jump,

Since you obviously saw the image, I have deleted it; being there for demonstration purposes only. I probably should delete the instructions, too, but those who want the images bad enough will find out how to get them anyway.

As Morrigan mentioned, if it can be viewed, it can be screen captured. Watermarking, IMHO, will only deter those who are unfamiliar with image touch-up techniques.

A few scripting possibilities that could be tried:

Have a script (presuming it is on a Unix/Linux based system due to potential processing performance) that does any/all of the following:

• Uses some coding that prevents the caching of the page/content on the client's system. This would prevent people from retrieving the image from their browser's cache. The HTTP 1.1 specification has an entire section (13) on Caching and Cache-Control Mechanisms.
• Places the watermark overlay automagically over the original uploaded image or embeds the watermark within the image itself and transfers that image to a separate location. This would mean only the image with the watermark would have any URL associated with it.
• Cuts the original image into a large number of small pieces (ie, 10x10 pixels) and recreates the image. This would deter anyone who did not want to "jigsaw" the image back together again.
• Uses only scripting to refer to the file. This would prevent people from finding out the image location, leaving them with only the watermarked image.
• Alters the binary image information producing a randomization of color replacements (ie, one time red becomes blue, next red becomes green, etc) and using the display script to correct the color associations. This would produce an altered image (the more hideous the better) as the source deterring people from saving it; however, this would probably consume a ton of processing power.
• Encrypt the binary image information with a random encryption/decryption scheme.

These are only a few possibilities I have thought of in my head after reading this thread. As I am not proficient in scripting, I cannot really say if any/all possibilities could even be accomplished through scripting alone.

As people tend to be "resourceful", I highly doubt that any methodology that is devised would ever be 100% foolproof. If the process is known and used by web masters, you can bet that others will find out and use that information to circumvent the protection. The most anyone can hope for is to make the process of obtaining the source image (or a copy thereof) so difficult and/or time-consuming that the person wanting the image will give up trying. An issue with that is some programming will allow the eventual retrieval process to be automated, therefore by-passing the difficulty and time-consumption hurdles. It is analogous to the software developer's dilemma of trying to protect their software from piracy.

I am not saying that image (or software) protection will never be accomplished, just trying to highlight how difficult it is going to be to find something that works the majority of the time.