Re: the approach

    We have two initial states already:
      a) Current Practice
      b) What Roy Wrote
    
	[...]

    Insofar as What Roy Wrote is different
    from current practice, it suffers from the same difficulty; since it
    doesn't contain design rationale, it's hard to evaluate whether we
    believe the problems it was attempting to solve were really problems
    or whether the proposed new protocol elements solve the problem, etc.
    
We also seem to have some lack of shared understanding about
current practice (for caches).  I keep hoping that someone who
is deeply familiar with current practice would be willing to
write this down.  Not in a fully eloborated form, but just
enough for us to avoid generating incompatible changes (and
to avoid pointless arguments).

-Jeff

Received on Thursday, 11 January 1996 02:33:41 UTC