From Wikimedia Commons, the free media repository

[edit] July 27

[edit] Photographers/ Color accurate folks- your opinions?

[copied to Commons talk:Color accuracy by Wikipeder 10:50, 9 August 2006 (UTC)] Please make responses there. (Mak)

A color calibration target is used for including an object with known colors in a photograph so that it is possible to adjust the delivery display (or printer) to reproduce the photograph's colors with the highest accuracy.

It would be nice if we could have some sort of suggestion for folks who are going out shooting objects where color accuracy is critical (cultural artifacts, paintings and so on). I guess the pro way to do scene matching would be to recommend taking the image with a color bar in the scene, eg Kodak Q-60 Q-13, GretagMacbeth ColorCheck, or even some Pantone based target.

That way, a hundred years from now, folks will be able to look at a Commons image shot in 2006 and have a pretty good idea of what the colors realy were because there are known colors in the scene. But those are proprietary. I suppose we could use a SMPTE color bar, but color spaces of various devices are hugely non linear, so really that is probably not enough data points. Can anyone recommend a public domain color bar target for this purpose that would be readily available to anyone doing some Museum shooting? -Mak 22:15, 27 July 2006 (UTC)

My thought is that we suggest folks use either the Kodak or ColorCheck targets, and cut the target up so that is is one long band, then photograph the object with the target at the edge(s) of the frame with the most representative lighting. Prior to upload, the photo is cropped so just a few pixels of each color is visible on one of the sides of the photo. This way, there is no motive to crop out an extraneous color bar because it shall not be noticable. A comment embedded in the EXIF metadata note which color target was used.

Commons will then be able to host authoritative images of paintings. Folks making prints of Commons images, or displaying images on electronic screens will be able to calibrate their output so that they can have color accuracy that is simply not possible without such a practice among commons volunteer photographers.

Of course the big problem is how to hold up a color band in the Louvre without the guards getting freaked out. Cameras are so small that they are easily concealed- but for low light, a wheelchair would be ideal for excellent stability and concealment of even the most bulky digital with a big light gathering lens- then take very long exposures, but I digress. But how to conceal that color bar? Hmmmm. Perhaps a helper who stands in frame with a Q-60 banded walking stick?

Joking aside, I suppose really we don't need Q-60 or Pantone ColorCheck targets. They are rather pricey ($60 USD) and possibly there are rights issues. All we really need is an image that a person could take to a photo finisher and have reproduced accurately. The spectral values can be specified extremely accurately in photoshop, so if we give them the correct values, they should be able to use a colorimeter on the prints until they get it right. Or would that wind up costing $60 for the resulting print? Humph. -Mak 19:13, 28 July 2006 (UTC)

The practical truth about color targets is that they can only be used under very controlled conditions. What I have done in the past is to take two shots, one with the color target and one without... I adjust the image with color target and then transfer the the adjustments to the one without. As long as the image is digital, the color information is in the file and you just have to pray that the photofinisher is well calibrated. One hundred years from now the color will not change. Color calibration is a pain, it varies from monitor to monitor and output devices... In the absence of any calibration device, for output purposes, try a woman... due to the fact that women are not color blind or color biased, train them in detecting yellow/magenta/cyan bias. On the practical side, I eyeball the image I put out in my printer and tweak it... when I take to photophinisher, I ask them to run a test print and calibrate according to their machine and print my image without adjustments. If the image is not right, I adjust in photoshop and reprint. On another hand... worry not too much about color reproduction... most people do not notice....--Tomascastelazo 00:56, 7 August 2006 (UTC)

Tom, the issue I am addressing is not about small variations of color. Without a known color reference in these images, it is impossible to correct the image, or know which one is closer to the truth. I wish this were an uncommon problem, but take a coffee table art book to a museum sometime. Ok, most aren't as bad as the heavily red shifted one illustrated here, but it is absolutely reprehensible what these printers do with colors. I mean, they aren't even vaguely close. Worse, even if the color is close, you are obligated to be skeptical. One might assume that there was a printer error (eg on a Lautrec, where in fact the painter really did put green cast in the flesh colors. Anyway, regardless what the file says the recieved light colors were, the camera does not know what type of light (or possibly what goofy filter) may be screwing with the real colors. My thought was that only a known reference color will take care of this problem.

I know it is way over the top heavy accurate, but Jimbo at the wikimedia talk in Boston challenged us to crank up the quality rather than the volume. This will address that issue. Commons art should be highly highly color accurate. That is something that Britanica really gives short shrift to. The way to do it is to stop scanning books and get good photos in museums. Otherwise we will be stuck with image choices like these.-Mak 02:38, 7 August 2006 (UTC)

I don't have any background in color calibration, but I take a lot of photographs in museums, so I'm interested. I gather that shooting paintings (for insurance demands, for instance) is a pain in the ass for many professional photographers. I'm not sure we can take really accurate pictures of paintings in usual Commons conditions (no tripod, no lighting system, etc.). It often happens that lighting is not homogeneous on the same picture: a dominant of tungsten light on one part, a dominant of natural light on one another. I'm not sure a color chart would help us in that case.

Now, we can still improve tremendously the quality of our pictures if we use these kind of charts. As for myself, I now use I collapsible white chart to correct the white balance before shooting (I never got any problems when holding it before an object, though museum guards tend to watch me closely in case I start frothing at the mouth). I shoot in RAW and then tweak the white balance if needed. The results are quite satisfactory while the process being easy (even if a bit tedious) to manage. If you explain to me how to use a color chart, I will certainly try it.

That being said, a lot of Commons users do not even change the white balance setting on their camera, so there's still a lot of education to be done. Jastrow 08:16, 7 August 2006 (UTC)

The color charts are used in high quality reproductions. The charts are known values of color. The idea is to place a color chart someplace next to the subject and take the picture. This picture is then taken to the photolab where they analyze the colors of the chart and give you a filter correction factor. Then you go back to your subject and reshoot with the exact conditions and you are supposed to have a perfect color reproduction. This was during the slide era. Now what you can do you open the digital file, analize the colors of the chart with photoshop, for example, and adjust color so that you get the correct amount of magenta, cyan, red. If you adjust one color of the chart, theoreticaly all should fall automatically into place. That is a quick way of doing it and you save on the reshoot. That guarantees that at least the digital information will be correct, but there is still the problem of the output. You can have a digitally correct file that will not look good in the monitor and will not come out with good color reproduction. You have to calibrate for light temperature, exposure (because even slight over/under exposure will alter the color balance, and not at the same rate for different colors), monitor and output device. And even if you get all things aligned.... the display light will alter it! So, in practice, just shoot with good exposure and eyeball the output to avoid magenta, yellow or blue bias.... Unless there is a strong cast, most people won´t notice anything.--Tomascastelazo 18:30, 7 August 2006 (UTC)

Image with colors pretty close to the colors as printed in the art book

Shot under incadescent light with auto white balance

Certainly in studio conditions like Tomas's scenario, a color target only offers miniscule difference when used by professionals who already are pretty close because they understand white balancing issues and exactly control and understand the color temperatures of their lights. I am not discsussing use by professionals but amateurs, and if we were talking about small variations, I would have zero interest in this subject. I am talking about huge differences in evidence in practically every duplicate image we have of art paintings. Editors are left wondering. Even folks who have seen the originals may not recall well enough the particular hues.

To illustrate- I carried out the procedure of eyeballing the white balance. THe results illustrate why a digital camera with manual white balancing will produce much inferior results compared to a process using a color target.

Using an art book image as a stand in for an original painting, I made 3 images. The first on the right will give you a fairly close idea of the original colors if you have a calibrated monitor. It was scanned on a flatbed scanner and I compared the original image to my calibrated screen. Pretty close. Ok. Next I took a photo from tripod using incadescent light and the camera on auto white balance. The results are the photo on the right. I knew it would be bad, but didn't expect it would be this far off. Now. If a color target with known color hues were present in the image, anyone other than the original artist could color correct it. We don't have to know anything about the light source to make the adjustment.

On the image page, you can see the information that the camera stated about the image. Pretty voluminous, but on the critical white balance value it says nothing- merely that it was automatic. Ok. Being able to eyeball the results on the lcd on the camera, I could tell it was way over the top too warm. So I went to manual and cut it down to 2500Kelvin.

Manually controlled color temp.=2500Kelvin
Closer, a pleasant enough image, but only problem- it isn't a very accurate reproduction, and gives the viewer a way different impression of the intent of the painter.

This is the result. Problem is- it looked pretty close on the LCD but not on the monitor. You tell me. Is it close enough? I don't think so.

We aren't talking about small degrees here. We are talking huge differences.

If I may, we could do a do it yourself, poor mans solution where the final output of the digital file is as close to the original we can get without going to expensive professional techniques.

1. Download and print out a color chart with a gray scale similar to professional color targets. It doesn't matter how crappy the printer is, or how far off the original is from the print.
2. Scan the print out.
3. Adjust the monitor so it is as close to possible as the scanned color target. (Including contrast on the gray scale range). Result of this step is you have a pseudo calibrated monitor.
4. Take a picture of the subject and include this printed out taget in the image.
5. Using photoshop or Gimp or whatever, adjust the image data until the taget on the screen matches the color on your physical target.
6. The result of this step is a pseudo color corrected image.
7. Uploader certifies as part of the Image notes that they have used these steps (or better).

Result- a low cost, non technical trust based calibration scheme.

Granted- Tomas wasn't kidding- there are a huge number of other challenges to getting colors close to the original. It is true that without hard core calibration techniques, the resulting image will still be noticably different from the original.

But you won't have the egregious differences we saw in the first two Eckersberg images.

What do you think? Am I high?-Mak 05:35, 8 August 2006 (UTC)

I think having a colour target in art images is a very good idea. It seems the only way to easily insure colour accuracy, given that the vast majority of monitors, printers etc. that display the images will not be colour managed, and the bulk of cameras, scanners and imaging programmes creating the image neither.

I suggest a different approach: A simple but standardised card with a black and a white target to be put next to any colour crititcal object. This will have the advantage that

• instead of subjective white balance, the white and black points can be set by a defined target using e. g. the GIMP's pipette.
• colour errors from manual or automatic colour or contrast correction are avoided. A problem with reproductions is that the brightest areas on a painting do not at all need to be the 255 bright white that contrast junkies like humans or their computers' algorithems usually try ty achieve.
• the white and black points can be set (and corrected!) by anybody who has the image on his screen, not just the photographer. The burden of dealing with colour accuracy is thus shifted from the army of amateur photographers to those who bother and to colour management experts.

A print-out-yourself card will not do, I'm afraid. There are tremendous differences in colour composition (black target) and reflexion depending on the type of paper, ink and printing process used.

The A and B field of a Q-13 grayscale will do. I don't see a problem with copyright there: It's just shades of gray, no proprietary colours like in the Pantone sortiment. --Wikipeder 08:46, 8 August 2006 (UTC)

I own a reference card which looks like the one you're talking about, but I've never used it. If you could just write a little tutorial about how to use it with the GIMP, I'll go to the Louvre this week and try it. Jastrow 09:12, 8 August 2006 (UTC)

I haven't got GIMP around here, I'm afraid, but I'll try to tell you what to do as good as I can. Maybe someone can look over this?

Tutorial for a RefCard and GIMP

1. In case you have a spot meter: When taking the picture, put the RefCard in exactly the same lighting as the object you want to take a picture of, and spot-meter frontally on the middle, gray target. Use the exposure value for your shot of the object.
2. In case you can influence your camera's white balance: Manually set your camera's white balance with the RefCard's gray target in the exact lighting the object is. If you can't do a manual white balance, select the most suitable of the pre-defined white balance modes of your camera.
3. Put the RefCard next to the object, making sure it's in the same lighting condition as the object and that it is frontally facing the camera. Take your shot.
4. Load your image into GIMP, go Tools-Color Picker and click onto the white target of the RefCard. This will show you the actual RGB value of this target. To realistically depict colours, the value needs to be L 97, a 0, b 0 in Lab color space, which GIMP doesn't offer, so we'll use the RGB equivalent, which is 246,246,246.
5. Then, open the curves tool (Tools-Color Tools-Curves) and in each of the RGB channels horizontally drag the upper right hand control point until the color picker will tell you that the value of the white target now is 246 in this channel. Unfortunately—as far as I remember—, in GIMP the color picker does not work simultaneously with the curves tool, so you will have to switch back and forth.
6. Do the same with the black target of the RefCard. It needs to be L 15, a 0, b 0 in Lab color space, or 38,38,38 in RGB. Voilà!

This procedure is a lot easier in Photoshop, admittedly. Once you set the general white and black point value to L 97, a 0, b 0 or L 15, a 0, b 0, respectively, you simply click once on the white and once on the black target for this manoeuvre. The link you supply offers a good tutorial on what to do in Photoshop (in French).

--Wikipeder 16:08, 8 August 2006 (UTC)

Mak, you are correct. That is one way to do it. In fact, I use a very similar method. Just be careful with the settings in the scanner. The best way is to override all auto settings, that includes exposure and color balance, operate in full manual. From there, you can do two things, 1) adjust in photoshop the color bias or better, adjust the scanner settings to eliminate the color bias. I always prefer to input the image as close as possible to the real thing. A good image will require less work than a bad image. A lot of people think that you can alter images with impunity, but that is not the case. Color shift of YMCK does not occur in parallel with exposure variations, that is, if you alter, for example, exposure by 10%, you would expect that the relationship between YMCK to vary accordingly, that is to vary up or down in the color values/relationship, but that is not the case. The best way to get into this is by learning the zone system and move on to color. To calibrate a scanner for exposure, make a gray scale of 10 steps in photoshop, print it, scan it and adjust so all 10 steps show. Make sure you adjust scanner first, not the monitor. You will have then two files to compare, a digital original where you can read values and compare to values of scanned image. If you want I can make a gray scale and send it to you with instructions.

Jastrow, best thing to do is to determine light source, read camera manual, experiment with color temperature settings, determine the best one for the particular environment and expose with an 18% gray card. In museums you don´t have much choice. If you really want perfect color reproduction, you must photograph under very, very controlled conditions. The 18% gray card exposure method will start you off with assuring an exposure that encompases the tonal range of the image, if it within the dynamic range of your equipment. Shoot one in the middle and bracket up and down by 1/3 stops. You may have to take a lot of pix for each painting, bt htat´s the way it is.... and at least it is digital. It is really nor complicated, eventually it becomes second nature, just like riding a bicycle.

If you shoot in Museums, do the following: 1. Determine exposure wth an 18% gray card. The card must receive the same light as subject. Even is the exposure readings vary once you point toward subject, use the info provided by gray card. Learn about it. 2. With the exposure information, take a picture of a target with a known value (rgb, etc). 3. Take picture of subject without target using the same exposure. 4. In computer, analize pic of target and adjust. Take note of adjustments. 5. Apply adjustments to pic of subject.

--Tomascastelazo 15:15, 8 August 2006 (UTC)

The advantage of using a white and black target instead of a gray card (or in addition) is that you also know about the absolute tonal range. A white on an old painting with patina is actually rather a dark yellow sometimes, and a defined white target that's on the image, too, can put this in relation. --Wikipeder 16:17, 8 August 2006 (UTC)

I am stunned by the excellence of these responses. For the benefit of others, it would be nice if we could pull these together into a guidelines page for folks who are into making high quality museum photographs. I wonder if you folks would be willing to contribute your thoughts and knowlege to such a thing? Both an easy to do, low cost approach for folks on a budget and not interested in gearhead settings that many folks don't have on their cameras, as well as one for those who have the tools and the inclination to pursue an even higher level of accuracy.

Good luck on your experiments Jastrow. At my location we have works of less general interest, but the museums aren't buttheads about tripods either. On the other hand, it would be fun to do some guerrila photography- bypassing the Louvre guards by using a wheelchair as a makeshift tripod. I wonder if you can do better than the Yorck project CD ROM version of Le radeau de la Méduse. As I recall, the lighting there is from the side and uneven. Who designs these Marquis de Sade lighting schemes of these museums anyway. Some are almost deliberately designed to produce wierd reflections so the only way to view them is from the strange angle. -Mak 16:50, 8 August 2006 (UTC)

Point to consider, these types of instructions can also put people off providing an image as they find that meeting the colour criteria beyond the capabilities of their equipment or themselves. Reading the discussion so far it sounds a complicated requirement to adhere to, when the opportunity to photograph is restricted either by time or equipment limits(storage space). Commons would be better served to have some way of identifing images that need to be recreated in a higher res, colour balanced. We want in the first instance a useable image of xxx, if you can do x,y,z then thats a bonus. Then we highlight that an image exists of xxx but we would appreciate another taken using the coonditions x,y,z. If quality is the general concern for improvement currently we load images from sites like flickr maybe to improve quality we set minimum requirements from these type of sites. Gnangarra 17:10, 8 August 2006 (UTC)

That is a very important point. The answer to that should be a two-step approach as I outlined above:

1. The photographer simply sticks a RefCard or standardised grayscale next to the object so it will be on the image, too. Nothing more (except proper exposure) will be required from him.
2. An image editor will adjust colour representation in the image with the help of the RefCard/ grayscale that is in the image.

Photographer and image editor may be the same person, but they need not be. This way, photographers do not need to bother about colour management and tricky image manipulation. As long as there's a RefCard or standardised grayscale in the image, after they uploaded the image, image editors hanging round at the Commons can easily replace the image with a colour corrected and possibly cropped version. Anybody whishing to verify the accuracy of the colour rendition can just go back in the file history and re-check with the uncropped version that still has the RefCard/ grayscale on it.

--Wikipeder 23:40, 8 August 2006 (UTC)

I strongly suggest the use of an 18% gray card because by only using a black and white target does not necessarily render white in its "correct" zone in the gray scale, while the 18% is a known reference value as far as density/exposure is concerned. This is very important when you consider the color variations possible due to over/under exposure. Regardless of the method used to color balance, exposure is definitely a must first step. It makes things a lot easier down the line. This, of course, is for color reproduction purposes. Worry first about exposure, second color balance. Those are the two variables to consider. Instead of a spot meter under these circumstances, I would use an incident light meter in order to avoid wrong readings.--Tomascastelazo 17:12, 8 August 2006 (UTC)

You are right about the importance of a gray card for correct exposure. I just think it should not be used instead of a standardised grayscale or RefCard, because it does not allow for adjustment of tonal range as described above. --Wikipeder 23:40, 8 August 2006 (UTC)

Unfortunately, the Radeau de la méduse is exhibited on the first floor of the Denon wing, where it's now forbidden to take pictures :-( Jastrow 17:15, 8 August 2006 (UTC)

How to make a gray scale in Photoshop

1. Creat new document

2. Size: 2000 x 1500 pixels, 300 dpi is a good start.

3. Pick gradient tool, white on one end, black on other.

4. Click on one end, drag to other end horizontally.

5. In menu bar go to image – adjustments – posterize

6. On posterize, type in 10 in the levels box that appears.

7. There you have it, a 10 step gray scale. You can calibrate monitor for exposure values with this one, print it and use to calibrate scanner.

8. Once you have calibrated for exposure, calibrate for color using methods described earlier.

I have uploaded a gray scale at Image:gray scale.jpg --Tomascastelazo 17:44, 8 August 2006 (UTC)

Gnangarra- I was imagining completely optional measures that a person could take if they were motivated to do so. It certainly would not be a requirement.

If a person had taken these steps, they would use a template something like the following:

I, the uploader of this work, certify that this photograph satisfies the first set of requirements for a colour corrected image. For more information on these requirements please see Commons:Color accuracy.

-Mak 23:56, 8 August 2006 (UTC)

Wikipeder - The gray card should be an 18% gray card, it can be bought in photo stores, it is designed specifically for exposure, it is a known value. Of course the use of a Mcbeth color chart would be nice too. Ideally, I would determine exposure with gray card, shoot a pic with color chart in image, take color chart off, reshoot. Then analyze in computer. It may sound complicated, but it is not. I´d rather have one good picture than two bad ones. It is worth the time and effort.

Mak - Crutches give you a higher point of view too... they make good monopods :o)--Tomascastelazo 04:23, 9 August 2006 (UTC)

Tomascastelazo: I am trying to think up a workflow that would allow very accurate colours while putting as little a burden on the photographer as possible. The idea is to not deter people but to have them simply stick a RefCard somewhere in the image so others can do the colour correction. One shot, if there is a RefCard on the image, it not yet colour corrected, if it is cropped already, it is.

Personally, I would do something different, too, ideally: shoot RAW, exposing right, with a colour chart and grayscale on the object, then take the shot with the same exposure values but without the charts.

On a guideline page, we should probably have both outlined, the at-least-stick-a-chart-somewhere approach as well as a guide to an excellency. Where would we best put such a page? --Wikipeder 08:15, 9 August 2006 (UTC)

[End of copy to Commons talk:Color accuracy Wikipeder 10:50, 9 August 2006 (UTC)]

Could you guys take this elsewhere now? It could be quite valuable to keep as reference. Commons:Photography colour calibration or Commons:Colour calibration just Commons:Photography tips or something, please? pfctdayelise (translate?)

There has been developing an extensive guide at Commons:Quality Images, How about moving this onto the discussion page for Commons:Quality images guidelines once finalised it will become part of the requirements for QI portraits any way. Gnangarra 08:57, 9 August 2006 (UTC)

Wikipeder - Thanks for the link... great one. We are basically suggesting the same thing. People then could upload both images and have someone color correct them. Maybe we can upload an example. However, I do not think it should become policy, only suggestion, and at the begining have it for artwork reproductions. I will try to get a color chart sometime next week and we can run the experiment. The lengthly discussion was necessary to bring light into the subject. --Tomascastelazo 13:34, 9 August 2006 (UTC)

Let's take all further discussion to Commons talk:Color accuracy Don't post responses here in VP anymore but in Commons talk:Color accuracy. OK? See you there. -Mak 16:05, 9 August 2006 (UTC)

[edit] July 30

[edit] Please move another category

Category:Native Americans flags to Category:Native American flags -- Himasaram 11:45, 30 July 2006 (UTC)

Never mind. I did it myself, manually... Feel free to delete the empty category. -- Himasaram 05:52, 21 August 2006 (UTC)

[edit] July 31

[edit] Where is "Vandalism in progress" (COM:VIP) link?

There seems to be an active Matt314 (talk · contribs · page moves · block user · block log · upload log) who is tagging things for speedy deletion, as the deletion process seems to be undergoing some automation, the vandalism is probably harder to catch. These images and probably many others, are in use, as CheckUsage plainly shows. Therefore, marking them for speedy deletion is obvious vandalism. --Connel MacKenzie 04:19, 31 July 2006 (UTC)

Hi! I did mark some images as {{Morguefile}} yesterday which redirects to speedy delete, please see en:Template talk:MorgueFile. Sorry, I was not aware that these images are heavily used, but since there were a lot of these images that I found via Special:Linksearch I checked usage for only a few. --Matt314 07:32, 31 July 2006 (UTC)

Listed them at Template:Deletion_requests#Images_from_Morguefile. --Matt314 07:40, 31 July 2006 (UTC)

Well, marking used files for speedy deletion is OK if they are clear copyvios or or clearly against policy. However, in the case of images that have been around for a while, with their license status being disputed, a regular deletion request would of course be better, and unlinking would be good too (preferrably with the help of the local communitites).

Calling this vandalism is a bit far fetched; As far as I can see, the images have to be removed. -- Duesentrieb^(?!) 08:40, 31 July 2006 (UTC)

Characterizing it as "vandalism" is perhaps less than productive. But your new template "morguefile" is a new twist, and quite unfamiliar. Having read the above pages, it is completely unclear if the name "morguefile" refers to "dead images" or to some website, or both.

I still don't understand commons. Why do you do this to sister projects? You've got that cached "red X" replacement method you can use if you're not going to ever use the CheckUsage tool. The same links to commons don't exist when there is no image.

The reason this (or, rather, these images) don't meet commons' criteria is still very unclear to me. Does tagging it with 'morguefile' somehow magically identify a copyright violation source? Why doesn't it identify that source? The veracity of whether it actually is a copyvio can't even be checked. The template name "morguefile" implies only that the image has already been orphaned (which obviously is not the case!)

Commons' policies again seem to be exceeding their remit. Even with some sister projects using CheckUsage to identify major errors on the part of commons' sysops, the knee-jerk reaction to identified problems is unacceptable. This image was re-vandalized and deleted without a replacement placeholder, nor without being orphaned. The overall policies that commons are permitting are completely out of hand.

How is it that any commons sysop actually believes actions like this are helpful? There can be no such thing as a speedy delete on commons. Replacement of images with a cached placeholder sidesteps all the problems your current abusive policies permit. Why is commons so resistant to doing things right?

Perhaps you don't believe that these policies are frustrating, infuriating and disruptive? Do you (collectively) just not see the problems this causes? --Connel MacKenzie 16:48, 1 August 2006 (UTC)

Morguefile used to identify images from http://www.morguefile.com/ which was believed to be a free source. After several discussions (including as linked above), it was decided their license was not sufficiently free. The terms of use are not acceptable. So Morguefile images should not be used.

The template was thus deleted which is rather unfortunate as non-admins cannot now see its history. It was recreated as a RDR to template:noncommercial to alert users who tried to use the template, that it was no longer acceptable.

This is a pretty standard process that I don't see much wrong with. We have to be flexible in response to changing understandings (and also changed licensing terms).

User:Matt314 is not an admin. I'm not sure if you're angry about the tagging he has done or deletion someone else has done. If it's deletion I suggest you approach that admin directly.

Commons has to balance copyright concerns against the desires of the projects. It's an extremely long, unfun and involved process to not upset anyone. With all due respect, no copyvio would ever be deleted if that was our only concern.

I'm pretty sure we've never claimed to be perfect or above reproach, either. So we make mistakes, we're human, sorry we haven't solved that one yet and managed to run the Commons solely on bots. pfctdayelise (translate?) 09:17, 2 August 2006 (UTC)

I'm not saying that human sysops should be perfect (obviously, I am far from it.) I'm saying that something is still systemically wrong with commons. Note also, that I'm not asking you to balance copyright concerns against anyone's desires. I'm only asking that images not be deleted when not orphaned. This can be accomplished a variety of ways. But I still think the best method offered to date, is to replace the copyvio images with the red "X" image. That way, image linkages on sister projects would be able to find the page on commons.

Take me, for example. I contribute content to commons, in the form of audio pronunciations I record, properly categorized through the commons' maze. I also upload occasional pictures from my digital camera (again, properly licensed, properly categorized through the commons' labarynth.) I focus my efforts in the area where I can help the most: the Wiktionary written for English readers. With about 50,000 manual edits in my main account, and many thousands more edits from 'bot accounts, one might say that I really know my way around en.wikt:. I helped get Commons Ticker turned on, on the English Wiktionary. But I, myself, am flummoxed, when it comes to finding a commons conversation about an image or audio file, if it is linked from en.wikt: but deleted on commons.

Am I really that poorly informed about how commons works? I'm sorry, but I have to doubt that conclusion. Is commons really that poorly organized? Well, no, internally it is mostly self-consistent. The only major, wholesale problem with commons is that sister projects don't have any of the needed, proper links to commons' discussion pages. Once an image or sound file is deleted, the only links that remain on sister projects are to LOCAL upload pages. Finding your way to commons then, is nightmarish, to say the least.

For other sister projects, such as the English Wikipedia, where no CommonsTracker is possible, the difficulty is increased dramatically. Especially if, like me, you are not a regular contibutor there.

The one single most helpful thing commons can do, is to replace files, instead of deleting them. The continuous stream of commons admins claiming that it is difficult, implies that commons admins don't participate in sister projects anymore, at all. I don't understand the resistance. There simply should not be a delete function on commons (except for previous revisions.) But the delete function, as it is used today, causes more problems than it solves.

A recurring theme I hear espoused on commons is that sister projects don't participate in commons activities enough. I think it is clearly because the links don't work. Commons Ticker is of interest only to the small handful of people involved with it. But almost all pages are affected by commons activities. If you'd like more participation from average user, then please do something to make commons accessible. The links can work properly, if you replace files instead of deleting them. And if the links work, you may find out just how many are affected by deletions.

Thank you for your patience in reading this. --Connel MacKenzie 19:53, 6 August 2006 (UTC)

I agree that it is very hard to find an image that was deleted on Commons, from redlinks on local projects. However I see this as a fault of the poor software. It should be trivially easy to provide a link to the Commons deletion log for pages in the Image: namespace. I actually opened a bug about this recently but unfortunately bugzilla: is throwing errors at me. If you can access it, search for "redlinked images" and you should find it.

I'm only asking that images not be deleted when not orphaned. Well, actually, I think that's a reasonable request and it is a practice I and many other admins choose to follow. It's not policy, though. I will ask about it on the mailing list (commons-l) and we will see.

However I strongly disagree that a better solution would be to replace images rather than delete them. Masses of extra work and double-checking is not going to help out our backlog any. And I am not sure why you see it as a better solution! Take Image:Vatican coa.png. It has been 'crossed' since 21 July. But it is still used in at least 5 wikis! Isn't this cross equally damaging to the projects, left in their articles, as a redlink from a deleted image? I don't understand why a deletion is more damaging. They seem the same to me. So... is being a Commons admin just doing this dirty work, manual labour? If so, I think you will understand why Commons admins don't support it!!

Thanks for this patient and calm post. pfctdayelise (translate?) 08:41, 7 August 2006 (UTC)

In the example, Image:Vatican coa.png, the projects using the image now have a direct link to the commons page (from where you can find the COM:DEL discussion.) While it is true, that the red X appears on the rendered project page equally as disruptive as a broken link, the linkage it contains is invaluable. --Connel MacKenzie 17:16, 7 August 2006 (UTC)

See my post to the mailing list Deletion of still-used images. pfctdayelise (translate?) 09:27, 7 August 2006 (UTC)

It is good good practice to delink images before deletion. So there should be no redlinks. But you claim also, that it is deficult to find out why the image was deleted. Well, not that difficult: You can copy and paste the image file name from the article history into Commons and then read the deletion log, which usually gives a hint, why the image was deleted. Or you can use "What links here" and you will surely find out. So, where is the problem? --ALE! 09:42, 7 August 2006 (UTC)

Actually, I have to agree, it is pretty difficult. On en.wp and I guess all projects, when you click on a redlink image, it actually takes you to the upload page. So first it takes some manouvering to check the logs. (For many users even this much would be impossible.) Then if it's not in the logs you guess it must be a Commons image. If you're not an admin, you won't see the nice deleted edits here, which can give a lot of information. Many reasons given in the deletion log are pretty cryptic. And if you are deleting from a discussion on COM:DEL, you can hardly summarise the whole discussion in the reason. So looking up the discussion, possibly in the archives, also takes some hunting. I think this is a fair criticism: there are definitely ways to improve the software to make all these things link together better. For example I think the local project deletion logs should also contain a link to Commons deletion logs. That would be a good start. pfctdayelise (translate?) 09:52, 7 August 2006 (UTC)

• ALE!, that is precisely the point. Most things in Wikimedia just work. How often do you cut-n-paste into a search box on a sister project? Even in IRC I don't do that; the links are clickable and go right to the proper page. But, using your method, I'm still not going to find a page that has been deleted. Finding the COM:DEL conversation from there (when you don't even know that the shortcut is "COM:DEL") is more than a little frustrating. Likewise, finding the deletion log. Don't forget, the mediawiki-savvy users will already be frustrated from having spent some time looking for the image on their local project. None of these steps should exist. When a person reading sister-project article encounters a red X, they can click it. With all those linkages in place, they will soon see the tag on the commons image page, and its link to COM:DEL (usually also with a sufficient explanation right there.) That person reading the entry can then click [back] and get to the broken article. From there, they can ask sensible questions and search for replacement images.

But when the image is simply deleted/missing, the scenario is different. Clicking the image link brings one to a local image upload page, without the filename. Cutting and pasting from the URL/location bar, a computer literate person can then get the url-encoded filename. Then, from memory, type http://commons.wikimedia.org/ (since there is no link, remember?) and paste the url-encoded filename into the search box. When the search finds nothing, the person can correct the url encoding and still not find the page. Cutting-and-pasting the title-name, they can then find "Special pages" and from there find "Logs" and from there find "Deletion log" and enter the proper title name in the third box on the right. At that point, they may discover (if every previous step was followed perfectly) only who deleted the image, but usually not why. Following the ever darkening trail, they end up at that users talk page, only to discover that it is a 'bot account "owned" by someone else. In a few more 10 second page loads, they may then find their way to some admins talk page. That admin will invariably have no recollection of deleting the image in question (how could they?) After a day or two, that admin will reply on their talk page with a link to COM:VP and COM:DEL. If the user (who has followed every step perfectly so far) manages to return to commons two days after that, (nearly a week elapsed in real-time, now,) they can finally find the original nomination for deletion. The original nomination for deletion was probably politely worded, and nominated with a very reasonable justification. But in the meantime, the sister project has not had its image (in fact, has had its pages mangled inexplicably) for over a week. An "undelete" request at this point will often fall on deaf ears; the time for discussion has long since passed. Sister project sysops (such as myself) may now be involved in the process, perhaps trying to find a replacement image. But they probably never saw the original image, and have only a rough idea of the replacement image desired. So they embark on the nightmarish foray through the commons: categories to find a new image. With luck, the sysop will be able to break out their camera, go to their local zoo, and take a picture of the Puma themselves. Then, they need to learn the commons' uploading license requirements, the commons' categorization requirements, and their own project's linking requirements to get the replacement image properly back in place. (That is, of course, assuming every step above is followed *perfectly*.)

Note also, the more common occurence: the drummer from a band finds a Wikipedia article about his band, but three out of five album covers are broken links. The drummer is high on heroin, and starts several flamewars on Wikipedia before being blocked. One week later, the same is repeated. Several months later, someone realizes that the stoner had a legitimate complaint to begin with...and then start the above detective work.

I understand perfectly, that it is a pain for commons: sysops to upload a small dummy file. It can add minutes to a routine deletion. It can slow a 'bot down to 1/10th its original speed. But that small upload will save multiple people hours of frustration, days/weeks/months of elapased time (where the final rendered page is broken, for all to see) and spread good karma far and wide. People previously unheard from, (perhaps the original photographers who don't understand mediawiki very well) would now suddenly have a voice on commons. This would facilitate the goal of encouraging sister project participation enormously. --Connel MacKenzie 17:16, 7 August 2006 (UTC)

I've got two words for everyone: CommonsTicker :) -Samulili 18:49, 7 August 2006 (UTC)

But, didn't I mention a couple times that I use CT? Let me say it again: I do use CT: how else would I have any idea about where stuff happens in Commons:? --Connel MacKenzie 08:20, 13 August 2006 (UTC)

I encourage everyone here to vote and comment on bugzilla:6909: "Redlinked images shouldn't go to Special:Upload" - I really feel this technical shortcoming is the major part of this problem. And also note some of the responses to my mailing list post such as Raimond's excellent point that we gain a lot of moral upper ground by being able to say, "We delete copyivos ASAP". If we delay deletion, we cannot say that. I still oppose totally the 'cross' method. pfctdayelise (translate?) 04:42, 8 August 2006 (UTC)

But, why? --Connel MacKenzie 08:20, 13 August 2006 (UTC)

Totally acknowledge User:Pfctdayelise! And with the regards to the CommonsTicker. I did not know it. Very useful. It should be shown on any edit and upload page for images. --ALE! 07:46, 8 August 2006 (UTC)

Connel MacKenzie's idea, applied literally, would in my opinion not work because it is not possible to see reverts onto image on ones watchlist. Within minutes, someone would have gone to the images and reverted the newly uploaded red X, and we wouldn't spot it.

However, it would theoretically be possible to upload a red X, then delete the old revision, leaving the image description page intact. I guess we could implement this system, if the admins are prepared to do it, and -- more important -- the projects support it. I can only speak for myself as an admin, and I am willing to do it when deleting images on the commons:deletion requests: if I don't have to orphan the image, it might save 30 min. / image.

Fred Chess 00:19, 9 August 2006 (UTC)

I did not know that file reversions behaved that way. That sounds like a separate bug. The second method you suggest seems very reasonable to me. You are able to restore deleted old revisions right? If so, then your method sounds optimal. --Connel MacKenzie 08:20, 13 August 2006 (UTC)

• Have you experimented with the "save 30 minutes per deletion" method you described, yet? Did you get positive results from that experiment? --Connel MacKenzie 07:11, 17 August 2006 (UTC)

[edit] Erin Silversmith's conduct

I didn't want to report this here, but I'm left with no alternative.

Erin Silversmith replaced Image:Yes check.svg with a new graphic that she created from scratch. I preferred the original image, but I didn't feel that it was my place to revert (thereby unilaterally declaring that the previous version was superior). Therefore, I decided to upload it under a different name: Image:Green check.svg.

A short time later, Erin deleted this image as a "duplicate." Obviously, it was not a duplicate, because it was no longer available for use under its original name.

I responded by politely requesting on Erin's talk page that she restore the image. This led to a discussion (partially quoted below) that she then removed in its entirety.

Erin indicated her unwillingness to allow both icons to co-exist (referring to my desire to do so as "idiotic"), and she reverted to the previous version of Image:Yes check.svg (implying that I was too stupid to figure out how to, despite my explicit indication that I didn't want to). She stated that "if [I] can't accept an improved version and insist on multiplying junk, then [she]'d rather the slightly inferior one remained as the sole version." Instead of restoring the proper description and licensing information, she added the statement "Gregory Maxwell's version has been reverted to to please User:Lifeisunfair."

In response to my request that she reduce the image's size by removing extraneous information (which she previously mentioned doing), she replied that she "won't make any more improvements to the image, as there is the risk that [I] will persist in reuploading old versions under new file names and [she doesn't] want to have to follow [me] around with the delete button."

Meanwhile, a related deletion discussion produced this exchange, in which Erin was similarly rude.

I don't know about the Commons, but this type of behavior isn't tolerated at the English Wikipedia. I'm an administrator there, and I would never abuse my tools in this manner (by unilaterally deleting an image in favor of a different one of my creation), nor would I dream of addressing another user with the incivility that Erin has displayed.

I've attempted to resolve this dispute amicably, and I'm posting this as a last resort. —Lifeisunfair 06:27, 31 July 2006 (UTC)

If Eryn's only complaint was that the old version was bloated, someone could go back to my orignal and change the stroke width by hand... My orignal was around the size of her improved version but someone later loaded my file into an editor to change the stroke width and in the process bloated the file. Really though, size shouldn't be a concern... we don't send the actual SVGs to the users, and someday when we do we will be sending them gzipped. --Gmaxwell 06:37, 31 July 2006 (UTC)

Erin believes that her rendition looks better. She described yours as containing "strange flat bits," but I don't see anything "strange" about them. In my opinion, her icon's shape is peculiar and unconventional. (Checkmarks usually aren't so rounded and uneven.) Nonetheless, I was more than willing to allow both files to remain, and I even allowed hers to take over the original name. To Erin, this was utterly unacceptable. She felt that the idea was "idiotic," so she reverted from her "superior" version to "please" me. —Lifeisunfair 06:50, 31 July 2006 (UTC)

I think she's only some kind of ignorant... have you ever tried to contact some of the other admins who could talk to her? Usually, normal admins shouldn't act like this... --Sven 20:23, 31 July 2006 (UTC)

In fact, I posted this in the hope that someone (perhaps an admin) would intervene. —Lifeisunfair 23:06, 31 July 2006 (UTC)

As I read it, the image policy says that you don't overwrite a file, unless it's a very minor correction. Re-drawing the image in a different style is not such a case, so overwriting the image is the wrong thing to do. However, jumping up and down about it is also the wrong thing; I'm going to upload Erin's version as Image:Green check.svg and let both versions exist. Alphax (talk) 11:14, 11 August 2006 (UTC)

I did not "[jump] up and down." I calmly and politely raised the issue on Erin's talk page, and I attempted in vain to reason with her until she removed the entire discussion. I turned here as a last resort, but it seems as though Commons sysops are free to behave however they please with impunity.

I wanted to allow the two images to co-exist, but Erin indicated that this idea was "idiotic" (preferring to remove her version instead), and she threatened to "follow [me] around with the delete button." —Lifeisunfair 00:30, 14 August 2006 (UTC)

In fact, Erin has nominated her own image for deletion, and she's pretending to be entirely unaware of the reason behind the new file's existence. —Lifeisunfair 02:32, 16 August 2006 (UTC)

Going by what is stated at Commons:Deletion guidelines one should not speedy delete images based on ones perception that it is superior to another. Personally I would say that artistical creations, such as "Yes"-signs, have all right to exist in multiple versions if they are different. But I can't view svg files on my computer and so I can't tell how they differ...

Fred Chess 21:09, 31 July 2006 (UTC)

One has a straight left edge, the other has a curved left edge; one has one slope, the other has another. Firefox 1.5 can view them. Alphax (talk) 11:14, 11 August 2006 (UTC)

[edit] Commons and multilingualism

In my opinion languages on commons are badly supported outside standard help pages. We have categories in different languages, mostly in english. Categories are limited through mediawiki's implementation but normal image pages for example are not. Everytime I upload a picture, I have to decide which language I should use. Recently I am deciding in favour of german, as it is my mother-tongue and there is no point in making understanding harder for my fellow wikipedians. I could write in different languages but I don't wanna be bothered with missing support or unreachable documentation. And please don't ask me to fix anything myself, cause I am only using commons instead of local (german) wikipedia. All I say is: I would help ppl with other languages to understand my pictures, if there would be an easy and satisfying possibility to do so. What is unsatisfying?

1. Documentation on the upload page how to specify descriptions in different languages.
2. Including different languages in general (technicaly impossible to tell which language is used cause no xml-lang tag is used).
3. Only displaying local language and ignoring other languages on display.
4. Handling of discussion threads in different languages ("Hey guys I don't understand what you are saying, please speak english")
5. Support of non-english-speaking ppl on pages like deletion request and so on. I might be wrong here, but a often heard complaint on the german version of village pump is that the requester needs help with something cause he doesn't speak english well enough.

I don't wanna blame anybody with this, just wanna point out some problems of an average user. If there are easy solutions for my problems, I would be thankful for any hints. --Chrislb 18:32, 31 July 2006 (UTC)

• There are varying Multilingualism philosophies. It is useful to keep in mind that not everyone has the same goals or approach when they discuss multilingualism.

Regarding point two, perhaps you are unaware of the way you can indicate language using a language template like so:

Français : Je parle français en petite peu.

• Issue- the more languages on a page, the more the visual clutter. Commons is on a bit of a grand adventure regarding omnilingualism. Consider the image Image:Hitler and Mussolini propaganda Luce photo.png. Let us assume the world 10 year from now, and each of these captions where hand created describing the particular scene of Hitler and Mussolini's meeting in Rome. (In this particular example, these actually aren't and are topic sentences copied from the wiki articles on Hitler in Mussolini from the respective language wikis). A question that immediately hits you is: What if most of the languages were represented? Look at the number of interwiki links on the sidebar. What if every single one of those languages had a caption? The description would be enormous.
  • One solution advocated by some is a text folding feature. Right now, if your browser supports stylesheets (nearly all do), then the individual users can customize their view of the articles to hide all text marked with the language templates (eg {{fr| text}} for languages they are not interested in. The problem with this is that it is a bit of a hack and is not a solution for casual visitors to commons. There are other solutions for text folding, but would require non trivial changes to the mediawiki software. Since this change would primarily only benefit Commons- and then only the few pages that have large numbers of multilingual passages, that positions the feature pretty low on the priority list.

• This is a frequently discussed subject so perhaps some of the old timers can fill in more here. -Mak 20:13, 31 July 2006 (UTC)

Hello Chris, Sorry for my late reply, but here are some thoughts. We don't want to add too much information to the Special:Upload page, because it becomes bloated and most users don't read it anyway. We have a Commons:Help desk (as well as this village pump) where you are welcome to ask questions.

As Mak said, the best way to add descriptions is to use language templates like {{en|this (insert caption here)}} -->

These templates have the SML language description stuff built in, so they should be easy to use. To show only certain languages using these tags, add this to Special:Mypage/monobook.css:

.description { display: none; } 	 
.en { display: block; }
.de { display: block; }

This would show only English and German tags. Keep in mind many people don't use the tags at all and if tags only exist in Spanish, for example, you wouldn't see anything at all.

You are frustrated at the multilingual support, I have to say we are too, and most if not all of the troubles are software/technical. (The majority of admins are not native English speakers.)

For discussion such as deletion requests in multiple languages, you are of course welcome to discuss in any language, but I don't see any way to support this better unless you suggest we all learn German, and Spanish, and Chinese, and Italian, and... If you want more German discussion I guess you simply have to kick the German-speaking admins up the butt and get them to help out more. You're welcome to do that. :P

I hope this helps a bit. pfctdayelise (translate?) 09:40, 7 August 2006 (UTC)

Thanks for the answer but I guess the developers have to work on that till everybody can be ok with it. As long as we have the state now we should urge everybody to learn more languages, what's not a bad point anyway ;) --Chrislb 06:15, 9 August 2006 (UTC)

[edit] August 3

[edit] Catalan claims

I have a disagreement with User:Martorell about category names. In spite of this rule, this user wants categories with catalan names and reverts some of my last changes ( example : http://commons.wikimedia.org/w/index.php?title=Category:Cities_and_villages_in_the_Land_of_Valencia&action=history ).

See the talks on my page and his page.

I think that Martorell is a catalan activist and desorganizes the project with partisan claims (he classifies Catalonia categories under french ones!). What do you think about this ? --Juiced lemon 20:15, 3 August 2006 (UTC)

[edit] Chauvinism claims

The only one difference that I see between rules applied to articles and rules applied to categories, is that "others" must be in English. In the rest of the case, I consider it's the same rules applied to articles. Am I wrong?. It's not desorganized, it's organized with other criteria. The actually desorganization is that you don't asked anyone before to do very wide changes, first to put template purposes and to talk in the discussion page. --Joanot Martorell ✉ 20:17, 3 August 2006 (UTC) PD: I'm not a Catalan activist, I'm a Valencian Wikimedian.

I don't need your permission to make changes, i. e. create more specific categories in order to classify haphazardly arranged images. You didn't ask me, and you have reverted my changes and put again images in imprecise categories (example : Girls in historical Valencian costumes.jpg (diff)). --Juiced lemon 22:27, 3 August 2006 (UTC)

It isn't about permission, it is about consensus. You came to a work already done that anyone found it any problem for a long time. Why do exist templates such {{Move-category}} or {{Move}} specially indicated for very populated categories? These aren't ornamental. Haphazardly arranged images were very few images, the category Category:País Valencià was organized well enough altough it was needed to do some improvings, specially in typos, and you done it, ok, very good for you. But you fell in cultural arrogance with me. There are a lot of toponyms in native form in categories. Here exist Category:Sevilla and not Category:Seville. The same matter for Category:Mallorca and not Category:Majorca. Or Category:A Coruña and not Category:Corunna. Or Category:Granada and not Category:Grenade. And anyone see it as a trouble, only you. Only in the case of non-English sentences it was seen as a problem, such as Category:Escuts de Suècia being moved to Category:Coats of arms of Sweden. --Joanot Martorell ✉ 22:54, 3 August 2006 (UTC)

PD: And what about moving[1] Alacant - Alicante to Alicante? It isn't a category, it's an article. According to language police it says specifically to respect native form. Another exemple is Bruxelles - Brussel and Brussels as a redirection. Your reasons against me are lacking in truth. --Joanot Martorell ✉ 23:46, 3 August 2006 (UTC)

As far as I understand, (and anyone knowing different please correct me), you can do as you please On article pages. Regarding Category names the two main bits of guidance are-

• use Ascii characters in category names
• use english language cat names (for now)

So to the point of this dispute- What about proper nouns? I came across the same issue regarding which place name to use when working on the By location category scheme. I don't know that there is a broad standard for this. Maybe that is wrong. I considerred the following approaches

• Choose the version used the most on the net (with the greatest search hits on google).-
  • advantages- uses the term that most users are most likely to use when searching for something.
  • disadvantages- It's majoritarianism, and if imposed broadly means that minority cultures are at an extreme disadvantage since the terminology of the majority will always squelch the minority.
• Choose the version on some online authority that allows input. This could be any of the wikis or even a nonwiki source, but since english was chosen for the category language, it makes sense that English WP be the authority.
  • advantages- Minority views have a chance because motivated individuals can get the title accepted on EN:WP by being persistent.
  • disadvantages- People may not be able to find the item since they are unfamiliar with the alternate spelling of the proper name. By insisting on local terms, the general public has a more difficult time finding information on the local places and culture that the advocate presumably is attempting to assist. Their actions have the opposite affect.

I chose option 2- go with EN:Wikipedia, and folks on that scheme went along with it. To be honest, there weren't a lot of folks there to argue with so I don't know how representative that decision is. Really, I could have gone along with the google scheme too, but I strongly prefer inclusiveness of minority views. Let the people fight it out there, and whoever wins, and whenever it flip flops, fine- we can rename the cats here. As far as Commons search goes, it doesn't matter, because for example you can make catalonia and catalunya go to the same place. That's what I wrote in the policy.

Any educated person knows that its Wien, Firenze, Venessia or Catalunya. Personally, "Catalonia" hurts my eyes and my ears, but that's what it is in En:WP, so tough break for me. If I felt strongly- I would argue the case on En:Wiki. -Mak 23:50, 3 August 2006 (UTC)

It's Commons, it isn't Wikipedia in English. It's a multilingual project. I'm agree that categories must be in English for sentences in order to a standard language. But, I suppose it's an eventual solution until a tech improvement multilingualism in categories. And, in adding, if we would read en:Wikipedia:What Wikipedia is not we were noticed that Wikipedia doesn't considers as a dictionary itself. Specifically it's telling that "Wikipedia is not in the business of saying how words, idioms, etc. should be used", so it shouldn't be the authority.

But proper nouns belongs to persons and their culture, and some persons aren't English-speakers. Would do you translate "Don Quijote" as "Sir Quixot"? Or "José Luis Rodriguez Zapatero" to "John Louis Rodericks Shoemakers"? I think it should be respected in native form, and using the {{Category redirect}} template for alternative categories.

Anyway, all the characters you find in "País Valencià" used are ASCII, as Category:José Luis Rodríguez Zapatero. Aren't these? --Joanot Martorell ✉ 23:56, 3 August 2006 (UTC)

Right. You mean the place would be called New York, not ca:Nova York. Heh heh. Small joke- I get your point and my personal opinion agrees with yours. If Chicago were translated from the Native American Potawatomi language, we would call it in English "Skunk Cabbage". Personally, I much prefer Chicago, and that people be faithful to to how things are named and pronounced locally.

We are making progress here on Commons. But in the meantime we will have to endure some of the absurd transformations that people make of local pronunciations and spellings. As I said- everyone is guilty- the catalunyan wiki names for "Venice" and "Florence" are just as awful as the english names. And you think catalunya is treated badly- Consider the awful approximations that non slavs make of slavic language names. "Khruixtxov" doesn't come anywhere close to the correct pronunciation- even though all of the sounds necessary are regularly made by speakers in catalunya. Even take the easy names- how about the capital city of Russia. Moscou? There should be a v sound in there (transcrit Moskvà)... Why not spell it the way it sounds. Nevermind- hard to break habits, unfamiliar consonant combinations, lazy mouths and all that I know.

In regards to Commons, we do not yet have a multilingual category translation but it is not unlikely that we will see it one day. In the meantime we have to pick one language for internally encoding the lookup names. What the display names are will be handled at display time- but natively, the database has to have a single encoding form. It so happens that our only display form is identical to the encoding form. But that will change. While we are only able to use english, to determine the encoding for proper names we must use some authority, and you must admit that using en:wiki is better than google as an authority. You are absolutely correct that en:wiki is not a dictionary but the wiki dictionaries currently have insufficient coverage- of course it will change but today, it just doesn't. I personally believe that an evolution of the wiki dictionary solve much of the single language user interface problem so that everyone can interact and display Commons category space in their own language.

For the time being we need to have a single language to map to, and there are much worse choices than english.

Regarding Ascii- you are conf