webber@aramis.rutgers.edu (Bob Webber) (07/10/87)
First I would like to thank allbery@ncoast.UUCP (Brandon Allbery), mjl@tropix.UUCP (Mike Lutz), dovich@ge-dab.UUCP (Steven J. Dovich), reid@decwrl.DEC.COM (Brian Reid), and lou@hoxna.UUCP ( L. Marco ) for their news postings critical of my quota proposal (or related issues) and even more so thank the people who mailed to me criticisms of the quota proposal. I will send personalized replies to the mail messages, but all of you have addressed some closely related issues to which I will outline herein why I do not feel they are over-riding objections. I would also like to anonymously thank the person who sent me a message explaining how the system could be implemented under the current version of news without re-coding (at the moment I certainly wouldn't want to brand them by indicating that they were the person that took the wind out of the sails of the `implement it first' people). Of course, most of you are aware that I am not proposing a parallel news system, but rather an alternative way for individual sites to handle the current problems (both quota and news.group name restrictions could coexist). Some of you are also aware that when I speak of communications bandwidth, I am referring just as much to limitations on the cpu's at the ends of the wires (such as disk space) as I am to the communication hardware that the computer views as a peripheral. Practically everyone has realized that both the quota system and the moderator system claim to solve the same problems with regards to resources dedicated to news. The question seems to rest with the fact that these two approaches yield very different environments for news to continue under. What I will address below is why I feel that the environment created by quota is to be preferred to that created by moderation (even `good' non-censoring moderation). It seems to me that people who have been proposing moderation have been using as their vision of news some kind of online library where people go to get reliable information. While this is a worthy goal, I find it completely alien to what the news system has been about for the last 5 years or so. First I should introduce myself, I am a reader of news. I have been reading news off and on for the past 5 years and yet prior to the moderation discussion, my postings have never generated any but the smallest of reply streams (incidently, I am quite happy with this and look forward to it once this topic has passed). Hence, I am much more a `reader' than a `poster.' When I ask myself what is the essence of Usenet, I say that it is direct unfiltered communication between a very large number of people (most of whom have above average educations and experience with computers). Reading the soc, talk, and rec groups, you will realize just how varied the people on the comp and sci groups are (many people contribute in both sections). In my opinion, placing a filter on news (which is what both the moderators and the people who use group names as a basis for deciding what to get are doing) is completely unwanted. I use a news reader that notifies me of subject lines and authors of all messages and I am quite happy to pick and choose what messages I read on the basis of this information (and the experience I have built up from doing this). Of course, it is claimed that the filter will raise `quality' and remove redundancy. This is exactly what I do not want. When I go to an expert on computer monitors and ask them what is the best monitor for displaying CAD work, they will tell me their opinion. When I go to the net and ask this same question, I see not only their opinion, but how other people react to that opinion and whether or not and to what degree other people share that opinion. Although a moderator might feel quite justified in simply saying that here is a typical message and 20 other people said basically the same thing, I would actually want to read them. The same thing applies to other topics in the sci, soc, misc, news, and talk groups. This is something that only net news has been able to reliably handle because of the size and type of user community it has attracted and because the postings are completely unfiltered. (Although this may appear to address only the redundancy issue, I believe it also addresses the quality issue). Some people have indicated that they strongly feel that it is in their company's best interest to subscribe to certain technical groups, but not the other nontechnical groups. This is difficult to evaluate. When I see a posting to a newsgroup that says ``Please mail replies, I do not read this group,'' I figure there is someone who is attempting to benefit from the group as if it were some kind of library. However, if someone is actually tracking a group, then that person would probably also be interested in other postings by the same people on other subjects (both technical and nontechnical) in order to get a context within which to evaluate the opinion that is being offered to that group. So assuming that a company has employees that have the time to process the net, I think there is some true benefit to getting the whole thing. In most cases, I suspect that people would be better off purchasing better documentation, hiring qualified consultants, and buying source licenses, than trying to use the net solely as a `library.' I personally think that people have overvalued the usefulness of sources and binaries that are over 64k (current net message size limit) and usually aimed at some specific application. My proposal did not say that one couldn't use uucp to move such items around, only that this should be separated from the function of news. I am particularly distressed by the posting of binaries. Although I can understand how for a few months after a computer comes out this might be the only reliable way to distribute code, non-optimizing C compilers for standard architectures (such as 68000s, 6502s, and 8088s) just aren't that hard to write. Access to the source of compilers is not that difficult to come by, besides the Gnu C implementation, I have also seen books containing full source to P-Code pascal compilers. While these may not be everyone's idea of `good' languages, they have certainly been sufficient for people running on `larger' machines (when was the last time you saw a SunIII binary or Pyramid binary posted to the net?). So it is really just a matter of a small portion of a user-community that is clearly doing alot of coding deciding to sit down and port a compiler. Some people have felt dismay that under the quota system they will miss messages from certain `experts,' since the quota system will inevitably loose messages (unless it should magically turn out that there really is enough bandwidth to support news unchanged, which is something we will probably never know if the net goes moderated and technology improves). I can only say that that is true. The implementation of the quota system is the turning point at which we must decide whether we want to build a futuristic interactive library or we want to continue to provide the traditional netnews service of unfiltered access to a broad range of opinion. I hope people will realize the worth of the latter. I think there are plenty of other people already working on the former, and so it is not necessary to convert netnews over to that task. Finally, let me indicate that I often benefit from `incorrect' comments that have sometimes been posted to the net. The reason is because they have sometimes motivated me to think about why they are incorrect and why someone would say them and have thus lead to some thoughts that I have enjoyed (and which I have occasionally shared with the net). --------- BOB (webber@aramis.rutgers.edu ; rutgers!aramis.rutgers.edu!webber)
eric@hippo.UUCP (Eric Bergan) (07/14/87)
In article <1006@aramis.rutgers.edu>, webber@aramis.rutgers.edu (Bob Webber) writes: > It seems to me that people who have been proposing moderation have been using > as their vision of news some kind of online library where people go to get > reliable information. While this is a worthy goal, I find it completely > alien to what the news system has been about for the last 5 years or so. > [...] > When I ask myself what is the essence of Usenet, I say that it is direct > unfiltered communication between a very large number of people (most of whom > have above average educations and experience with computers). Its not clear to me that your vision of the use of usenet is shared by the majority of the people on the network. I certainly view usenet as much more of a library or distributed bulletin board than a "town hall forum" for carrying on conversations. Its definitely true that moderation favors the former model, and quota systems favor the latter system. But I think that the network as a whole will have to decide which model they prefer. This may be extremely difficult to judge, since such a small fraction of readers of the news ever post anything. The rest of your posting is primarily defense for your position that moderation is not an effective control of resources, given your view of what the network should be. I agree, no question that moderation will tend to dampen the kind of interaction you are looking for. On the other hand, quota systems will dampen the kind of interaction that I am looking for - namely that questions be answered, but only once - rather than 9000 postings of dictionaries of different "face" symbols possible under ASCII character sets. Who is to say which of us has the right to impose their model on the other? I think the answer is that neither of us, deciding on our own, has the right to impose on the other. The net as a whole needs to decide. Those that are unhappy with the decision will have to either live with it, or form an alternative mechanism for the kind of model they are seeking. > In most cases, I suspect that people would be better off > purchasing better documentation, hiring qualified consultants, and buying > source licenses, than trying to use the net solely as a `library.' I don't think this is necessarily the case. When I request information from the network, it tends to be for small "snippets" that really aren't available in any other form. Examples would be finding out other users experiences with a given product, or how to get system x to work with system y, etc. There may not be better documentation, it may not be worth hiring a consultant (who may not know the answer, either), and it may not be solvable from source (when's the last time you saw source code for the firmware in the LaserWriter offered?). -- eric ...!ptsfa!hippo!eric
allbery@ncoast.UUCP (Brandon Allbery) (07/17/87)
As quoted from <1006@aramis.rutgers.edu> by webber@aramis.rutgers.edu (Bob Webber): +--------------- | the function of news. I am particularly distressed by the posting of binaries. | Although I can understand how for a few months after a computer comes out | this might be the only reliable way to distribute code, non-optimizing | C compilers for standard architectures (such as 68000s, 6502s, and 8088s) | just aren't that hard to write. Access to the source of compilers is not | that difficult to come by, besides the Gnu C implementation, I have also | seen books containing full source to P-Code pascal compilers. While these +--------------- Heh, heh, heh. News for you. The Amsterdam Compiler Kit costs $1000 plus. Gnu C will not run on small machines, as it's designed for VM architectures. And, I have *used* that P-code Pascal book (THE BYTE BOOK OF PASCAL, for those inter- ested); if you think its execution speed is acceptable, I challenge you to use a PD spreadsheet "compiled" under it. "Just sit down and write a compiler"??? Look at the Minix compiler some day (better yet, compile something with it -- bring a copy of THE LORD OF THE RINGS with you, you'll have read it twice through before it finishes compiling a modest-sized program; and _that's_ based on the ACK!). Look at how long it has taken Borland to come out with Turbo C (it's been "in progress" ever since Turbo Pascal came out), look at how long it took Microsoft to get their C compiler bug-free enough to be worth using. It is NOT easy to write compilers THAT WORK ON SMALL MACHINES. And such compilers are rarely cheap enough for the people who BUY small machines -- if I could afford Microsoft C in my software budget, I could afford an Altos box in my hardware budget. And -- your large machines COME with C compilers. This is not and never willbe true of MS-DOS boxes, and probably won't be true of OS/2 boxes if and when. I'm sorry, but judging small computers by large-computer standards is bogus. "Let them use sources" is rather like "Let them eat cake", and is just as incendiary to the computer-variety peasant. -- [Copyright 1987 Brandon S. Allbery, all rights reserved] \ ncoast 216 781 6201 [Redistributable only if redistribution is subsequently permitted.] \ 2400 bd. Brandon S. Allbery, moderator of comp.sources.misc and comp.binaries.ibm.pc {{ames,harvard,mit-eddie}!necntc,{well,ihnp4}!hoptoad,cbosgd}!ncoast!allbery <<The opinions herein are those of my cat, therefore they must be correct!>>