From: utzoo!decvax!ucbvax!CAD:tektronix!zehntel!sytek!menlo70!hao!cires!nbires!crp Newsgroups: net.followup Title: Re: people who need mail paths - (nf) Article-I.D.: nbires.113 Posted: Thu Mar 17 22:52:09 1983 Received: Wed Mar 23 22:02:32 1983 References: ihlpf.109 Speaking of accounting and charging for net costs... I would like to offer the opinion that usenet is as successful as it is (and you may take your own estimate of its success) in large part due to the fact that it is not a budget line item at almost any site and that news and mail forwarding are provided gratis. It certainly isn't as efficient as it could be; auto routing based on link cost, for example, could make it much more cost effective (as well as more convenient). On the other hand, I think the net would not exist if it had to be tremendously efficient and services had to be specifically paid for. Consider, those who deal in money things, what it would take to convince your company or school to pay another site actual real money for such a nebulous service as net.general or mail forwarding. Automobiles are a terribly inefficient form of transport, but the auto will survive for a long time because it is tremendously convenient -- and that is worth a lot to an individual. Calling long distance on demand during the day is exactly the same sort of convenient inefficiency and it will survive too. Note that some sites bear a disproportionate amount of network traffic and phone costs and thus help to make usenet work. Two sites leap to mind - decvax and ucbvax (not to ignore others). I think it is appropriate to say "thank you" to these sites and their administrators for helping to make the usenet electronic community a reality. As a closing note, I *do* think that network efficiency is a concern now that the network is a reality. The network has reached a point where mail, at least, is very unwieldy and most mail no doubt travels further than it would need to if the sender understood the way all of the sites were connected together. A transport system which would perform automatic routing based on knowledge of node connections, node traffic, and link cost is non-trivial. Only if someone is willing to build such a system (Mark?????) gratis will it get done in any complete way. Sites might be willing to pay a nominal amount for such a piece of software, but probably not enough to make it a business proposition. Charlie Price -- NBI (The Word Processing Folks!) {allegra|princeton|ucbvax}!nbires!crp