ron@brl-tgr.ARPA (Ron Natalie <ron>) (03/05/85)
> Net.music is only one of the newsgroups that is gated to the ARPAnet; the > gateway is "INFO-MUSIC". Those on ARPAnet receiving the information have > no place to post followups, except to the parent group. (That is, unless > the subgroups are enabled at each of the ARPA sites.) Up until recently, > the net.music subgroups were mis-forwarded into net.music. > > (I might be wrong about some parts of the gateway concept; I'm not at an > ARPAnet site. Perhaps Ron Natalie or Gene Spafford can elaborate further > on what is being forwarded, and how the gateways work.) > Your analysis is correct. Up until very recently net.music.* was posted to INFO-MUSIC. INFO-MUSIC was posted to net.music. If someone on the arpanet responded to a net.music.folk article, it would appear in net.music. INFO-MICRO had this in a much worse way, so I turned off the gateway. I finally figured out how to block the subgroups about a month ago. Still INFO-MICRO is plagued by the lack of mailing lists for groups like the Commodore and a lot of it's traffic is commodore specific garbage. I am in the process of getting all the net.micro subgroups paired up with their Arpa counter parts. Finally, I would like to point out that I would like to divest BRL of doing the INFO-MUSIC anything distributions. That is, I will still provide forwarding from USENET to wherever the list gets rehomed to, but I want the ARPANET side of INFO-MUSIC run from some other host. Any ARPANET volunteers? -Ron
cem@intelca.UUCP (Chuck McManis) (03/12/85)
Jim Poltrone (a/k/a Poltr1, the Last of the Raster Blasters) writes : > > Net.music is only one of the newsgroups that is gated to the ARPAnet; the > gateway is "INFO-MUSIC". Those on ARPAnet receiving the information have > no place to post followups, except to the parent group. (That is, unless > the subgroups are enabled at each of the ARPA sites.) Up until recently, > the net.music subgroups were mis-forwarded into net.music. > Having dealt with Usenet for a while and ARPAnet for still longer I may be able to clear this up somewhat. The ARPAnet does not have the same mechanism for distributing "news" as the Usenet community. Whereas all of the net.* and *.* newsgroups on the usenet are maintained by the news software at the individual sites the ARPA groups are *all* mailing lists. What distingushes the group from an individual is usually the INFO- prefix. (HUMAN-NETS is of course an exception) In all cases the mailing list has a moderator who maintains the list. His/Her is always list-REQUEST@sitename, where "list" is replaced by the group of interest (INFO-MUSIC in the above example). There is no concept of the "subgroup" in this model. So when the average ARPA user logs in in the morning, if he or she "subscribes" (read: is on the mailing list for) to say INFO-MUSIC, then chances are there will be a message to the effect of "You have new mail from INFO-MUSIC@BRL-BMD". The messages appear in the persons MAIL.TXT file and are usually interspersed with other messages (see Digests below) from other people on the system or other mailing lists. Digests : Sometimes, when the volume of the mail in a mailing list gets quite heavy, and the moderator is very ambitious (or an insomniac) he/she can begin to digestify a mailing list. Of the two most common ways to do this is digest by volume or digest by date. In the first case, messages are collected by the mailer on the host system into a "holding" mailbox. When sufficient mail has collected to make a xxLine or xxK chunk the contents are concatenated together, the subject lines copied into a mini table of contents at the top, and then the whole kit and kaboodle is mailed to everyone on the mailing list. In the second techinique, the holding mailbox is "digested" at say midnight every day and sent out the next morning. There is software that handles most of the drudge work of assembling digests, but the moderator will usually read through the resulting digest to remove the "please add/drop me to/from the list" messages that should have gone to the -REQUEST address. Usage: In typical day to day use, on the ARPAnet, one would use any of a number of mail programs to peruse their messages, some of these would be from a mailing list such as INFO-MUSIC. When something of interest is read, the user will typically reply to the sender, and if he/she feels there is enough interest use the CC: field to send a copy to the entire news group by adding INFO-FOO@WHEREEVER to it. In some cases the Reply to All command will accomplish the same effect. If this person wanted to start or stop receiving messages from some newsgroup they would send an appropriate message to the -REQUEST address to be added or dropped. The UseNet Connection : So how does all of this relate to Usenet? Well as is often the case, some able programmers said "Gee, we have people on this machine asking the same questions as the people on this other machine. If only they could exchange mail everyone could be more enlightened." And the question of how one could forward mail between networks was born. Essentially all it takes is two machines on different networks connected together by a third network. Such is/was the case at UC Berkeley and SRI international, a couple of the more well known gateways. When one sent mail to "amd!intelca!cem"@Berkeley on the arpa net, the berkeley mailer program was modifed to recognized this a usenet address and forward the mail to me via the UUCP network. Identically, when I sent something to ucbvax from the usenet side like amd!ucbvax!MCMANIS@USC-ECLC the rmail software was appropriately modifed to recognize the "MCMANIS@USC-ECLC" part as an ARPA address and forward it via the TCP/IP network. How does this relate to news? Well, when people move they take there old habits with them, and some took their mailing list subscriptions with them too. Many people would send a message though the gateway to the moderator of their favorite list and ask to be re-added. Which, they would be and there were no problems at first. However, when 75 people on an ARPAnet machine all get the same messages from the same mailing list it slows things down, but the software is generally smart enough to maximize the network bandwidth. When those same 75 or even 100 to 150 people all have the same messages funneled through a 1200 baud modem 75 to 150 times it gets to be quite unacceptable. So the system admin people yelled "Wait a minute, whats going on here" and said "Why has this message just been sent out 150 times?" when they connected it with the ARPA mailing lists they decided to come up with a better way to do it. Which they did, and the fa.* (From Arpa) newsgroups were created. Now, the ARPA mailing list could send one copy to the gateway and the gateway would post it to the appropriate fa.x newsgroup (Say fa.info-music) This worked and still does, but the people reading the net.* news were still missing the fa.* stuff. So the software was modified to try to post the ARPA stuff to the appropriate net.xxx group. Which is easy from the arpa side, INFO-MICRO --> net.micro, but they other way was real tough. So as an interim fix (in the case of net.micro) net.micro.* went to INFO-MICRO and INFO-MICRO went only to net.micro. This causes those really strange postings that were being complained about in net.music. Using net.micro as an example, if I posted an article to net.micro.atari, the gateway combines it with all the stuff from the other groups and forwards it to INFO-MICRO. Someone on the ARPAnet sees my article and replys to INFO-MICRO which always forwards into net.micro. Poof! There is an answer in net.micro for a question in net.micro.atari. This can only be alleviated by either the group being "gatewayed" to a more specific ARPA mailing list (say INFO-ATARI if it exists) or disconnecting the subgroup from the Gateway and making the connection between the mailing list and the news software 1:1. This has gotten much longer than intended, but hopefully will clear up some of the misconceptions around (like you can't send mail to someone on the ARPAnet, etc) Of course none of this is documented, you had to be their to see it happen. I sometimes think the computer industry is developing a whole new generation of human history books, passing info on from generation to generation. I guess we'll see in a couple of decades... --Chuck -- - - - D I S C L A I M E R - - - {ihnp4,fortune}!dual\ All opinions expressed herein are my {qantel,idi}-> !intelca!cem own and not those of my employer, my {ucbvax,hao}!hplabs/ friends, or my avocado plant. :-}