xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (02/06/91)
Please notice and respect the followup line! dillon@overload.Berkeley.CA.US (Matthew Dillon) writes: > (1) Add a global net restrictions to the number of groups an article > may be cross posted to. For example, 2 or 4. Basically throw > away any additional specified groups automatically. > This will get rid of the more blatent cross postings. If you > really need to cross post to more, you have to work a little > harder to do (i.e. two separate messages) rickert@mp.cs.niu.edu (Neil Rickert) writes: > No! No! No! > An article cross posted to 20 groups may be annoying. But it is still > much less annoying that posting the identical article 20 time, once > per group. And being computer users most of us know how trivial the > multiple posting is to do. Don't even think of the idea. I didn't like several of Matt's ideas: * Automatically pass through newgroups, and you get alt.psu.girls.have.big.tits showing up on the rn "add this group" messages again; hand application is better, best would be an intermediate status between doesn't exist and does exist, say "writable not user readable" so that fast arrivin postings wouldn't go to news.junk but the sysadmin would have time to nuke groups with obscene names, e.g. * nuking newsgroups automatically that miss a month of posting would kill many low traffic groups that are nonetheless widely subscribed, and so would be unsuitable as mailing lists. Comp.theory.cell-automata is one such group, which can go idle for months, then begin a strong thread or several in response to new software, new research results, and so on. * While 20 separate posts is more expensive than 1 article crossposted 20 times, it still saves bandwidth, since 20 little newsgroup wars die out a lot faster (most groups will completely ignore most articles) than one huge, crossposted, everybody gets mad at everybody newsgroup war. In the case of the particular abuse that started the present unpleasentness, if the articles by "Roger Rabbit" had been posted as separate postings, they could have been caught by t.b kill files and ignored, since none of the follow-on a.t.t-t responses would have been seen in t.b. No crossposts, no war. I would like to make a counter proposal that only requires one data bit per newsgroup, though some code effort and lots of politics would be required. Add a new status bit, like "moderated", that says "accepts crossposts"; if you put a group among the crossposting list whose members have voted not to accept crossposts, your article would be rejected; if you forged it and got past the test, downstream sites would drop it cold. This would let alt.tv.tiny-toon protect itself, as could talk.bizarre, or any other group the frequent target of malicious crossposts, and would eliminate some of the most frequent abuses toward/by talk.bizarre. I would expect that most groups would choose to refuse crossposts, given the option; most net conversations are far too wide spread anyway, and stay crossposted while the threads of conversation have long segregated into portions appropriate only for each separate group. Many talk.bizarre habitants like to abstract articles from other groups, misread or read dead literally or take out of context part of the article, and repost it to make a humorous point. While we in talk.bizarre laugh milk up our noses at such cleverness, the group of the original posting could probably take a pass on seeing the way their stuff got mangled, as another example of a benefit. Marking _all_ the test groups "does not accept crossposts" except perhaps one special one just for testing crossposting that has exactly one appointed reflector site, would eliminate a frequent and netwide crossposting abuse. And certainly no one could conduct a campaign of the sort I have been against a group that chose not to accept crossposting; note that almost nothing else will stop such a campaign, since the campaigner does not have to post _any_ of the articles seen in the target newsgroup. A more complex proposal like the unix "groups" mechanism could limit crossposting to named sets of groups, while another one could say "only accepts crossposts from sibling groups", a means of forcing rationalization of the net hierarchies, but both of these would require a lot more political haggling than a simple yes/no switch on crossposting. Your comments solicited. As usual, I can design, but not implement, this stuff. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>