Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version Vortex 1.0 6/6/83; site vortex.UUCP Path: utzoo!linus!decvax!vortex!lauren From: Lauren WeinsteinNewsgroups: net.news Subject: Usenet, Inc., etc. Message-ID: <73@vortex.UUCP> Date: Mon, 25-Jul-83 04:18:51 EDT Article-I.D.: vortex.73 Posted: Mon Jul 25 04:18:51 1983 Date-Received: Mon, 25-Jul-83 13:23:59 EDT Organization: Vortex Technology, Los Angeles Lines: 179 Greetings. Some of you out there may recall my comments on this subject a couple of Usenix's ago, and my opinions have not altered since then. In short, I'm sorry gang, but I just don't buy it. I have nothing against the concept of information utilities (and in fact I have been involved in the planning of such beasties) but the Usenet framework simply cannot be coerced into a "commercial" pattern without total disruption -- with the end result most likely being FAR fewer sites distributing even less useful information at HIGHER cost. There is a simple rule that pervades all aspects of technology, and it applies here as well: "If something isn't broken, DON'T TRY FIX IT!" And, contrary to the feelings that some persons may have, Usenet (in general) is NOT BROKEN. Let's consider some points, both positive and negative: 1) My contention all along has been that, contrary to appearances, USENET DOES NOT EXIST. Yes friends, even though you are reading this message, there is no "real" Usenet. What we DO have is a collection of many random computers happily calling each other (essentially at random) and spilling various messages onto each other's disks. There is no central authority, no central control, no FORCED costs. Each site makes an individual decision regarding what resources (if any) they can devote to participation in Usenet activities. If some sites cannot absorb heavy telephone costs, other sites are free, as they see fit, to subsidize those less "wealthy" sites. Nobody FORCES sites to poll other sites -- such polling is done when it is considered beneficial to all involved, and may involve as few or as many newsgroups as desired. If a site feels that it is no longer beneficial to continue polling a given site or set of sites, they may discontinue at any time. The point here is that Usenet provides a very efficient system for ensuring maximal freedom AND maximal fairness, since each site only calls those sites that it has decided it can afford to call and that it wishes (for whatever reasons) to call. This all takes place without a central authority and proceeds well. There is no reason to add a layer of bureaucracy! I've seen similiar situations in the past where well-meaning individuals have set up such mechanisms to try bring "fairness" and "order" to a situation. The result is almost always an increase in costs and a decrease in value. However, some bureaucrats DO get jobs out of the deal. Whoopee. 2) A central Usenet authority would find itself to be LEGALLY responsible for most Usenet interactions. This would include possible lawsuits over slanderous materials and possible copyright violations, dissemination of possibly obscene material into areas that have strict rules against such materials (encryption makes no difference from the legal standpoint), and so forth. I don't think that anyone seriously wants to take this on in this environment. Right now, since there really is NO network, there is no single organization that can be sued or otherwise pinpointed for blame. In fact, since Usenet is really just lots of separate "private" messages which are being relayed from point to point, it is probable that materials could not even be considered to be "published" at all -- which might be a very desirable situation. Create Usenet, Inc., however -- and this all changes. You've got an entity to blame now, and the messages have been formalized into a framework that would probably be considered to be publishing -- subject to the same rules and regulations as more conventional publications. I don't think we really want this at this time. 3) You cannot simply gloss over the fact that Usenet costs are typically bundled in with other UUCP and communications costs. This lack of formal differentiation of costs allows many sites with "hard-nosed" administrators to participate in a fairly free environment with a minimum of unnecessary bureaucratic interference. If you turn Usenet into a "line item" on the budget, you are going to find many sites subjecting news to intense scrutiny, and all but the groups considered most "relevant" by the adminstrators, regardless of their actual merit, will be dropped. Maybe this would be more "efficient" in terms of budget outlay and win you points with the head office -- but is this really what you want? Probably some "rich" sites could avoid this fate, but I'll give you odds that many sites will find themselves either cutoff completely or severely restricted. 4) How do you establish a fair "charging" system? Some sites could afford to pay thousands of dollars a year -- IF administrators would agree to such an explicit item. Others couldn't afford anything -- just paying the monthly bill for power and phone lines (not even including any long distance calls) can be an extreme strain. However, the persons who submit the best quality messages cannot always be expected to be located on such "rich" machines -- they're going to be spread all over, including at sites with little or no budget and little or no hope of getting approval for formal payments to Usenet, Inc. Should all of these "less fortunate" people be cut off? This might leave us with a lot of messages from rich sites -- but would the overall quality of the net be the same? I say no -- it would be lower. I've given this quite a bit of thought, and I do NOT believe it to be possible to derive a "fair" charging plan for Usenet. The only entity that would be sure to come out ahead would be Usenet, Inc., itself. 5) The idea of charging persons for submitting items is not reasonable. Persons who submit useful material should not be penalized for such practices. A person who wants to answer a question in an area in which he or she has expertise should not have to pay for the privilege. It's not even clear if the person ASKING the question should be charged, since the question (and answers) might be of interest to a wide range of persons on the net. I could go on, but I think you get the idea. Now, let's get to the bottom line. The real "problems" with Usenet don't have to do with organization or overall costs. As I've pointed out, individual sites optimize calling and costs as they see fit, and the lack of a formal organization has certain very significant benefits. So what's the problem? THE QUALITY OF MESSAGES! That's the bottom line, in my opinion. We don't REALLY need 30 people telling us that "BTW" stands for "By The Way". ONE would have sufficed -- and that's assuming that the question should have been asked on the network in the first place. A very large percentage of the items on the net should really be sent as direct mail -- not as netnews articles! If all of the messages that essentially say the same thing or that really shouldn't be present at all could be expunged from the net, I believe we'd find an immense lowering of overall costs, and a MUCH more useful "network". I am continually hearing from people who have stopped reading most or all of netnews since they just don't have the time to filter through the garbage to find a few reasonable and useful items -- and who can blame them? It's not even the NUMBER of people that causes the problem -- the network could handle many more users successfully, since most people don't generate many messages, most none at all in fact! There are elements to dealing with this problem successfully. None of them involves Usenet, Inc. or anything like it. One element is clearly user education. There really has been no formal attempt to make sure that all users are properly informed as to the matrix of customs that make up Usenet. These "customs" should be put into some sort of more formal framework as quickly as possible. Another important element (here we go again) is to make better use of DIGESTS!!! Yes, digests! It must be becoming apparent by now that some form of editing of many materials to cut out repetitious and nonsense messages is increasingly necessary. This isn't necessarily censorship. Would you expect "Time" magazine to publish every manuscript sent to the editor? You wouldn't have the desire (or time) to read anything if they operated in that manner. The use of digests would mean that only ONE copy of most messages would be MAILED to the digest moderator, and only the finished digest would be broadcast to the entire net. There are certainly technical issues to be worked out, but I believe that a massive improvement in quality and a lowering of costs could be realized by such a technique being used for even a few of the more verbose newsgroups. I've taken up far too much space already, and I apologize in advance. However, I've seen situations in the past closely analogous to this one, and Usenet, Inc. is simply the wrong way to go. There are solutions to our specific problems, but Usenet, Inc. is not one of them -- it would only cause its own problems. I respect Usenet, and I want to see it continue -- but that means dealing with our specific problems in the correct manner, and Usenet, Inc. just isn't it. I hope some of you agree with me. --Lauren-- {decvax, ihnp4, harpo, ucbvax!lbl-csam, randvax}!vortex!lauren P.S. As for satellite circuits and such... I do work in this area, and I don't believe that Usenet generates the "correct" form of traffic to be cost effective for the leasing of satellite channels. Intrastate telephone costs in many cases are similar to long-haul interstate costs at various hours, and my calculations in the past for other clients have convinced me that satellite channels are only useful for extremely high volume traffic that is heavily centralized. There are a couple of hops in Usenet that might have enough traffic to almost be suitable, but the costs would be similar if not higher in most cases. There are alternatives to satellite channels, of course, some of which might be more cost effective than what we're using now. I can elaborate on this topic later if anyone wishes. --LW--