Re: Still trying to make sense of HTTP caching model

Jeff said:
] Actually, I would probably combine #1 and #2 into
] 	Expires: value < deskewed current time
] given that the Date: value serves as a lower bound on the "current time"
] when it is received.  When I wrote "Expires: <yesterday>" what I really
] meant was "Expires: some time obviously in the past"; I don't think
] we need to pick a particular distinguished value.  In fact, we probably
] ought not to pick a particular value, since it can only lead to one
] more possible bug in an implementation.

I was listing the three ways that it seemed the spec would allow based 
on current list thinking.  I was objecting to there being three ways.

I agree with combining 1 & 2, and eliminating "max-age" -- it is 
completely redundant.

That would also get rid of any ciunter=intuitive "Expires: <yesterday>" 
distinguished value.

Paul