![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Tuesday 10 February 2004, Jeremy A. Greene wrote: > Username/pw can be tied to their existing ones (like existing aaa pw > or through MS domain etc.). And username/pws tend to be easier to > remember and update. And a md5 key is not the easiest thing to > remember or type in. Anyway, it seems that it's a more traditional > approach that people have already dealt with in one way or another. > This seems like another thing to deal with. I see no reason why a one-time username/password (or a pre-existing one, if it is deemed strong enough), cannot be run through a hash function to give you your initial AAAH-MN key. As long as the hash function is the same at both ends, the entropy of the username/password used as input is sufficient, and the method of distributing the username/password is deemed secure enough for the application in hand, you're home free. There's no reason to bother a user with entering a string of hex digits, for instance. Henrik
Attachment:
pgp00083.pgp
Description: PGP signature