If I wanted to send a document, or more specifically, a picture to someone, but I wanted to makre sure that they couldn’t keep it or reproduce it or post it publicly (unless they acted really really fast, and I know that super slick people aren’t going to be thwarted, but I’m talking about jsut an average person), are there any kinds of software that will make the document self-erase or something? (Especially Mac software…)
Not really, no. You can’t both give something to someone and not give something to someone. People always seem to forget that computers run entirely on making copies, not inscrutable Zen koans.
Not by computer, no. At least not without tamper-proof hardware and software on the receiving end that just did that.
Which can be fully and finally defeated with a good camera and possibly a tape recorder.
The poor man’s solution to this problem is to display the picture as an avatar on Yahoo Messenger or MSN Messenger.
Braaaaaaaaiiin implants. No monitor, just straight into your head. 
I saw a prototype monitor the other day that could only be easily read with the supplied set of glasses - the monitor would show the real image, then alternate it with the negative - very, very fast - so that an ordinary observer would just see a blank grey screen, but an observer with the glasses (that switch out the negative frames) would see the picture.
I expect that could be defeated with the right kind of photographic equipment, but I think it could be developed so as to be quite difficult - for example, it could be made to specifically take advantage of human persistence of vision (so the camera, unless carefully tuned, would smear or miss parts of the picture), and the blanking interval could be dynamically changed so that a camera would have trouble keeping in sync.
If I understand you right, any half savvy user will print the screen and take a copy that way. To be fair, this can be done with anything that ever comes up on screen - so is a good reason why you shouldn’t send anything you wouldn’t want to be copied and potentially distributed.
A user with more savvy would disable anything preventing him from saving the image directly, or go into the browser’s cache directory until he found the file he wanted. Really, circumventing these things is so simple it barely justifies being called a trick.
This would be a very simply programming task. Just make an App which displays the photo, and deletes itself when done.
There is a technology called either IRM or DRM (information/digital rights management) that is intended to (mostly) solve this problem. See http://www.microsoft.com/windowsserver2003/technologies/rightsmgmt/default.mspx for an example.
The basic idea is everything is always encrypted and the recipient can only decrypt by asking your server for the key each time they need to process the document. That [get key & decrypt] process is built into the OS & user software so the naive end user doesn’t really know or care whether the document is protected or not.
There are guts in the IRM module in the OS which let the key server say “printing is OK until 9/28, viewing is OK until 10/15, and forwarding or saving is never OK.” And all software on the machine must abide by those rules; the OS enforces them.
As folks upthread have said, this setup can protect up until the moment the document leaves the digital realm. Once it’s an image on a screen they can’t prevent you pointing a camera at the screen & getting a crappy but unprotected copy. Ditto for audio; once it’s piped to speakers, you can record a reasonably good copy with a microphone and recorder. If you’re allowed to print it once, you can then photocopy a million copies, or just scan it back in.
Even within the digital realm, there is great effort being spent to try to figure out how to make all this work robustly when the bad guys have computers that may have their IRM/DRM kernel compromised.
As a practical matter for a home user who wants to protect a doc sent to another home user, no good tools exist today. The real answer lies is in the human realm, not the digital.
Have them give you a $10,000 deposit which you will refund in 10 years, with interest, only if the document never appears in public during that time. If they aren’t willing, perhaps they can’t be trusted. Obviously the size of the deposit needs to be related to your economic harm if the document leaks & their value gained from access to a non-public document.
Seems like that could be defeated in several different ways - for example, taking a screenshot, taking a photograph of the screen, running the program in a virtual computer that doesn’t permit it to delete itself, taking a copy of the program before running it…
The OP SAID it didn’t have to be foolproof! He just wanted a way to provide the casual user with a one-time use photo.
BTW, it’s possible to disable screen capture in “OS X” by software control.
I just want to point out that the great majority of Dopers are WAY more computer savvy than the great majority of users, who have no idea why nothing comes out of the printer when they press the PrintScrn button.
Ever receive a fax with a disclaimer on the bottom?:
Would something like this work? 
Personally, when I get a “wrong number” fax, I immediately disclose, copy, distribute, and misdirect the contents and I’ve never been called to task.
To stop someone reproducing something publicly, you could always watermark the image with the intended recipient’s private details. No one is going to stick that embarrassing photo of you on the company bulletin board if it has their name, address, email, phone and names of their kids written across it.
OK, but there are still problems - getting your self-deleting application to the other end (many email servers just delete executable attachments), getting it to run at the other end (what OS is the recipient running)…
It’s not just a matter of the user being computer-savvy - it’s a matter of whether the user expects the image to be tricked up for one-time display beforehand. If not warned, even the very savvy user may look at it, then dismiss it without capturing the image or anything like that because they don’t expect it to behave that way.
Even then, you’ll want to think hard about the mechanism. If you run an app which simply deletes itself, the app can probably be recovered through any disk recovery program. If the image winds up displayed in a browser, a copy may be in the browser cache.
Off the top of my head, what I’d do is provide a link to a page containing the picture on my own server. The URL would include a long key that I could check for validity, and that the user wouldn’t have an easy means to generate another of. I’d use the key on the server end to prevent serving the page more than once, and I would send it back with appropriate headers to prevent browser caching. The rub comes in preventing it caching an embedded .jpg. You would have to make into some sort of presentation that wouldn’t be cached.
I think the obvious answer is it’s IMPOSSIBLE to prevent a sufficiently savvy computer user from being able to copy data that is sent to him. Look at all the broken DRM protection schemes out there. Depending on the OP’s level of paranoia, it might be possible to come up with a scheme that provides enough protection to satisfy him. Note that Watermarks are not the solution - Visible ones can be destroyed with Photoshop, and invisible ones simply let the owner track the photo back to the “leaker.”
And since it’s just about impossible to prevent simple screen capture solutions, you probably do not want to TELL the person up front that the image is “for one time viewing” or anything like that. That’s a red flag telling them to capture it. Continuing on my suggestion above, the mail would simply contain a link to something like:
http : //hypotheticalserver.com/photoshare?ID=Foh3UoI89Relk17U9ffXsaDFr8T78Uewqu7
(space added to keep VB from presenting this as an actual link. That I don’t mind - a link to a nonexistent server is harmless - so much as the fact that it truncated the display of it.)
without any warning as to its behavior. They click on it once, and view and dismiss it, presuming that they can do so again. To further the illusion, I wouldn’t even tip them off on the retry - I’d hand them a garden variety 401 error page, and not answer any mail about it. That way, they assume I’m just running an unreliable server, and not being responsive to email. That way, I might even be able to send them another “one shot” picture without them twigging.
Of course, a side effect of trying to conceal the “one shot” feature from the recipient is that you are royally pissing them off.