Eric Rescorla wrote:
> Eliot Lear <[EMAIL PROTECTED]> wrote:
>   
>> Eric Rescorla wrote:
>> That the password is at all related to the hash result at all is an
>> (IMHO) unnecessary risk that would in our scenarios impact more than a
>> single service.  There exists methods where this is NOT the case.
>>     
>
> Yes, there do. But they all involve lugging some object around,
> in which case the problem becomes vastly easier. We need 
> a system which doesn't require a token.
>   

That's not true, Eric.  Anything you can lug around can be "lugged"
around in software.  It doesn't solve the malware/bot problem, but the
two issues are separate and distinct.

Eliot

_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix

Reply via email to