. Unfortunately the image suffers from $OBVIOUS_FLAW and as per featured picture criterion $CRITERION [where it helpfully says that "at full resolution... the image must not have visible compression artifacts"], is not likely to become a featured picture. Please read and understand the criteria before reviewing any further images" or "Thanks for reviewing File:Example.png. Unfortunately your reasoning does not address why the image should (not) be promoted as per the featured picture criteria. Please read and understand the criteria before reviewing any further images"). We should be a bit more strict in an obvious way on reviewers who don't respond to such a gentle poke.
The latter includes newcomers who are genuinely interested in participating but aren't very familiar with what the current criteria translate to in practice. The new criteria will help these people get up to speed. (We can say "lurk moar", but I'm not sure whether learning by example is the best way to get familiar to our expectations.) MER-C 06:30, 23 November 2009 (UTC)
Point of order... This problem of compliance and awareness will never go away unless we grasp some stinging nettles, and this ain't the time or the place to do it. We've repeatedly rejected appointing expert reviewers and the payoff there has to be the acceptance of more discretionary closing, which paradoxically permits exactly that. Here we saw a split in the consensus in this year's FPC RfC, with vote-counters on one side and consensus-weighers on the other. Unless we can agree to allow trusted expert reviewers to weigh up criteria compliance in closing, we're basically stuck with those problems and need to consider wider participation when we complain about them. However real the issue is, it has little bearing on the criteria, so let's leave it for another day. mikaultalk 22:14, 23 November 2009 (UTC)
To address the original comment in this thread, I don't see any need to thrash out anything fundamentally new here, nor arrange things in any way contrary to the consensus arrived at during the earlier FPC RfC. A re-jig of the criteria was floated at the end of that but never completed. As I mentioned at the time, the substantial changes to our closing methods and consensus assessment never really touched the criteria, just showed them up to be a little unfocussed and out of step with current priorities. The changes above seem to me to achieve that refocus and tune in well to the FPC zeitgeist, give or take a few more tweaks. I still think this discussion belongs on the WP:WIAFP talkpage, but it's no biggie. mikaultalk 22:14, 23 November 2009 (UTC)
@MER-C, it's flattering to see one's name invoked in the starting of a process RfC, if one actually wants the RfC. Sure, the criteria could use an occasional update. But that wasn't what I was driving at and don't think this process is really needed to solve the existing criteria shortfalls. The pressing matter is to encourage a reviewer culture where people check the galleries to see that we already feature a type of something before asserting we don't, and to read up on the difference between noise and grain, and to articulate not just how they feel about a nominee but also to articulate which elements create that impression. What worries me is that a second RfC in under half a year might drive off a portion of our better reviewers, and I actually don't like to see my name invoked in its inception. Durova369 22:24, 23 November 2009 (UTC)
  • I don't like the bit about animals being cut off. I fear it will be interpreted to narrowly, and animals that have merely a foot, part of their tail, or some small part of their body cut off will be voted down. We've promoted plenty of incomplete animals in the past, and for good reason - they illustrate their subjects well, despite limitations. Mostlyharmless (talk) 23:06, 20 November 2009 (UTC)
  • New licensing criteria: (b) doesn't make any sense - "share alike" IS a restriction on how the image can be reused. It also seems a backdoor way to bar GFDL and anything else not CC => instant delist for hundreds of FPs = not good. Agree with no watermarks, but a message (this image (C) by X, You may use it under free license Y) in the metadata is probably OK. MER-C 06:44, 23 November 2009 (UTC)
    • Yeah, I'd agree with that. I seem to remember making a bigger deal of licensing in the criteria was deemed unnecessary a while back. I'm quite happy with the substance of the proposed 5(c) and see no point in (or consensus for) expanding it further. mikaultalk 22:20, 23 November 2009 (UTC)
      • I disagree. The freedom of our images is the primary thing which distinguishes us from every other photo gallery or image collection on the internet. If that's not important, I don't see why we're bothering to exclude fair use images. It is true that "share alike" and "attribution" requirements are restrictions on how images can be reused. They do not, however, impede the ability to reuse the image whatsoever. What if I changed the wording to say "... which would impede practical reuse of the image"? Kaldari (talk) 22:31, 30 November 2009 (UTC)
        • Yeah, you could certainly make that clearer, I don't see that being a problem. No-one is suggesting licensing isn't important, just that there's no need to have a detailed, multi-point criterion on the issue. It's a requirement, we can link to more detailed consideration of what free licensing means, but there's no need to do that here; basically, as an important part of image description page info, it belongs in a section on image description page requirements. Much neater and easier to tick off as you go, I think. mikaultalk 23:11, 30 November 2009 (UTC)