Re: Confusion about Age: accuracy vs. safety

Roy T. Fielding writes:
 > > I can speak for myself here: as I am not as wise as you guys, and as I
 > > don't have much time for discussions, I did implement *exactly* (as
 > > faithfully as possible - see code below), Jeff's pseudo code.
 > 
 > Anselm, I believe that what you have implemented is just the calculation
 > of the Age upon receipt, right?  The question is: does the proxy code
 > add an Age header field (or to the value of an existing Age header)
 > when it forwards a message not from its own cache?
 > 
 > There is nothing wrong with the calculation of Age upon receipt.  The
 > problem would be if intermediaries added to the Age value (the age as
 > perceived by the outbound clients) when no aging occurs. If they do add
 > to the age, then Age becomes meaningless as a mechanism for ensuring a
 > lower bound in the presence of clock skew.

Oops, sorry for the confusion, you are right. I am getting tired these
days...

Anselm.