greid@adobe.UUCP (Glenn Reid) (10/31/85)
The NET: The net relies on UUCP links. UUCP links rely on phone lines and the existence of special software on each end to transfer information. These phone lines and UUCP links have existed a lot longer than USENET. Many of them are long-distance and were set up to facilitate the exchange of useful technical information among computer students and professionals. Somebody thought of "newsgroups", and discovered that the UUCP transport system could be used to distribute such text in a manner similar to electronic mail, but using the "chain-mail" approach to distribution. News works great. The problem is that its life blood depends upon the existence of UUCP links which are very likely established for entirely different reasons, and which may become newsfeeds as a matter of courtesy. The problem now is that too much news and not enough technical information is passing along some of these links, and something needs to be done about it. Some of the newsgroups are arguably technical, but the entire system is arguably distinct from the mail system and the "raison d'existence" of many of the UUCP links. However, it is handy to have news and mail travel the same paths, for many many reasons. I think that a completely separate connectivity should be gradually built for the USENET news system, with minimal centralized control, in such a way that the all-or-nothing scenario which is cropping up is not a problem. There is no reason to destroy net.anything other than for the sake of phone bills, and most of the problem with that is historical. We should work to improve that. Another thought: have an "option" to news that allows a site to turn off a newsgroup with the following behavior: Any article posted to it that hits that site will be returned to the sender informing him that the news got no farther than that particular site. That should discourage people from posting the value of pi. Or any article > n bytes. Or whatever. Let's have a few constructive ideas, people. All I see in this group is one of the following arguments: 1) The net is not free and there are a few great guys doing a great job and I support them 100%. 2) The net and the whole world are free and egalitarian and I happen to like net.movies and I think I may just ignore that rmgroup and I pay very few phonebills, myself. 3) I have a great idea about how to reconnect everybody's computers that will make all this a moot point. :-) Glenn Reid decwrl!adobe!greid -- You have new mail.
wls@astrovax.UUCP (William L. Sebok) (11/02/85)
In article <787@adobe.UUCP> greid@adobe.UUCP (Glenn Reid) writes: >Another thought: have an "option" to news that allows a site to turn >off a newsgroup with the following behavior: Any article posted to it >that hits that site will be returned to the sender informing him that >the news got no farther than that particular site. That should discourage >people from posting the value of pi. Or any article > n bytes. Or >whatever. This cure would be worse than the disease. I can envision a flood of such returned articles overwhelming the mail links. If the article is mailed back along the arrival path then a site that does this article returning has just doubled the cost of the article to all of the intermediate sites. If instead the article is mailed along an optimized path: the average mail link tends to be more expensive than the average news link (which is more often local), and the article returning site has just flooded some expensive mail links with news articles. -- Bill Sebok Princeton University, Astrophysics {allegra,akgua,cbosgd,decvax,ihnp4,noao,philabs,princeton,vax135}!astrovax!wls
joel@gould9.UUCP (Joel West) (11/02/85)
In article <787@adobe.UUCP>, greid@adobe.UUCP (Glenn Reid) writes: > I think that a completely separate connectivity should be gradually built > for the USENET news system, with minimal centralized control, in such a > way that the all-or-nothing scenario which is cropping up is not a problem. It would help. I'm not sure this isn't being done already. But if backbones are bitching about paying the feed for some groups, they should only feed those groups to "free" sites (ie, local or dedicated lines) and let their long-distance connections find another path to get net.flame.all. Eventually, only the people who want to pay the bill would be paying for it. > Let's have a few constructive ideas, people. All I see in > this group is one of the following arguments: > 1) The net is not free and there are a few great guys doing a great > job and I support them 100%. > 2) The net and the whole world are free and egalitarian and I happen > to like net.movies and I think I may just ignore that rmgroup and > I pay very few phonebills, myself. > Glenn Reid Amen. It seems that there's no meeting of the minds between the net-anarchists and net-gods, so it makes it difficult for net-moderates to reach a consensus. As a result, each do what they want to do and the rest of us suffer. Joel West CACI, Inc. Federal {cbosgd,ihnp4,pyramid,sdcsvax,ucla-cs} !gould9!joel {sun!suntan,decvax!sdcsvax} gould9!joel@nosc.ARPA The ultimate form of recent anarchy was cancelling someone else's posting!
ron@brl-sem.ARPA (Ron Natalie <ron>) (11/03/85)
> > Another thought: have an "option" to news that allows a site to turn > off a newsgroup with the following behavior: Any article posted to it > that hits that site will be returned to the sender informing him that > the news got no farther than that particular site. No, no, no, no. It's bad enough that I get letters from idiot ARPANET sites that don't obey their own protocols and manage to mail me failed mail because some user's mailbox was over quota. I don't need a shit load of "your mail has been blocked" messages from everyone who installs a restrictive forwarder. I have no control on where my message gets distributed. Ideally now, it goes anywhere, but restrictions on distribution should be handled silently as they are now. > people from posting the value of pi. Or any article > n bytes. Or > whatever. Let's have a few constructive ideas, people. All I see in > this group is one of the following arguments: To really screen lists appropriately, you're going to have to moderate them. You'll have to balance these advantages with the problems of a moderated list. -Ron
friesen@psivax.UUCP (Stanley Friesen) (11/04/85)
In article <173@gould9.UUCP> joel@gould9.UUCP (Joel West) writes: > >It would help. I'm not sure this isn't being done already. >But if backbones are bitching about paying the feed for some groups, >they should only feed those groups to "free" sites (ie, local or >dedicated lines) and let their long-distance connections find >another path to get net.flame.all. Eventually, only the people >who want to pay the bill would be paying for it. > Actually this can be carried further, after all "local areras" *overlap*, so it should be possible to develope a chain of local, or at least low-cost zone, calls all the way across the country. Then there would be little need for the expensive long distance calls at all! Perhaps the European system of having the *recieving* site pay for the news if it is sent over a long distance feed would be good. -- Sarima (Stanley Friesen) UUCP: {ttidca|ihnp4|sdcrdcf|quad1|nrcvax|bellcore|logico}!psivax!friesen ARPA: ttidca!psivax!friesen@rand-unix.arpa
herbie@polaris.UUCP (Herb Chong) (11/09/85)
In article <830@psivax.UUCP> friesen@psivax.UUCP (Stanley Friesen) writes: > Actually this can be carried further, after all "local areras" >*overlap*, so it should be possible to develope a chain of local, or >at least low-cost zone, calls all the way across the country. is the added propogation delay worth it? remember that the primary reason for the backbone sites is the reduce the propogation delay as well as to ensure that a large group of sites have a decent chance of getting ALL news. Herb Chong... I'm still user-friendly -- I don't byte, I nybble.... VNET,BITNET,NETNORTH,EARN: HERBIE AT YKTVMH UUCP: {allegra|cbosgd|cmcl2|decvax|ihnp4|seismo}!philabs!polaris!herbie CSNET: herbie.yktvmh@ibm-sj.csnet ARPA: herbie.yktvmh.ibm-sj.csnet@csnet-relay.arpa ======================================================================== DISCLAIMER: what you just read was produced by pouring lukewarm tea for 42 seconds onto 9 people chained to 6 Ouiji boards.
friesen@psivax.UUCP (Stanley Friesen) (11/11/85)
In article <260@polaris.UUCP> herbie@polaris.UUCP (Herb Chong) writes: >In article <830@psivax.UUCP> friesen@psivax.UUCP (Stanley Friesen) writes: >> Actually this can be carried further, after all "local areras" >>*overlap*, so it should be possible to develope a chain of local, or >>at least low-cost zone, calls all the way across the country. > >is the added propogation delay worth it? remember that the primary >reason for the backbone sites is the reduce the propogation delay >as well as to ensure that a large group of sites have a decent chance >of getting ALL news. > Yes, or at least a slight increase in propagation delay would be worth it if it ameliorated the net transmission costs. Actually, just a careful analysis of long distance links to eliminate unnecessary ones(like multiple links into the same aerea), and to find two or three step paths that are cheaper per site would be a big help. Other good suggestions include having the *recieving* site make the long distance call, not the sending site, and I saw an article about a new type of network link based on radio communication. All I am saying is there aught to be a way to *optimize* the net news transmission to minimize cost, and that *would* be worth an increased propagation delay. -- Sarima (Stanley Friesen) UUCP: {ttidca|ihnp4|sdcrdcf|quad1|nrcvax|bellcore|logico}!psivax!friesen ARPA: ttidca!psivax!friesen@rand-unix.arpa