You can always create completely separate accounts and then use them over Tor one at a time blah blah blah, but that's a huge usability failure. What you really want is a UI that allows you to sign in to multiple accounts without the server knowing they're all the same person (which means native software, probably with something like Tor automatically using separate circuits for each account), and then allows you to seamlessly receive messages for any of them and clearly delineates which account you're sending from when you post something.
Most of the time, it would be enough to have some sort of unified Google login (in a Google iframe or whatever), offered as a service to different sites on the Web, and different pseudonyms available that __don't reveal your 'true' ID or other pseudonyms to the site owner__.
I'm a little surprised that Google haven't offered some kind of 'Google Identity' service to the web at large based on this sort of idea. It wouldn't be good enough for activists in oppressive regimes, but it would offer a convenience to general people who like to keep separate identities on different sites.
The problem, I think, is that, given the current "don't be evil, just creepily self-serving!" attitude that seems to guide Google, no one would trust them. Even if one could trust them to keep one's identity private at launch, I am sure that anything like this would come loaded with the kind of "T&C may change without notice at any time" codicil that would allow them to permit themselves to disclose your identity later.
> different pseudonyms available that __don't reveal your 'true' ID or other pseudonyms to the site owner__
(which isn't offered by Facebook—right?—and which does require trust in the issuer) would be part of the putative Google Identity.
On the other hand, having spent this much time arguing that it wouldn't happen, I'm now wondering why OpenID with Google as 'vouching agent' (or whatever that part is called) doesn't count.
That's quite an understatement.
The other aspect, is the lack of transparency Google has when making policy and product changes. Like suddenly trying to change app reviews and youtube comments from a pseudonymous to a real-name activity. Without any confidence in their service and policy stewardship, it's hard to trust them not to change this stuff randomly too.
1. You have a canonical URL that identifies the user.
2. The user can have any name they want, and the name will only change if you are in a chat room, adding a (2) if they are a duplicate name.
3. Finally, you can add tags and a nickname (separately) so you can always see whatever name YOU assign to the person and the person in question can represent themselves anyway they want.
Steam is one of the few social networks I participate in because I truly believe it adds some value to the basic digital distribution model. Say what you will about the ability to resell games or DRM or what have you, it is a compelling product.
If they had that from the start and later decided to remove it, people would go insane.