a few random thoughts

[<DEVANS%COLOLASP.BITNET@CUNYVM.CUNY.EDU>]

Gosh, maybe I'm not such a stoop after all. Most of what Brian says
I have thought about myself, but been too intimidated by the masters
to bring up. Especially the assumption that packets are lost only due
to collisions, whereas it is my experience that while that happens,
it sure isn't the mechanism responsible for those multi-minute
backoffs. I like Brian's idea about pP (I hadn't thought of that
one!), but I need to think a bit more about it before I'm really convinced
about it...as a frien of mine says...can 10,000 lemmings be wrong?

Mostly, my thoughts, such as they are, can be summed up by saying
that most of the routing and timing algorithms we use were definitely
NOT designed for radio, and it is highly unlikely that they are
remotely optimal for radio as we use it today (that may change in
the future); consequently, it is only smart to do as much experimenting
as we can with all aspects of routing and timing. The "optimal"
algorithms and parameters will likely seem obvious after we find 'em,
but they sure don't look obvious right now.

73 -- Doc