just a thought, how will tincan handle manipulation of messages by malicious clients?
the messages are stored in a encrypted lump. your signature will be unique. Even if both of us want to be called steve tin-can will keep us separate. by assigning a 2 utf-16 suffix on your tag. I may be steve-山蛋 and you steve-산ಡಿ (steve mountian eggs and steve acid D if any one is playing) the suffixes are optionally hidden
no centralized database. When you name yourself you your suffix is a hash.
Great question
the outcome would be spoofing - writing messages and miss-attributing them (can you still do that on email?)
The user name is folded in with message so that would be a bit harder do than in email were one just has to change some text.
accidental collisions will of course be rare although the soccer pitch problem
4294967295 /4294967296 * 4294967294 /4294967296 ... (4294967296 - n - 1)/4294967296 should we tack on another char?
Exactly accidental collisions even if the hash was random would be rare there would have to be more than 2 billion steves before the chance of collision became likely. as it stands tin-can is more spoof resistant than email.
The signature is not in plain text standard editors would not be effective in changing it.
It is assigned at installation making forced selection quite a bit of a chore. I need to thank you for this line of questions it is making me think that the encryption and the hash should perhaps not be made public.