Password-Protected PDFs

[flouran]

Hi,
I noticed that when I password protect a PDF from being copied or printed in Adobe Acrobat, evince bypasses these restrictions. Is it possible at all to password protect a PDF in evince or somehow password protect a PDF such that no other software can bypass the restrictions set in place?

Thanks,

[CRGreathouse]

There are several types of password protection for PDFs.

First there are strong encryptions. You can't open the PDF at all unless you know the password, and if the password is strong the file is essentially impossible to hack (barring advances in cryptographic technology).

Second there are weak encryptions (40-bit?). These are like the above, but easier to break. These were once standard but now very easy to break. It's possible to cycle through the possible hashes until the right one is found, regardless of the length of the password. Using rainbow tables, these files can always be cracked in a matter of hours.

Finally there are advisory passwords. The contents of the file are not encrypted (though they are obfuscated slightly), since otherwise the file couldn't be read. The PDF program is asked to disallow printing, or modifying, or the like. Of course this can't be enforced, thus my name for this type of protection.

With this background, I don't think I need to actually explain the answer!

[flouran]

I am using what you dub an 'advisory password'. I am distributing a pdf to other people whom I am weary to trust, and thus I do not want them copying or printing the contents of the pdf. In other words, they can look but not touch. The encryption I am using is 128-bit AES. However, evince simply bypasses this restriction. The reason (I think) is that evince doesn't care for advisory passwords; however, evince could most definitely *not* open a document which requires a password to open unless the password is entered.

[frmky]

Ghostscript/GSview also ignores "password protected" restrictions on printing and copying PDFs. It's easy to even "print" to a non-password-protected version of the PDF that can then freely be used in Acrobat. I have on occasion taken advantage of this feature.

Of course, what you are seeking is fundamentally impossible. If they can see the content, then they can image what they see using a screen grab. They can then freely print the images. If they want copyable text, Acrobat includes surprisingly good OCR software.

[CRGreathouse]

Originally Posted by flouran

I am using what you dub an 'advisory password'. I am distributing a pdf to other people whom I am weary to trust, and thus I do not want them copying or printing the contents of the pdf. In other words, they can look but not touch. The encryption I am using is 128-bit AES. However, evince simply bypasses this restriction. The reason (I think) is that evince doesn't care for advisory passwords; however, evince could most definitely *not* open a document which requires a password to open unless the password is entered.

The contents of your file are *not* encrypted: if they were, those untrusted people could not read your file. But what you want (as frmky pointed out and I already explained) is technically impossible: if they have the information, you can't stop them from using it. If they have enough information to display it onscreen they have enough information to print it.

You can include information in the file asking programs not to do certain things, but you can't stop them. Further, given such a 'password-protected file', it's easy to produce another file with the same information but no password, so that even printing from Acrobat is possible. (It's not quite so easy that you could do it with a hex editor and scratch paper, but nearly so.)

[flouran]

Originally Posted by CRGreathouse

The contents of your file are *not* encrypted: if they were, those untrusted people could not read your file. But what you want (as frmky pointed out and I already explained) is technically impossible: if they have the information, you can't stop them from using it. If they have enough information to display it onscreen they have enough information to print it.

You can include information in the file asking programs not to do certain things, but you can't stop them. Further, given such a 'password-protected file', it's easy to produce another file with the same information but no password, so that even printing from Acrobat is possible. (It's not quite so easy that you could do it with a hex editor and scratch paper, but nearly so.)

I'm not surprised.

[Uncwilly]

What about making a video slideshow of the info on a DVD? Once you burn it as a regular DVD, most computers will display it, but make copying the view cumbersome at best.

Send it out on VHS.....
35mm

[CRGreathouse]

So get a digital camera.

Certainly the process can be made more difficult, but you can't make it impossible. If the user is presented with the information, it's available for whatever end.

[cheesehead]

Originally Posted by CRGreathouse

So get a digital camera.

... or take out that SX-70 resting on your closet shelf, get a couple of film packs from the resurrected Polaroid instant film factory (http://www.nytimes.com/2009/05/26/te...laroid.html?em), and warm up your digital scanner.

[joblack]

Originally Posted by CRGreathouse

The contents of your file are *not* encrypted: if they were, those untrusted people could not read your file.

You can encrypt the content - it depends which protection scheme you're using. Most of the password protected content uses older pdf versions so you can use it with older Acrobat Reader.

I doubt that you can crack pdf-1.7 files without knowing the key (e.g. password).

[CRGreathouse]

Originally Posted by joblack

You can encrypt the content - it depends which protection scheme you're using.

Right, I'm well aware of the schemes. But any PDF that can be opened without a password can be printed without a password -- the only thing stopping you is software choosing to listen to the PDF's recommendation that it not be printed.

This is fundamentally different from an encrypted PDF, where it's not obvious how to get the content at all. (As I mentioned, though, the older schemes are easy to break with rainbow tables -- it takes a few hours, tops. Even without rainbow tables it can be done in a few days on a slow laptop.)