Re: Single model or framework for reputation/trust

der Mouse <[email protected]> Tue, 4 May 2004 16:16:51 -0400 (EDT)
Newsgroups gmane.ietf.asrg.smtpverify
Message-ID <[email protected]>
>>         Reputation is a strong indicator of future behavior.
> Not in the real-world.  Reputation is a measure of past behavior.

Any system based on doing things differently for different hosts is
assuming that past behaviour is an indication of future behaviour[%].
Thus, for practical purposes, reputation, which is a measure (a
relatively rough measure, but still) of past behaviour, is valuable as
an indicator of future behaviour.  Not a perfect indicator, true - but
significantly better than chance, which is all that's necessary for it
to be useful.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       [email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B

[%] Actually, this isn't quite true: for example, accepting user
    submissions for your own dialup pool has nothing to do with past
    behaviour being an indicator of future behaviour.  My comment
    really applies only to arm's-length systems.