webber@brandx.rutgers.edu (Webber) (07/03/87)
In article <8235@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes: > > The message you are replying to specifically mentioned that it was > > always assumed that some people would choose not to carry some things. > > Unfortunately, it is *precisely* the moderated groups that we *want* to > carry... so long as they truly are moderated. From our viewpoint, you > are basically proposing to destroy their usefulness. Well, I am proposing to allow other people a different mechanism for controlling their flow problems than doing it by group. Thus if you are doing it by group, at some point, you will be getting the combined effects of both. Of course, I understand that you actually get all the groups, it is just other people that you don't pass them all on to, so doubtless it wouldn't be that big a problem for your site. > > ... I definitely favour improved security and quotas > > for controlling the flow of news... > > Quotas, as you have been told repeatedly in private mail, are unworkable > nonsense. Improved security carries a heavy price, but is looking necessary. Yes I have been told that repeatedly. I have proceeded to repeatedly reply that they do. I have then proceeded to explain how. People have asked questions. I have answered questions. People have asked more questions. Doubtless you haven't seen all of these. However, the only person who actually probed the concept among the backbone was kre, who you should know. He seemed to think the idea was ready for a trial implementation. Of course, he has such significant flow problems that he may be desparate enough to encourage anything that might work. > We are aiming at a somewhat higher level of security, specifically one that > is good enough that Bob Webber in particular will not be able to routinely > break it even if he tries. This is unfortunately a lot of work, both for > us and for the software, and we would have preferred to avoid it. Alas, due > to some shortsighted actions, both by the backbone and by certain individuals > (notably you), the consensus that has kept Usenet running smoothly is > starting to break down. Well the net certainly was not running smoothly from all I heard. Weren't there people complaining about too much flow and not enough news groups and stuff like that? I think you are overly ambitious with regards to security -- I would offer to test it out for you -- but then we would probably never see the system delivered. > We see nothing terribly wrong with being compatible with the existing setup. > There isn't anything really wrong with it except that it doesn't scale well > to really large user communities. We have thought a great deal about this, Oh, it's only problem is that it doesn't handle the current situation. > and our conclusion is that the only long-term solution is 100% moderation. Ahhh. If you completely change the content of the net then it will work on your system. Gee thanks. > However, it seems worthwhile to speed up traffic handling anyway. Yeah. What the heck. Might as well. > > This would yield > > a new net that was initially smaller and preserved many of the virtues > > of the old Usenet... > > What you're suggesting, as I understand it, is deliberate incompatibility > specifically aimed at forcing most existing sites off the network. You > have not explained why (a) the same thing would not happen to the new > network after a year or two, and (b) how this is consistent with your > disparaging comments about the "alternate backbone" concept. Actually, the problem with the `alternate backbone' concept is that as soon as it becomes successful, it will get merged back into Usenet and once this happens, it will loose its focus and the same problems will reappear. The incompatibility is that some sites will base their flow control on quota and others by group. The net result will be that the total flow control will be some funny function of both and neither side will be quite happy. However, this is the same as where the net is headed anyway because there is a great variance in the way different sites rank the various groups and hence the order with which they will choose to cut (if they take it into their head to actually make the decision for themselves instead of just going along with your favourite set). > Let us not mention such silliness as confusing the desire to see new groups > moderated with unwillingness to create them. I would hardly call a moderated group a group in the sense of Usenet up until the recent messup. The better ones look more like magazines. The worse ones are just a pain with delays, frustrations, and sometimes even indigestable digests. Any day now I expect to hear that comp.misc.unix has decided that it really wants to be digestified this week. ---- BOB (webber@aramis.rutgers.edu ; rutgers!aramis.rutgers.edu!webber)
kre@munnari.oz (Robert Elz) (07/06/87)
In article <281@brandx.rutgers.edu>, webber@brandx.rutgers.edu (Webber) writes: > In article <8235@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes: > > Quotas, as you have been told repeatedly in private mail, are unworkable > > nonsense. Improved security carries a heavy price, but is looking necessary > > However, the only person who actually probed the concept among the backbone > was kre, who you should know. There's some common misconception here that the "backbone cabal" is a tight knit group of collaborators .. not true, I don't know almost any of the backbone admins. I don't think I've ever actually met Henry, I do think I've "seen" him at a usenix or something like that. Apart from electronic communications, that's the extent of our acquaintance. > He seemed to think the idea was ready for a trial implementation. That's no quite what I said. What I recall saying was that the idea was nonsense, but if you want to try and implement it, to demonstrate that it really does work, then go ahead, do it. Its your time you'd be wasting. > Of course, he has such significant flow problems that he may be desparate > enough to encourage anything that might work. No, compared with most sites on the net I have no flow problems at all. We simply get only a very reduced number of groups, and we try to pick the ones with real value (that is, we want some assurance that what we pay for is going to be worthwhile, so we like moderated groups, the quality control is much more even, we know what to expect from week to week). kre