Re: Loops (RE: CPIM changes)

Mark,


Wednesday, November 13, 2002, 11:49:45 AM, you wrote:
Mark> But I am trying to contrast a "sound architecture" with an "easy to adopt
Mark> architecture".

In my experience, they are the same thing, or at least very tightly related.


Mark>  It would seem like a bad
Mark> thing to (say) adopt current bit patterns used by X or Y

Mark, we have a very basic disconnect about the nature of gatewaying and
nature of what I have been saying.  I never said anything even remotely like
"adopt current bit patterns".  This is not an effort to standardize AOL or
MSN.  However interconnecting with them IS a goal.

Interconnection among independent, heterogeneous services requires common
semantics, not common syntax.  (Common syntax is always convenient, but not
essential.)

The more of the semantics that are common, the better.  However real
experience with networking says that it is essential to focus on the
smallest possible amount.  Requiring more than that minimum serves to
exclude participant services.

To be specific:  I have been consistently suggesting that propagation of
specific bit patterns should absolutely minimized, for the gateway task.
Obviously, it is the essence of doing the Homogeneous Relay task, but it
utterly defeats the Heterogeneous Gateway task.

d/
-- 
 Dave Crocker  <mailto:[email protected]>
 TribalWise <http://www.tribalwise.com>
 t +1.408.246.8253; f +1.408.850.1850




  [reminder: [email protected] for non-technical discussions, please]