mo@seismo.CSS.GOV (Mike O'Dell) (12/06/85)
For the person complaining about wanting backbones yesterday so damn the engineering, as a person working on a possible backbone switch, I can tell him that many of the existing problems stem from not doing the necessary engineering and simply wishing at the problems. The currently-used protocols were never designed to be used over multi-access media and work correspondingly well. There are things that can be done to the protocols to improve their behavior, and that is being looked at. Another very fruitful approach would be to use full-duplex RF repeaters, instead of simplex digipeaters, but everyone says that is technically difficult and economically impossible. It is more likely challenging in both areas, ignoring the geopolitics of voice versus data bias in band plans and "coordination committees." I must also mention the Phil Karns and Terry Foxes and the other people who are trying to do the real work. The number of people working is small; the problems are very difficult, given the irrational shape of the feasible solution space; and perfection certainly isn't the goal - but simply spraying bits around and claiming "gee, look what I can do with my new TNC and my 2m HT", and calling it "packet radio" while utterly ingoring the engineering issues is not worthy of the enterprise which has done other fine work like transforming SSB from mathematics on the blackboard into an operational reality. A great deal is known about the problems which must be solved, and try as we might, the laws of physics are not subject to supplication, at least in my neck of the woods. -Mike O'Dell (learning the bloody code in spite of his better judgement)