It's a heck of a lot easier to shout non-unique names to identify someone though in most situations... I mean, I'm pretty sure no one wants to have to yell at their kid "76995cdc-9825-4b13-b66a-a60b26b6cc41! Stop drawing on the wall!"

The problem with typography on the Internet is the desire of font creators to limit distribution of their work (as the font files themselves are copyrighted in the US, and the font designs elsewhere). Thus, DRM is most likely inevitable in some form or other. This is why PDF obeys the 'no-embedding' bit in TTF, and has the option (if not the requirement) to embed only parts of fonts.

But why stick with a proprietary format? I always wondered what the problem would be with establishing some sort of private/public key signature/encryption method of DRM.

In this way, one would use a signature on the font to ensure that the font can only be used on one domain/rooted-URL and to also 'affix' some sort of source on the file (so that taking the raw font-file won't work elsewhere, and if the decrypted data is redistributed, the source domain/site is plainly visible) and would, furthermore, only be able to be decrypted with a one-time key unique to the session (transferred with SSL?). The end user is ALWAYS going to be able to theoretically pull out the decrypted TTF or rewrite the 'tag' on the decrypted TTF marking its original source, but you're never going to get around that problem in open-source implementations, as black boxes aren't going to be kosher either. At the very least though, you could build on the idea to make it difficult enough for others to crack without trying, and prosecute with the DMCA when they do...

I don't agree with this in the long term, but it's a better solution than a proprietary black box, and is perhaps a reasonable compromise for open-source implementation as well as meeting the rather restrictive demands of the font foundries...