mark@cbosgd.UUCP@mcnc.UUCP (Mark Horton) (08/01/83)
A couple of notes on moderated newsgroups: The real issue involved with moderated newsgroups is information quality vs speed of distribution. (My experience with moderators agrees with Lauren's - the moderators don't "censor" in practice.) It is unlikely that a moderator would go through their list of submissions more often than once a day, so the added delay is the time for mail to get to the moderator plus the time it waits for the moderators attention. The item is then distributed from a different point on the network, but a more central point (if the moderator is well connected) and will probably reach the net at about the same speed. Mail can take anywhere from a few minutes (in the case of a direct link and an autodialer) to overnight (in the case of a polled connection) to 3 or 4 days (mail to/from Europe seems to take this long - I haven't figured out why). In the case of a flakey link, it could sit in some spool directory arbitrarily long. However, a good average figure is probably overnight. If the moderator times things right, and does moderation in the morning, the total added delay would probably average one day. (Of course, if the moderator gets lazy or goes on vacation, it can take longer. I've seen this happen on the ARPANET many times. If the moderator is a grad student, and he suddenly gets a burst of energy and works on his thesis for 2 months, the digest may suffer.) By contrast, news travels at varying speeds. Within Bell Labs, it's not uncommon to see discussions where turnaround is a couple of hours. I saw one recently where the originator of the discussion posted 4 separate articles in one day, all in response to someone elses followups to his various articles. This only matters to a person who checks for new news every hour or so - I personally check it once a day. I claim that for most informational newsgroups (e.g. net.general, net.wanted, net.lang) an extra day of delay won't matter much. For some newsgroups the extra day would be fatal (e.g. net.news.config). For discussion newsgroups (net.flame, net.singles) such delay would be undesirable, but a moderator is not needed or wanted. For internal discussion newsgroups such as btl.all, given that almost everybody in BTL has a dialer, a moderator is probably a bad idea. The point is that only some newsgroups want or need a moderator. net.announce is going to be the first moderated newsgroup. It will be an experiment to see how the whole concept works out. Since this will take the place of net.general (which will continue to exist), the moderation delays probably won't be a problem at all. The last test message I posted to net.announce produced failure messages from only 7 systems. The Usenet administrators of those sites have been quite cooperative in upgrading their news to 2.10.1 (if you're still running 2.10 you'll have the same problem), and I will try again in a couple of weeks. Hopefully by the end of August we'll be able to start using net.announce. There is code in 2.10 to support moderated newsgroups which have names mod.all, all.mod.all, or all.announce. Thus, we have a good deal of flexibility in choosing newsgroup names for other newsgroups that wish to be moderated. Once we decide if this is a good idea, the next step would be to either create (say) net.mod.misc, or else create the newsgroup class mod.all by having all the system administrators put mod.all in their sys file as a class they forward everywhere. Mark Horton
teus@mcvax.UUCP (Teus Hagen) (08/04/83)
I basically agree with Mark that for some groups a moderator is needed. The delay will be about one day. However I do not see it is needed to have only one moderator around. Perhaps having more moderators will speed up the posting and is less depending on the workload of someone. Of course the moderators will disagree about certain articles, personally I do not care about that.
Lauren Weinstein <lauren@vortex.UUCP> (08/05/83)
I'd like to clarify one point -- it is in the Usenet, Inc. environment that "censorship" is most likely to eventually occur (due to liability and monetary considerations). Such an outcome is not likely within moderated newsgroups operating in our current environment. The lack of a central authority prevents any centralized attempts to "steer" the course of discussions, and the current "spread" costs of the network avoid monetary restrictions from unnecessarily preventing reasonable discussions from taking place. In my view, the purpose of moderated groups (and/or digests) within our current Usenet would *not* be to restrict the flow of information to "useful" topics nor to prevent certain individuals from participating in discussions. However, moderators would be able to perform a number of useful functions. Most of our current "problem traffic" is the result of "timing problems" (e.g. most of the "multiple messages" we see that essentially duplicate each other), or lack of education regarding appropriate newsgroups for various messages. Both of these problems may be classed as "accidental" -- they're not the result of a determined effort to harass people or clog the network. So, our moderators (distributed around the net) would simply act to prevent the "accidental" incidents from getting out of control. When a particular topic generates large volumes of "duplicate" responses (e.g. the "BTW" topic) he or she would return most of the later messages back to their senders, with a note explaining that the topic had already been covered by previous messages. Of course, the moderator would not treat messages with significantly different content in this manner -- such messages would be distributed. If the author of a "duplicate" message really wished it to be sent anyway, even after being notified as to the current situation, the moderator would of course be willing to do so, probably with an appropriate explanatory note for the network readers. In some cases, the moderator might find it useful to distribute "summary" articles that incorporated some of the various replies, or to simply note (or list the names of) the various people who contributed similar statements on the current question. The whole point here is simply to allow questions to be answered with a minimum of repetition through careful, but non-intrusive, moderating of incoming materials to some groups. The moderator would perform a similar function when it came to "mis-addressed" articles. In this case, a message would be sent back to the sender suggesting the "proper" group for the original message. Once again, if the author insisted, the message would be distributed in any case. In our current NON-Usenet, Inc. world, the freedom to distribute any message does not rest on the monetary or liability considerations of some central authority -- I for one would sure like to keep it that way! I believe that most net users would greatly appreciate the ability to avoid having their messages (unknowingly) duplicate other messages already in the pipeline, or to be informed when the "incorrect" newsgroup was used for a submission. As I mentioned above, most of our problems are the result of "accidental" interactions, not the result of deliberate acts. --Lauren--