bobvan (12/09/82)
Thanks to all news administrators who've responded to the user count query. I've received almost 1300 lines of mail with data for 107 sites. A few awk scripts boiled this down into the data that follows. Counting lines in /usr/lib/news/users yields an average of 56 users per site, with 32 sites reporting. Counts from other more accurate sources yield an average of 20 users per site, with 86 sites reporting. The larger size of this sample leads me to have more faith in it than in the figure derived from /usr/lib/news/users. If we (probably fallaciously) assume that 20 users per site is an accurate net-wide average, we get about 7,500 net news readers. Data is scarce, but it looks like the average net news reader spends about 8 to 12 minutes per day reading news (7-day average). Here is that data base I've built from your contributions (order of arrival): wc site lib/news/users "accurate" ==== ============== ========== tpdcvax 32 15 ittapp 44 38 ittvax 52 50 decwrl 9 ? floyd 87 ? eagle 15 ? u1100a ? 6 machaids ? 18 ariel 34 ? lime 29 ? houxe 23 ? houxg 53 ? cwruemp ? 21 burdvax 45 25 microsof ? 30 harpo ? 124 zeppo ? 30 hocda 24 ? utcsrgv 165 146 rochester 48 ? msdc ? 4 hp-pcd 46 ? bunker 69 ? utah-cs ? 25 we13 144 ? mi-cec 27 ? ll1 23 ? spanky ? 2 tekmdp-crimson ? 60 tekmdp-azure ? 90 tekmdp-bronze ? 30 teklabs ? 80 tekcrd ? 30 tekid ? 80 tekchips ? 20 iddvax ? 30 tektinker ? 10 crl44 ? 5 thor ? 10 hp-cvd ? 1 yale-comix 323 ? grkermit 20 37 hplabs1 ? 45 hplabs2 ? 35 cbosgd 34 19 u1100s 7 ? sask 27 23 utah-gr 78 ? cbosg 55 36 utcsstat 83 ? mitccc ? 28 pyuxll 29 14 altos86 ? 5 cfib ? 2 uw-beaver 32 ? uw-june 73 ? uw-vlsi 8 ? mcnc 49 ? Please report any errors of more than 5-10%. A "?" means data not available. The figures above include averages for 50 Bell Indian Hill sites that are not shown for brevity. The rumors of net news being distributed on paper within Indian Hill turned out to be false. They stem from one user who likes to have a "newspaper" to read while waiting for compilations to finish. Once I stop receiving updates and additions to the data base, I'll do a report for net.followup. Bob Van Valzah (...!decvax!ittvax!tpdcvax!bobvan)
lepreau (12/10/82)
Bob's readership survey is certainly interesting, and my initial reactions were: one hell of a lot of readers-- what a resource and audience, and two, but 10 minutes a day isn't bad at all... the net must not be such a waste after all. It was also a provocative survey, and later I got into bc using Bob's estimates of 7500 readers at 10 mins/day, 7 days a week and came up with the observation that: Time spent reading news equals the working hours of 220 full time employees. I was going to write a long flame at this point but I'll just say that I'd rather have the benefits of 220 people writing software for us all than whatever benefits I get now. The latter are non-trivial, and include the obvious benefits of sharing software, making contacts, and getting bug fixes, as well as less tangible ones. But it just doesn't compare-- I don't think we really appreciate the true costs involved. Never mind the cpu/disk/phone costs (news is a dog, in case you hadn't realized). Please don't flame by counting up the hours spent on Reader's Digest and claim that we should therefore abandon that recreation for one million gnomes writing BASIC pgms for us. [Besides, the Digest is an important instrument of public policy-- Ronnie would be bereft of information without it.] I guess netnews is openly becoming "recreation", and therefore all is justifiable, but that sure isn't the way its costs used to be justified. [As someone (ber?) said on the net long ago: it's amazing what I'll read if it pops up on my screen, stuff I'd never glance at on paper. I only read the most well-edited magazines, but here.... I suddenly know my New Year's resolution-- but of course, that's in the future.] Jay Lepreau
trb (12/11/82)
I've heard plenty about what a waste of resource netnews is. I had a conversation recently with a coworker who had just brought up netnews at a site which many would consider devoid of technical interest. Shortly after bringing up netnews, he made an extremely pertinent comment; he said something like this: The thing I really like about netnews is that it's a bunch of people THINKING! You might not appreciate this too much if you work in some open-minded den of blue sky research, but there are lots of work places where you don't have easy access to any kind of creative thinking and communication about your own projects, not to mention other people's. Working in such a place is like being imprisoned, and a communication medium, even as sleazy as the one offered by the lowly netnews/mail system, is tantamount to intellectual freedom. (I'm dead serious.) I offer my sincerest thanks for netnews system, bugose though it might seem, cuz it's my most convenient path to all the brains out there. Is there a communications medium that exists anywhere that offers the price/performance of netnews? I certainly think not. Andy Tannenbaum Bell Labs Whippany, NJ (201) 386-6491
mark (12/11/82)
The problem with Jay's suggestion ("I'd rather have 220 people writing software for us all than the benefits of netnews") is that those 220 people wouldn't be writing software for us all. Most of us are owned by some company that brands everything we do proprietary. You can't have what we do. Those of you at Universities who actually write software (there must be a lot of you who do something else) are probably mostly very specialized and write stuff of use to a small community (or produce prototypes that are primarily useful for the papers they generate). There are certainly people out there who write software primarily for public domain purposes. If you add 10 minutes to each of their days, I think all you've accomplished is to give them an extra 10 minutes to answer the increased phone calls. (There is a big demand on these people, everybody wants their work, and their phones ring off the hook.) If they can't reach people with netnews, they will have to talk to them individually on the phone. For the rest of the people, what would you do with your extra 10 minutes? Your job? Or hang around the coffee machine and chat with your coworkers? (USENET really is just a huge "old-boy network".)
essick (12/11/82)
#R:tpdcvax:-23200:uiucdcs:10900014:000:1119 uiucdcs!essick Dec 11 15:22:00 1982 Just for kicks, here are some PLATO statistics I had handy (gathered from September 15, 1978 through October 19, 1982): Number of users: lots, literally thousands of signons Simultaneous users: max: about 650 average day 400 to 500 late night (2am) still 150 Notesfiles/newsgroups: 1573 (many are ``error logging'' and such) (probably only a few hundred used lots) (All of the mean figures are over the 1495 days that the machine was up during this period) mean sessions/day: 5080 mean session duration: 19.67 minutes total people-days spent: 103,773 mean people-days per day: 69.41 (these are 24 hour people-days!) mean notes/resps per day: 1408.7 total notes/resps written: 2,106,097 note/response ratios: notes: 38.4% resps: 61.6% Most of these were LOCAL to the U of I PLATO system. Only 6.81% of the notes were ``networked'' notes. These statistics are kept automatically by the PLATO notes system. They can be printed whenever needed through a program called ``notesys'' which gives just about all the breakdowns you could want. -- Ray Essick, University of Illinois at Urbana-Champaign