Re: Greetings and comments on MCP 2.1

[Michael Brundage <brundage@piaf.ipac.caltech.edu>]

On Thu, 7 Aug 1997, Erik Ostrom wrote:
> > A server driven approach (or a client driven approach for that matter)
> > has the benefit of reducing the amount of traffic to shift before the
> > session can begin passing normal messages.
> 
> Yes, but we're talking about bandwidth vs latency here.  In a perfect
> world we would have empirical evidence to help us make our decisions,
> but for now, it's just my intuition that the cost of, in your example,
> sending six extra lines from the client is less than the cost of
> waiting for the lines from the server to show up.

Well, we do have empirical evidence of some sort; Jupiter of course, plus 
the clients you're all doing and ours too.  But I agree that the 
bandwidth is small in the grand scheme of things; feature negotiation 
(in any form) is swamped by the traffic for updating GUI components, even 
high level ones.  For me, having a very flexible, growable, dynamic 
scheme is more important than the having the most efficient one possible.


michael