CAD:ricks (12/08/82)
#N:ucbcad:16100001:000:724 ucbcad!ricks Dec 8 00:24:00 1982 Another viewpoint on the proliferation of newsgroups: If people (even a small number) want net.bob.wombat, why should we stop them. If the news group turns out to be quite active (tens to hundreds of submissions), then it was a good idea to create it. If on the other hand, the number of submissions is quite small (1-10), very little disk space and phone/ethernet/arpanet tranmission time is used. Therefore I do not understand why people get so upset over the "proliferation" of small news groups. In fact, the number of flames about this could have filled a few small news groups. Rick L Spickelmier ...!ucbvax!ucbcad!ricks (USENET) ucbcad.ricks@berkeley (ARPANET) ricks@tektronix (CSNET)
pag@sri-unix (12/10/82)
It is not the traffic that many lightly-used, soon-to-be inactive newsgroups generate that is the problem. The problem is that they clog up the news directory, fill up the "active" file and individual .newsrc's, causing the overall execution time of readnews to suffer. Perhaps a different overall design would make the number of newsgroups independent on execution time, but that is not the case with our current software. --peter gross seismo!hao!pag decvax!brl-bmd!hao!pag ucbvax!hplabs!hao!pag
dan (12/10/82)
If the times of most recent modifications to .newsrc lines and the times of creation of the newsgroups were recorded, inactive newsgroups could be safely deleted. Readnews could automatically remove lines for defunct newsgroups from .newsrc files and detect recreation of deleted newsgroups. The news system doesn't have to be as slow as it is. A master dot file (perhaps formatted like a .newsrc) could be maintained by periodically merging all of the individual .newsgroup files. Readnews could look only at the master dot file and would become a cheap program to run. Most of the other news inefficiencies can probably patched over with less effort than we currently expend agonizing over the problems with large numbers of newsgroups. (The news system should probably be redesigned rather than patched, but ...) So, why are we even discussion this?
aron@sri-unix (12/10/82)
I would like to bring up a case history as a response to those who argue "if you don't get 1001 yes responses don't start a group". A long time ago we had an inactive fa.info-micro on the net. I suggested (and concurrently someone else did) a net.micro. I got about 10 yes responses which I thought was alot, ignorant soul that I am. The group was started, and is now one of the most active on the net. We've got gateways to arpa, five sub-groups, and daily contributions. Given our human tendency towards laziness and apathy unless hit over the head (look at how many people vote in national elections), I think that if you get 10 or more yes responses to start a new group, that represents at least a 100 interested people - enough to warrent a group . aron shtull-trauring seismo!harpo!presby!aron
sjb (12/11/82)
OK, I now realize that there is no real way to get an accurate account of general net-wide attitude to a new group by mail responses, as a lot of people just aren't going to respond. However, if we allow new groups to be created with this 'little' bit of support, I think we must do it under the assumption that, if the group turns out to indeed NOT be productive, we remove it. In this way, the worthwhile groups will stay, the dead ones will go, and everyone SHOULD (!) be happy.
mclure (03/09/83)
#R:ucbcad:16100001:sri-unix:8200008:000:323 sri-unix!mclure Dec 9 03:38:00 1982 The solution is for the notesfile and readnews maintainers (or some random hackers) to add "automatic deletion" of unused newsgroups to their software and see that it gets distributed. Then we would have no need to argue about potential Usenet dictators. A three-month timeout on newsgroup life seems reasonable. Stuart