gnu@l5.uucp (John Gilmore) (10/08/85)
Erik, I don't understand how your "frequent calling" method produces lower phone bills. Every message has to go between you and every news neighbor exactly once. Forwarding news quicker will change the ratio of news that you receive to news that you send, but it will all still have to traverse a phone line in one direction or the other.
fair@ucbarpa.BERKELEY.EDU (Erik E. Fair) (10/12/85)
In article <184@l5.uucp> gnu@l5.uucp (John Gilmore) writes: >Erik, I don't understand how your "frequent calling" method produces >lower phone bills. Every message has to go between you and every news >neighbor exactly once. Forwarding news quicker will change the ratio >of news that you receive to news that you send, but it will all still >have to traverse a phone line in one direction or the other. What exactly didn't you understand about the explanation? The key point to realize is that when your system says `Duplicate article <1234@frobozz.UUCP> rejected', and frobozz is one of your n full-feed neighbors (where n > 1), the article in question has traversed the link between you and frobozz not once, but TWICE (once from frobozz to you, and once from you to frobozz) because both of you got the article from other sites within a window when you and frobozz weren't communicating. The more often you communicate, the smaller this window is, and therefore, the lower the probability that articles between you and frobozz will go over the link twice. In short, your neighbors don't know that you have the article unless you tell them (by sending on the article), and the longer the period of time that you have an article without telling them, the larger the probability is that they will get the article from elsewhere, and tell you that they have the article (by sending on to you when you already have it [and have probably queued it for delivery to them]). Does this clear things up? Or have I further muddied the waters? Erik E. Fair ucbvax!fair fair@ucbarpa.BERKELEY.EDU