Commons talk:Quality images candidates

Color or black and white?

One topic keeps bothering me. It would be good if we could also make a basic statement on this in the guidelines. Photos can be edited in different ways. This creates derivatives, but also fundamentally different ones. I'm talking about these in particular: Color and black and white. Anyone who is currently working with black and white knows that these photos have a fundamentally different set-up than the corresponding color images. Not every color image is suitable as a black and white version and vice versa. Since the two approaches are clearly different, I think it is permissible and sensible to nominate both variants as QICs. (Quite apart from that: the software I use to process the photos also distinguishes between the two methods. This also shows that the development of the photos is different).

It is certainly more difficult with small changes, such as a different section. It is always difficult to find a clear boundary, as transitions are sometimes fluid. I can well imagine that with an opening, some photographers could come up with the idea of nominating any number of only minimally altered images. It won't be possible to find a really clear boundary, so it's up to the reviewer to decide.

For those who think that black and white photos (category Black and white photographs) are simply created by pressing a button on black and white, I would ask you to take a closer look at black and white photography. In my photography courses, I always point out the possibilities and use examples to show the diversity of black and white photography. In this respect, I would also like to point out the creative possibilities at this point. XRay 💬 10:46, 16 February 2025 (UTC)

Hi @XRay is your photography course online? Riad Salih (talk) 20:19, 17 February 2025 (UTC)

Unfortunately not. I give photography courses at our local adult education center. --XRay 💬 20:27, 17 February 2025 (UTC)

I'm well aware that creating an appealing b/w image is more than just removing the color. What I'm unsure about in the context is:

• Don't color images usually suite Commons's purpose - providing educational images of documentary character - better than b/w ones?
• Can b/w images comply with Commons' QI guidelines on color space and the use of 'natural' colors?
• Can two versions of the same image, one color and one b/w, be QI?

And don't get me wrong, I don't have an answer nor a clear opinion on any of these questions. I've recently supported a b/w image because the lack of color emphasized the scene in my opinion, but to be honest, in most b/w pictures I don't such benefit. Plozessor (talk) 17:26, 18 February 2025 (UTC)

Thank you for your comment. There is one phrase that I have a problem with again and again and more and more: "educational". This term is often used in a rather restrictive way, but as far as I know it has a broad meaning. There are so many ways to use images. It's impossible to keep track of them all. I don't want to presume to draw realistic boundaries here. --XRay 💬 18:16, 18 February 2025 (UTC)

I think "educational" means that the image should be useful to illustrate something, which can be a place, an epoch, a style of art, a type of photography, whatever - just more than 'See this beautiful picture I took!' or 'See this work of art I created!'. But say, we want to document how a church looks like, a color image of the church serves that purpose better than a b/w image.

While agreeing that b/w photography is a special technique, we can also call it art, I do not think that b/w images are very useful for Commons' purpose, with few exceptions of course. But besides for explaining photo techniques, I can't imagine many areas where a b/w image would serve the 'educational' purpose better than a color image. Plozessor (talk) 05:07, 19 February 2025 (UTC)

As far as I know, documentation is not a primary goal in the project objectives. People often only think in terms of Wikipedia dimensions. But Wikimedia Commons is more: an archive for all (!) Wikimedia projects! And not every project focuses on documentation. I always see pure documentation as a limitation. What certainly doesn't belong in Commons are purely private images such as private party pictures.--XRay 💬 06:21, 19 February 2025 (UTC)

Perhaps some basics about black and white. Black and white photography was regarded as the “real” pictures until around the 1960s. They were the basis for documentation. Color was seen as distorting. Then black and white became rather insignificant. Since the 1990s, black and white has been developing again, more as a genre and form of expression in its own right. It is often used when color is perceived as disturbing or distracting. Black and white photos help, for example, to emphasize structures more clearly or to avoid distractions. --XRay 💬 08:20, 19 February 2025 (UTC)

And a postscript: I can imagine several QICs from one RAW file in addition to color and black and white. If the resolution is high enough, the full image might not be QI - for example, if it contains too many distracting elements or the composition is not good enough. However, several extracted images could be QIC in their own right. In my opinion, nothing should be rejected across the board. And each nomination should be reviewed on its own merits. If someone - for whatever reason - overdoes it, you can address this person specifically. --XRay 💬 10:03, 19 February 2025 (UTC)

Documentation is not a primary goal in the project objectives

That may be correct, but the "educational purpose" is defined as an official policy in the scope of the project

Each nomination should be reviewed on its own merits

That is how I've handled it so far. And if I see a benefit in having this picture in b/w, I will not hesitate to support it. Plozessor (talk) 10:40, 19 February 2025 (UTC)

What about selective color shots, HDR or color alterations, or removing large objects by cloning? Which of these practices should be acceptable and which should not? Wilfredor (talk) 23:26, 20 February 2025 (UTC)

Another QICBot failure

QICBot failed to remove the promoted and declined images from the candidate list once again. I have the intention to fix the issue for the gallery of recently promoted images by removing the duplicates tomorrow, but the images will be archived twice (as usual) unless someone else has an idea to avoid this. --Robert Flogaus-Faust (talk) 10:44, 17 February 2025 (UTC)

Likely an edit conflict again. I could modify the script so it dies when it can't save the candidates list after extracting the new nominations, which would avoid the duplicates in the galleries, but would delay posting the new QI's by a day unless I can figure out a way to auto-restart it when that happens. Thanks. Mike Peel (talk) 18:59, 17 February 2025 (UTC)

I don't think that this is necessary. In fact, it should be really easy for me to remove duplicates from an archive page semi-automatically. I might give it a try with a small offline script (to be written by me tomorrow) and a single edit of the archive page for February 18. The gallery with the recently promoted images is not a problem anyway. I dealt with that before within a few minutes. All the best --Robert Flogaus-Faust (talk) 21:51, 17 February 2025 (UTC)

 Done My offline script for the elimination of the duplicates from an archive page worked nicely. I removed everything from the latest archive that had been archived yesterday. Two images archived yesterday were not in the archive for today, though. I will check again whether they are still on the candidate list. --Robert Flogaus-Faust (talk) 21:45, 18 February 2025 (UTC)

I found that these two images were archived as unassessed, but both of them received supporting votes before QICbot could remove them from the candidate list one day later. I removed them from the archive for February 17 and I also removed Category:Unassessed QI candidates from these files. So everything should be perfectly o.k. now, hopefully. --Robert Flogaus-Faust (talk) 22:52, 18 February 2025 (UTC)

Same issue today it seems (sorry @Robert Flogaus-Faust: ...). Thanks. Mike Peel (talk) 07:03, 28 February 2025 (UTC)

Thanks for your notice. I'll deal with this issue tomorrow in the same manner as last time, cleaning up the duplicates from both the archives and from the gallery of the recently promoted images. --Robert Flogaus-Faust (talk) 11:04, 28 February 2025 (UTC)

One possibility to solve this in the longer term might be to turn your offline script into something the bot can run daily after the main run to check and remove duplicates, if you think that might be feasible? Thanks. Mike Peel (talk) 12:02, 28 February 2025 (UTC)

Thanks for your suggestion. I am sorry, but my scripts are just two rather short Visual Basic Script (VBScript) files on my computer, each with less than 100 lines including some debug features that are not required any more. Duplicate images are usually rare except when QICbot encounters an edit conflict or very rarely for duplicate nominations. However, if any "File:" prefixes in the gallery of recently promoted quality images are missing, neither the QI categorization tool nor QICbot will recognize the respective images. Fixing the latter issue on a daily basis might be a good idea.

I have never really been a professional programmer, even though a few of my programs were considered to be quite useful by my former colleagues. I had some experience programming Basic, Pascal and a bit of assembly language for DOS. This was some decades ago, of course. All of my more recent programming efforts have been VBScript (.vbs) files (deprecated now, but still working on my computer). I have no idea whatsoever about programming Python or bots. But if you think that my code might be helpful for you you to add some useful features to the bot, then I will send my scripts to you by email if you contact me via Wikimail. However, my scripts were just intended for me. So a little bit of documentation what they are supposed to do and also a few minor changes (English instead of German) might be quite helpful for you. This might take some time because I will be really busy off-wiki during most of next week, unless everything can be done tomorrow, of course. --Robert Flogaus-Faust (talk) 18:29, 28 February 2025 (UTC)

Thanks, I don't know VBScript, and I don't think that can be run on toolforge. Perhaps you could document the logic of what you're doing, and I could try to write a Python script to do the same tasks at some point? I'm also really busy these weeks, so I don't think there's a rush. Thanks. Mike Peel (talk) 12:40, 1 March 2025 (UTC)

Same issue yesterday. I fixed the list of recently promoted images and the archive page for March 8. No changes to the March 7 archive or to image categorization were required this time. --Robert Flogaus-Faust (talk) 12:30, 8 March 2025 (UTC)

Description parsing bug

To prevent parsing issues with images that lack English descriptions, use non-description templates, or have overly long descriptions, I request this change in QInominator so that the nomination description is derived from the image name. This prevents problems with description parsing and generates shorter descriptions, making the QIC nominations page less bulky, while also encouraging users to create more descriptive image names. Wilfredor (talk) 15:37, 17 February 2025 (UTC)

Thanks for your contribution. However, the English description is more useful for me, because it also contains the English name (if any), whereas I avoid the common names in my file names. The stuff about the license that is frequently copied after that is not useful at all, though. Nevertheless, I prefer the gadget as it is now. --Robert Flogaus-Faust (talk) 16:35, 17 February 2025 (UTC)

A descriptive filename can still include the English name if you wish. This small change would reduce clutter, encourage better filenames, and avoid repeated license blocks Wilfredor (talk) 20:13, 17 February 2025 (UTC)

Sure, but I would probably just copy the names from the English description, which includes the scientific species names in italics. Anyway, your change would not be a serious problem for me. I am just content with the way it is now. Let's see what other people think about your suggestion. --Robert Flogaus-Faust (talk) 22:03, 17 February 2025 (UTC)

I understand what you mean, however, we still need to open the image and look at the description to evaluate it, so those long descriptions could mess up the process. Yeah, let's see what others think, maybe some other ideas will come up to improve it. Wilfredor (talk) 02:34, 18 February 2025 (UTC)

First, it is not a good practice to just apply the default description suggested by QInominator in the first place. But if, then the description is usually better than the filename. Also using the filename instead of the description would not solve all parsing issues because we also have filenames with special characters.

I can live with the suggested change, but I'd prefer the current approach. Plozessor (talk) 04:57, 19 February 2025 (UTC)

Ok, based on this thread I have modified the code change request so that it does not add license information and other things and if it does not find a description in English it will take the first one it finds but first it will look for the one in English if not it adds the one it finds Wilfredor (talk) 19:33, 19 February 2025 (UTC)

I am not convinced by the proposal. I consistently name my images using the native name (for example Uppsala domkyrka and Berliner Dom, not Uppsala Cathedral and Berlin Cathedral) but translate the image description into English. Now I can get the English description automatically when I nominate images here, but with this proposal I either need to manually change the text or nominate images here without an English description (since my file name is not in English).--ArildV (talk) 20:49, 19 February 2025 (UTC)

The current proposal changed, I'm just fixing the bug of adding license and anothers informations. It will take the english description but if the english description is not there, I will take the first description found. Wilfredor (talk) 01:58, 20 February 2025 (UTC)

Perspective issues "easy to fix"

Unfortunately, I missed the archiving of this QIC candidate, which is why I would like to leave a brief comment here:

• @Plozessor Skewing the façade in the foreground to make the proportions look more pleasing would be a fake, because it would result in two (ore even more) different vanishing points with parallel horizontal lines. The only thing that would help here would be a greater distance between the camera and the object and therefore a smaller angle of view. I have no problem at all with this QI candidate being rejected by others, but then please give a technically correct reason or simply state that the image does not suit your personal taste. I think the photo is flawless as far as geometric optics are concerned.

Smial (talk) 13:08, 18 February 2025 (UTC)

I'm not sure, @Smial. I understand that skewing is difficult or impossible if we have a deep object, but here it's more or less a flat surface. I would have done it roughly like here. Would you consider this an inacceptable fake? Plozessor (talk) 17:15, 18 February 2025 (UTC)

Then I misunderstood the various reviews in CR, I'm sorry. They sounded to me as if the difference in size between the part of the building in the foreground and the three-storey wing on the left were the main problem and that this difference should somehow be corrected separately. Your proposed version would be a viable solution, even if the tree has grown a lot and is cut off at the top. It corresponds to a partial horizontal correction. With 100% horizontal correction, as is generally required here for vertical perspective corrections, it would just look wrong in a different way. Smial (talk) 11:50, 19 February 2025 (UTC)

NP, I was referring to Photoshop's "Transform/Skew" function, Adobe's terminology is often different from natural language. Good that you brought it up! Plozessor (talk) 15:00, 19 February 2025 (UTC)

User:Spurzem

Unfortunately, I had to open a thread about Spurzem on the Administrator Noticeboard due to his recent comments on Consensual Review. I really wanted to avoid this, but I've warned him twice before, so I felt it was my responsibility.

Anyway, you can share your opinion in that thread. -- Jakubhal 15:30, 13 March 2025 (UTC)

Hello Jakubhal, it's amazing what you've gathered against me, and your sense of duty deserves recognition. -- Spurzem (talk) 15:43, 13 March 2025 (UTC)