page@swan.ulowell.edu (Bob Page) (02/25/88)
About a month ago there was some discussion on alt.config about how big the alt 'network' was. I sent out a sendsys message to find out. Responses have finally stopped coming in. I posted the message on the evening (to me) of February 8, and all the responses were generated (but not delivered to me) by February 16. Total sites responding: 550 total .uucp sites: 247 total .com sites: 153 total .edu sites: 121 total .org sites: 13 total .gov sites: 6 total other sites: 10 (arpa, bitnet, cdn, csnet, .net, unspecified) I suppose I could sort & report on the sites geographically, or topologically, but I don't know how much more information it would provide, and I don't have a nice way to to it automatically. (Read: I don't want to do it by hand). However, it is clear that alt does not very far out of North America, or the sites outside of NA don't respond to sendsys requests. The arbitron data last posted claimed that alt.sources was carried by 66% of the net. That's 2/3 of the 710 sites reporting, or 473 sites that sent in arbitron reports. At 8900 estimated sites (total) on Usenet, I should have received about 5900 replies to my sendsys message. Since I only got 550, I'd say rather than 66% of the network sites carrying alt.sources, it might be closer to 6%. It shows how the "self-selecting" arbitron sample can produce bogus results. If the 8900 site estimate for UUCP/Usenet sites is anywhere near accurate, and only 75% of those carrying alt.* responded to my sendsys message (a very low figure, considering the message is automatically generated by the news software), then the percentage of sites carrying alt.* is about 8%. The numbers are subject to a lot of play, of course, but I doubt the alt subnet is carried by more than 10% of the net, and I'm sure that the 66% figure for alt.sources propagation is far too high. Here's a list of the sites that responded, sorted by site name: 3comvax.3com.com abhg.uucp acad.uucp academ.uucp admin.uucp adobe.com agate.berkeley.edu agora.uucp ai.uucp aimt.uucp albanycs.albany.edu altura.srcsip.uucp alvin.mcnc.org alxfin.uucp amdahl.uts.amdahl.com amdcad.amd.com amethyst.uucp aob.uucp aplcen.uucp apollo.uucp apple.uucp aramis.rutgers.edu argent.uucp argon.uucp arthur.cs.purdue.edu asi.uucp ateng.uucp aucs.uucp aurora.uucp aw.sei.cmu.edu axiom.uucp babbage.acc.virginia.edu bbn.com bellcore.bellcore.com bellman.acc.virginia.edu blblbl.uucp blia.bli.com bloom-beacon.mit.edu bobkat.uucp bocar.uucp boole.acc.virginia.edu bpa.bell-atl.com briar.philips.com brspyr1.brs.com bsu-cs.uucp bu-cs.bu.edu bucket.uucp bucsb.uucp buengc.bu.edu bungia.bungia.mn.org bunny.uucp burdvax.prc.unisys.com caelum.att.com caip.rutgers.edu cartan.berkeley.edu cascade.stanford.edu cblpe.att.com cblpf.att.com cblph.att.com cblpn.att.com cbmvax.uucp cbnea.att.com cbnscs.uucp cbosgd.att.com cbrme.att.com cbterra.att.com cc5.bbn.com cci632.uucp ccicpg.uucp ccng.waterloo.edu ccv.bbn.com cdes.uucp cello.hpl.hp.com cg-atla.uucp cierra.uucp cipric.mn.org cit-vax.caltech.edu claris.uucp clyde.att.com coffee.stanford.edu coma.uucp compsys.mu.edu copper.tek.com cos.com countach.uucp cp1.bell-atl.com cpesac.uucp cpsc6a.att.com cpsc6b.cpsc6a.att.com crash.cts.com csri.toronto.edu csuchico.edu csustan.uucp ctnews.convergent.com cucstud.uucp cunyvm.cuny.edu cxsea.uucp daemen.uucp daimler.uucp dalcs.uucp dalcsug.uucp dale.acc.virginia.edu dalsqnt.uucp dalton.acc.virginia.edu dandelion.ci.com dartvax.dartmouth.edu dasys1.uucp dayton.uucp decvax.uucp decwrl.dec.com denning.uucp devon.uucp dino.ulowell.edu dogie.edu donk.uucp doug.uucp drexel.uucp druco.att.com druhi.att.com druin.att.com druks.att.com drumn.att.com drune.att.com drupa.att.com drusp.att.com drutx.att.com druxh.att.com druxm.att.com druxu.att.com druxv.att.com dsinc.uucp e.ms.uky.edu ece-csc.uucp eddie.mit.edu edison.ge.com edsel edsel.uucp elbereth.rutgers.edu eleazar.dartmouth.edu ellymae.uucp elmgate.uucp elroy.jpl.nasa.gov elxsi.uucp ems.ems.mn.org encore.uucp epimass.epi.com epiwrl.epi.com escargot.uucp etnibsd.uucp euclid.acc.virginia.edu fauern.uucp flight.foobar.uucp foobar.hf.intel.com fredonia.uucp friday.uucp frog.uucp fxgrp.uucp galaxy gamma.uucp gandalf.littlei.uucp gargoyle.uchicago.edu gauss.rutgers.edu geac.uucp genesis.att.com gethen.uucp glacier.stanford.edu glimmer.uucp gmdzi.uucp gpu.utcs.toronto.edu granite.bbn.com granite.dec.com granjon.uucp gronk.uucp gse.entity.com gumby.cs.wisc.edu gvgpsa.uucp haddock.isc.com hall.cray.com hammer.tek.com hc.dspo.gov helios.toronto.edu hi-csc.uucp hi.unm.edu ho2cad.uucp ho7cad.att.com hocpa.uucp homxa.uucp homxc.uucp hop.toad.com hotlg.att hotps.att.com hou2c.uucp hou2d.uucp hou2e.uucp hou2h.uucp hou3b.uucp houdi.uucp houem.uucp hound.uucp hounx.uucp houtz.uucp houxa.uucp houxh.uucp houxj.uucp houxk.uucp houxl.uucp houxs.uucp houxu.uucp houxv.uucp hp-sdd.hp.com hpcea.ce.hp.com hpcea.ce.hp.com hpcupt1.hp.com hpcvaab.hp.com hpcvlx.hp.com hpda.hp.com hpdsla.hp.com hpdslab.hp.com hpesoc1.hp.com hpfcdc.hp.com hpiacla.hp.com hpihoah.hp.com hpisoa2.hp.com hpisod1.hp.com hplabsz.hpl.hp.com hpsemc.hp.com hpsmtc1.hp.com ht.ai.mit.edu hub.ucsb.edu hudson.acc.virginia.edu husc6.harvard.edu hutch.uucp hydra.riacs.edu icus.uucp idsssd.uucp ig.ig.com iisat.uucp im4u.uucp ima.isc.com imagine.pawl.rpi.edu infinet.uucp intek01.uucp io.att.com ism780c.uucp ista.uucp itsgw.rpi.edu j.cc.purdue.edu jade.berkeley.edu jlsoft.uucp jolnet.uucp jpusa1.uucp julian.uwo.cdn jumbo.dec.com kaos.uucp kestrel.arpa killer.uucp kitty.uucp kksys.uucp kodak.uucp koko.uucp krebs.acc.virginia.edu labrea.stanford.edu lakeoz.uucp lakesys.uucp laticorp.uucp laura.uucp lclark.uucp leah.albany.edu left.att.com lehi3b15.csee.lehigh.edu leia.gwd.tek.com lindy.stanford.edu linus.uucp lll-winken.llnl.gov louie.udel.edu lsuc.uucp lzaz.att.com lzfmd.uucp lzfme.uucp lznv.att.com lzsc.att.com m2c.org macom1.uucp magic4.uucp mahendo.jpl.nasa.gov mandrill.cwru.edu marque.mu.edu maxwell.acc.virginia.edu maynard.bsw.com mccc.uucp mcco.uucp mcdchg.uucp mcnc.org mcp.entity.com me.toronto.edu meccsd.mecc.mn.org meccts.mecc.mn.org megatek.uucp mendel.acc.virginia.edu metavax.uucp mibte.uucp midas.tek.com mines.uucp mit-amt.media.mit.edu mit-caf.uucp mitisft.convergent.com mmlai.uucp mnetor.uucp moogvax.uucp mordor.s1.gov mozart.uucp mrst.uucp mtfmi.uucp mtgzy.uucp mtgzz.uucp mtunb.att.com mtund.att.com mtune.att.com mtunf.att.com mtung.att.com mtunh.att.com mtuni.att.com mtunj.att.com mtunk.att.com mtunx.att.com mtuxj.uucp mtuxo.uucp n3dmc.uucp ncc.uucp ncrcan.toronto.ncr.com ncs-med.uucp ncsugn.ncsu.edu ncsuvx.ncsu.edu ndcheg.uucp nebula.att.com necis.uucp necntc.nec.com nis.nis.mn.org noao.uucp noether.uucp nosun.uucp nucleus.uucp nud.uucp nunki.usc.edu nvanbc.uucp oberon.usc.edu obpa1.uucp occrsh.att.com oddjob.uchicago.edu odyssey.att.com oliveb.olivetti.com omepd orange.qtp.ufl.edu orca.tek.com orchid.waterloo.edu orstcs.cs.orst.edu oswego.uucp otter.hple.hp.com ovaltine.berkeley.edu oxtrap.uucp pacbell.uucp packard.uucp panda.uucp papaya.bbn.com parsely.uucp paul.rutgers.edu pbhya.uucp pbhyb.uucp pbhyc.uucp pbhyd.uucp pbhye.uucp pbhyf.uucp pbhyg.uucp pdp.cs.ohiou.edu pegasus.uucp peora.ccur.com percival.uucp philabs.philips.com phoenix.princeton.edu phoenix.uucp plnapc.philips.com plus5.uucp pogo.tek.com polya.stanford.edu polygen.uucp polyslo.uucp portal.uucp poseidon.uucp princeton.princeton.edu psu-cs.uucp psuvax1.psu.edu psuvm.bitnet psychsun.stanford.edu pt.cs.cmu.edu ptsfa.uucp pucc.bitnet puck.uucp puff.cs.wisc.edu puffin.uss.tek.com pwcs.stpaul.gov pyrnj.uucp qiclab.uucp questar.questar.mn.org ra.rice.edu radio.toronto.edu redwood.uucp reed.uucp rhesus.primate.wisc.edu ritcv.uucp rocky.stanford.edu rolls.uucp rosevax.rosemount.com rtech.uucp ruby.tek.com rutgers.rutgers.edu sarin.uucp sbo.uucp scobee.att.com scooter.uucp sdcjove.cam.unisys.com sdcrdcf.uucp sdcsmb.sm.unisys.com sdcsvax.ucsd.edu sgi.sgi.com shamash.uucp shark.tek.com shasta.stanford.edu shrike.austin.lockheed.com sialis.mn.org siemens.uucp sigmast.uucp silver.bacs.indiana.edu sis.uucp sjfc.uucp sli.uucp smegma.uucp snark.uucp snjsn1.sj.ate.slb.com sol.arpa soleil.uucp spar.spar.slb.com spdcc.com spectrix.uucp speedy.mcnc.org spool.cs.wisc.edu spp3.uucp spt.entity.com sq.com sri-spam.istc.sri.com step.dwr.com stsci studsys.mu.edu sunybcs.uucp super.upenn.edu sw1e.uucp swan.ulowell.edu swatsun.uucp swlabs.uucp t9103.uucp tandem.uucp td2cad.intel.com teemc.uucp tekcae.tek.com tekchips.tek.com tekcrl.tek.com tekecs.tek.com tekfdi.tek.com tekgen.tek.com tekgvs.tek.com tekig.tek.com tekig4.tek.com tekig5.tek.com tekigm2.tek.com tekirl.tek.com teklds.tek.com tekred.tek.com teksce.sce.tek.com tektronix.tek.com telesoft.uucp temvax.uucp terminus.uucp thorin.cs.unc.edu ti-csl.csnet tmsoft.uucp topaz.rutgers.edu trwrc.uucp tslanpar.uucp ttidca.tti.com tub.uucp turtlevax.uucp twitch.uucp ucbvax.berkeley.edu ucdavis.ucdavis.edu ucsd.edu udccvax1.acs.udel.edu uflorida.cis.ufl.edu ukecc.engr.uky.edu umbc3.umd.edu umd5.umd.edu umix.cc.umich.edu umn-cs.cs.umn.edu umn-cs.cs.umn.edu umn-d-ub.d.umn.edu uni2.bcm.tmc.edu unido.uucp unisoft.uucp uokmax.uucp uoregon.uucp uport.uucp ut-emx.uucp ut-sally.uucp utah-gr.uucp uthub.toronto.edu utmanitou.toronto.edu uunet.uu.net uvaarpa.virginia.edu uvacs.cs.virginia.edu uvaee.ee.virginia.edu v.sei.cmu.edu valid.uucp van-bc.uucp vaxine.uucp vice.tek.com videovax.tek.com viper.lynx.mn.org virginia.acc.virginia.edu vixie.uucp vu-vlsi.uucp wa3wbu.uucp watale.waterloo.edu watcgl.waterloo.edu watdcsu.waterloo.edu watdragon.waterloo.edu water.waterloo.edu watmath.waterloo.edu watt.acc.virginia.edu watvlsi.waterloo.edu wb3ffv.uucp westmark.uucp whuts.uucp wp3b01.uucp wucs1.uucp wucs2.uucp xanth.cs.odu.edu xroads.uucp yetti.uucp yunexus.uucp zeus.hf.intel.com zeus.tek.com zorch.uu.net Full data (all the mail replies) can be furnished upon request, but it's over a megabyte, so don't ask lightly. ..Bob -- Bob Page, U of Lowell CS Dept. page@swan.ulowell.edu ulowell!page "I don't know such stuff. I just do eyes." -- from 'Blade Runner'
karl@triceratops.cis.ohio-state.edu (Karl Kleinpaste) (02/25/88)
Just a little side note: We at Ohio State receive and redistribute the alt.* groups, but I see we didn't show up in your list of responding sites. One could question why not, of course (it should have been easy to get a reply back to you from the standpoint of network topology, both Internet and UUCP), but in any event it's clear that there are sites doing things with alt.* which didn't get sendsys msgs back.
jbuck@epimass.EPI.COM (Joe Buck) (02/26/88)
In article <5077@swan.ulowell.edu> page@swan.ulowell.edu (Bob Page) writes: >About a month ago there was some discussion on alt.config about how >big the alt 'network' was. I sent out a sendsys message to find out. >... >Total sites responding: 550 >... >The arbitron data last posted claimed that alt.sources was carried by >66% of the net. That's 2/3 of the 710 sites reporting, or 473 sites >that sent in arbitron reports. At 8900 estimated sites (total) on >Usenet, I should have received about 5900 replies to my sendsys >message. Since I only got 550, I'd say rather than 66% of the network >sites carrying alt.sources, it might be closer to 6%. It shows how >the "self-selecting" arbitron sample can produce bogus results. I sent this reply to Bob, and I suppose I'll post it as well. Bob, the flaw in your analysis (in getting your 6% figure) is assuming that your sendsys reached all the sites on the altnet and that all of them successfully sent a reply to you. Nope, not even close (as I know from trying sendsys in the Bay Area before). If INTERNET isn't defined (it isn't in most systems on the net, still) news will reply to the Path: header, and we all know how well THAT works. There are also paranoid netnews administrators who've turned off the feature. . . . -- - Joe Buck {uunet,ucbvax,sun,<smart-site>}!epimass.epi.com!jbuck Old Internet mailers: jbuck%epimass.epi.com@uunet.uu.net
page@swan.ulowell.edu (Bob Page) (02/26/88)
jbuck@epimass.EPI.COM (Joe Buck) wrote:
>I sent this reply to Bob, and I suppose I'll post it as well.
A couple other people sent me similar responses .. "we get the alt
groups but didn't show up in your report."
Hey, I report what I get. I thought about putting a UUCP-like
Reply-To: line, but didn't. However, the Path: line left here saying
ulowell!page, so I should have gotten a large number of responses
anyway. Ulowell isn't out in the sticks, uucp-wise.
If only 25% of the alt sites could respond to my sendsys request, it
still means about 25%-35% of the net gets alt, which is still a lot
less than *I* had thought originally, and less than half of what
the arbitron stats report.
Until some better method exists for measuring connectivity, the best
we have is sendsys, so that's what I based by report on.
..Bob
--
Bob Page, U of Lowell CS Dept. page@swan.ulowell.edu ulowell!page
"I don't know such stuff. I just do eyes." -- from 'Blade Runner'csg@pyramid.pyramid.com (Carl S. Gutekunst) (02/26/88)
>>Since I only got 550, I'd say rather than 66% of the network sites carrying >>alt.sources, it might be closer to 6%. It shows how the "self-selecting" >>arbitron sample can produce bogus results. >If INTERNET isn't defined (it isn't in most systems on the net, still) news >will reply to the Path: header, and we all know how well THAT works. Well, setting INTERNET didn't work so hot either. I saw our reply to the sendsys go out, as well as several from the 20 sites we feed alt.all. All took the internet route. None bounced. None showed up in the listing. My experience from my Bay Area sendsys is that fewer than 50% of the replies get through. If you tree that out on a national level, I would not be overly surprised if you got a reply from fewer than 25% of the sites receiving sites. However, the claim that arbitron shows a highly irregular number of alt sites makes sense. Obtaining alt and running arbitron both require deliberate action from the sysadmin. Sysadmins who care enough about USENET to run arbitron are almost certainly much more likely to be interested in alt. <csg>
reid@decwrl.dec.com (Brian Reid) (02/27/88)
Did you do a control-group test also? Send out a sendsys to the whole network (not just the alt subpart) to see what you got back there? I don't believe you have actually learned anything at all about the relative sizes of the alt subnet and the complete net until you do that part of the experiment. In the past when I have send out sendsys requests, I have gotten replies from less than 5%-10% of the sites that I thought should be responding. Once you have both the "whole network" and the "alt subnet" counts, then you can take their ratio to get an estimate of how much of the net gets alt groups.
jnp@calmasd.GE.COM (John Pantone) (02/27/88)
We too get alt newsgroups - calmasd.ge.com
--
These opinions are solely mine and in no way reflect those of my employer.
John M. Pantone @ GE/Calma R&D, 9805 Scranton Rd., San Diego, CA 92121
...{ucbvax|decvax}!sdcsvax!calmasd!jnp jnp@calmasd.GE.COM GEnie: J.PANTONEjbs@eddie.MIT.EDU (Jeff Siegal) (02/27/88)
In article <301@bacchus.DEC.COM> reid@decwrl.UUCP (Brian Reid) writes: >[about how you should send a sendsys to the whole net as a control] Also, you might want to try using a different control message (for both parts of the experiment) like "version". Sendsys is easily turned off so that news administrators can keep their connectivity (names of internal machines, etc.) informaiton private. Jeff Siegal
denny@mcmi.UUCP (Dennis Page) (02/29/88)
In article <5077@swan.ulowell.edu> page@swan.ulowell.edu (Bob Page) writes: >About a month ago there was some discussion on alt.config about how >big the alt 'network' was. I sent out a sendsys message to find out. Your data appears to be inaccurate. There are a minimum of four sites in our local area which receive at least alt.sources, do not block the sendsys message, and who's mailpaths are functional. None of these sites (including mcmi) appear in your listings. What newsgroup was the sendsys attached to? -- Denny Page Martha, the Clones are loose again! Don't just stand there... get a net!
rynes@mandrill.CWRU.Edu (Edward M. Rynes Esq.) (02/29/88)
Just a note: While our machine (mandrill.cwru.edu) was in your list
of sites that receive the alt groups, the site that feeds them to us (hal)
and the site that feeds them to them (ncoast) were not. Also, none of the
sites that we feed the alt groups to were listed. So me-thinks that your
data is no more valid than the arbitron data.
______________________________________________________________________________
Edward Rynes Unix Engineer | "Give me what I want, and all I
A. R. Jennings Computing Center | Can think about is losing it.
Case Western Reserve University | I'm losing it!" K.H.
Cleveland, Ohio 44106 |
(216) 368-2982 rynes@mandrill.CWRU.Edu | {cbosgd|decvax|sun}!mandrill!rynes
______________________________________________________________________________page@swan.ulowell.edu (Bob Page) (03/01/88)
Well, I learned a whole lot about the value of sendsys! I didn't send a sendsys to the whole net as a control, nor did I send one out with a UUCP address as a Reply-To: line. If I ever do another one, I'll certainly remember this. I'd like to see a new control message .. 'doesget', where I can send Control: doesget alt and get a response that says Subject: site foovax.uucp: doesget alt YES or Subject: site foovax.uucp: doesget alt NO This way, system admins wouldn't be so paranoid about protecting their sys files, so they probably wouldn't turn this feature off. And I'd only get the information I was looking for, easily parsable by some mail subject reader. This also means I'd have to assume everyone was running the new software that supports the 'doesget' control message! I did another size test in alt: counted every site that had passed or generated an article in the last six weeks. I looked at the Path: line and counted the number of unique sites I found. Total: 852 (I won't bother to list them here). This data isn't a whole lot more reliable than sendsys, since there are presumably lots of leaf sites who don't post, and they won't show up in the total. However, it is 300 more sites than responded to my sendsys request! I can't think of any better ways to find out how many sites carry alt. ..Bob -- Bob Page, U of Lowell CS Dept. page@swan.ulowell.edu ulowell!page "I don't know such stuff. I just do eyes." -- from 'Blade Runner'
reid@decwrl.dec.com (Brian Reid) (03/01/88)
I'm always looking for ways to improve or validate the arbitron data, so I thought I would compare Bob Page's sendsys response with my arbitron data for roughly the same period. I took the list of arbitron respondents that claimed to get the group alt.config, and canonicalized the host names (e.g. changing "3comvax" into "3comvax.3com.com") and then ran my list and Bob's list through "comm". Here are the results. Hosts in arbitron sample: 311 Hosts in sendsys sample: 550 Hosts common to both samples: 161 (29% of sendsys, 51% of arbitron) Total hosts: 711 Hosts in sendsys only: 389 (125% of arbitron) Arbitron only: 173 (31% of sendsys) I don't have any interpretation of this data yet, but I'm posting it so that other people can look at it and try to figure out what it means. It is very clear from this data that neither the arbitron data nor the sendsys data is getting anything close to a majority of the sites out there. As always, the commercial plug: can I convince any of the 389 sites listed in the second column to send in arbitron reports? ---------------------------------------------------------------------------- +Host shown in arbitron data only | | +Host shown in sendsys data only | | | | +present in both | | | | | V | V 3comvax.3com.com V abhg abvax.abnet.com acad academ actisb adelie admin adobe.com agate.berkeley.edu agora ai aimt albanycs.albany.edu alberta altura.srcsip alvin.mcnc.org alxfin amdahl.uts.amdahl.com amdcad.amd.com ames ames.arc.nasa.gov amethyst andrew.cmu.edu aob aplcen apollo apple aramis.rutgers.edu argent argon array arthur.cs.purdue.edu asi astroatc ateng aucs aurora auscso aw.sei.cmu.edu axiom babbage.acc.virginia.edu bacchus banana.uq.oz bbn.com bcd-dyn beach.cis.ufl.edu bellcore.bellcore.com bellman.acc.virginia.edu bigboy blblbl blia.bli.com bloom-beacon.mit.edu bnl bobkat bocar boole.acc.virginia.edu bpa.bell-atl.com briar.philips.com brspyr1.brs.com bsu-cs bu-cs.bu.edu bu-it bucasb bucket bucsb bucsd buengc.bu.edu bungia.bungia.mn.org bunny burdvax.prc.unisys.com cadomin caeco caelum.att.com caip.rutgers.edu calyx canopus cartan.berkeley.edu cascade.stanford.edu catfish catlabs caus-dp cavell cblpe.att.com cblpf.att.com cblph.att.com cblpn.att.com cbmvax cbnea.att.com cbnscs cbosgd.att.com cbrme.att.com cbterra.att.com cc5.bbn.com cci632 ccicpg ccng ccng.waterloo.edu ccv.bbn.com cdes cello.hpl.hp.com cfisun cg-atla chemabs chinet cierra cipric.mn.org cit-vax.caltech.edu claris clyde.att.com coffee.stanford.edu cognos coma compsys.mu.edu cooper copper.tek.com cos.com countach cp1.bell-atl.com cpesac cpsc6a.att.com cpsc6b.cpsc6a.att.com crash.cts.com cs.glasgow.ac.uk csri.toronto.edu csuchico.edu csustan ctnews.convergent.com cucstud cunyvm.cuny.edu cxsea daemen daimler daisy dalcs dalcsug dale.acc.virginia.edu dalsqnt dalton.acc.virginia.edu dandelion.ci.com dartvax.dartmouth.edu dasys1 dayton ddsw1 decvax decwrl.dec.com denning devon diamond.bbn.com dino.ulowell.edu dogie.edu donk doug drexel druco.att.com druhi.att.com druin.att.com druks.att.com drumn.att.com drune.att.com drupa.att.com drusp.att.com drutx.att.com druxh.att.com druxm.att.com druxu.att.com druxv.att.com dsinc e.ms.uky.edu ece-csc eddie.mit.edu edison.ge.com edsel edsel edvax elbereth.rutgers.edu eleazar.dartmouth.edu ellymae elmgate elric elroy.jpl.nasa.gov elxsi ems.ems.mn.org encore eos ephemeral.ai.toronto.edu epimass.epi.com epiwrl.epi.com eplrx7 ernie.berkeley.edu ers escargot etnibsd euclid.acc.virginia.edu euler.rutgers.edu exec fauern fedeva fermat.rutgers.edu flight.foobar foobar.hf.intel.com fredonia friday frog fxgrp galaxy gamma gandalf.littlei gang-of-four garfield gargoyle.uchicago.edu gauss.rutgers.edu gcm geac genesis.att.com genghis gethen gilroy glacier.stanford.edu glimmer gmdzi gondor.cs.psu.edu gpu.utcs.toronto.edu granite.bbn.com granite.dec.com granjon grebyn gronk gryphon gse.entity.com gumby.cs.wisc.edu gvgpsa haddock.isc.com hall.cray.com hammer.tek.com handel hardees.rutgers.edu harvisr hawkmoon hc.dspo.gov hechcx helios.toronto.edu helps hi-csc hi.unm.edu ho2cad ho7cad.att.com hocpa homxa homxc hop.toad.com hoptoad hoser.berkeley.edu hotlg.att hotps.att.com hou2c hou2d hou2e hou2h hou3b houdi houem hound hounx houtz houxa houxh houxj houxk houxl houxs houxu houxv hp-sdd.hp.com hpcea.ce.hp.com hpcea.ce.hp.com hpcupt1.hp.com hpcvaab.hp.com hpcvlx.hp.com hpda.hp.com hpdsla.hp.com hpdslab.hp.com hpesoc1.hp.com hpfcdc.hp.com hpiacla.hp.com hpihoah.hp.com hpisoa2.hp.com hpisod1.hp.com hplabsz.hpl.hp.com hpsemc.hp.com hpsmtc1.hp.com ht.ai.mit.edu hub.ucsb.edu hudson.acc.virginia.edu husc4 husc6.harvard.edu hutch hydra.riacs.edu icus idsssd ig.ig.com iisat ileaf im4u ima.isc.com imagen imagine.pawl.rpi.edu infinet intek01 io.att.com ism780c ista itsgw.rpi.edu izimbra.css.gov j.cc.purdue.edu jade.berkeley.edu jarvis.csri.toronto.edu jasper jlsoft jolnet jpusa1 julian.uwo.cdn jumbo.dec.com kaos kestrel.arpa ki4pv killer kitty kksys klinzhai.rutgers.edu kodak koko krebs.acc.virginia.edu labrea.stanford.edu lakeoz lakesys laticorp laura lclark leadsv.leads.lmsc.com leah.albany.edu left.att.com lehi3b15.csee.lehigh.edu leia.gwd.tek.com lindy.stanford.edu linus lll-tis lll-winken.llnl.gov louie.udel.edu lsuc lzaz lzaz.att.com lzfmd lzfme lznv lznv.att.com lzsc lzsc.att.com m2c.org macom1 madnix magic4 mahendo.jpl.nasa.gov mandrill.cwru.edu marque.mu.edu mas1 math.rutgers.edu maxwell.acc.virginia.edu maynard.bsw.com mccc mcco mcdchg mcnc.org mcp.entity.com me.toronto.edu meccsd.mecc.mn.org meccts.mecc.mn.org megatek mendel.acc.virginia.edu metavax methods mibte midas.tek.com mines minnow mit-amt.media.mit.edu mit-caf mit-eddie mitisft.convergent.com mitsumi mmlai mnetor moogvax mordor.s1.gov moscom mozart mrst mrstve mtfmi mtgzy mtgzz mtunb.att.com mtund.att.com mtune.att.com mtunf.att.com mtung.att.com mtunh.att.com mtuni.att.com mtunj.att.com mtunk.att.com mtunx.att.com mtuxj mtuxo mtuxo.att.com mtxinu munsell n3dmc navajo ncc ncoast ncr-sd ncrcae ncrcan.toronto.ncr.com ncrcpx ncs-med ncsugn.ncsu.edu ncsuvx.ncsu.edu ndcheg ndmath nebula.att.com necis necntc.nec.com netsys newton.rutgers.edu nicmad nis.nis.mn.org nitro noao noether nosun novavax nuchat nucleus nucsrl nud nunki.usc.edu nvanbc obdient oberon.usc.edu obpa1 occrsh.att.com octopus oddjob.uchicago.edu odyssey.att.com ogg olivea.olivetti.com oliveb.olivetti.com olivee.olivetti.com olivej.olivetti.com oliven.olivetti.com omepd onecom orange.qtp.ufl.edu orca.tek.com orchid.waterloo.edu orion.arc.nasa.gov orstcs.cs.orst.edu oswego otter.hple.hp.com otto ovaltine.berkeley.edu oxtrap ozvax pacbell packard palo-alto panda papaya.bbn.com parsely paul.rutgers.edu pbhya pbhyb pbhyc pbhyd pbhye pbhyf pbhyg pcrat pdp.cs.ohiou.edu pegasus peora.ccur.com percival philabs.philips.com phoenix.princeton.edu phoenix pixar plnapc.philips.com plus5 pogo.tek.com polecat polya.stanford.edu polygen polyslo popeye portal poseidon princeton.princeton.edu psu-cs psuvax1.psu.edu psuvm.bitnet psychsun.stanford.edu pt.cs.cmu.edu ptsfa pucc.bitnet puck puff.cs.wisc.edu puffin.uss.tek.com pwcs.stpaul.gov pyramid pyrdc pyrnj pyrps5 pyrtech qetzal qiclab quest questar questar.questar.mn.org ra.rice.edu radio.toronto.edu rayssd rayssdb rayssde redwood reed remus.rutgers.edu rencon rhesus.primate.wisc.edu richp1 ritcv rochester rocky.stanford.edu rolls rosevax.rosemount.com rtech ruby.tek.com rutgers.rutgers.edu sandoz sarin sauron sbo scgvaxd scicom scobee.att.com scooter sdcjove.cam.unisys.com sdcrdcf sdcsmb.sm.unisys.com sdcsvax.ucsd.edu sdn sgi.sgi.com shamash shark.tek.com shasta.stanford.edu shrike.austin.lockheed.com sialis.mn.org siemens sigmast silver.bacs.indiana.edu sis sjfc sli smegma snark snjsn1.sj.ate.slb.com sol.arpa soleil spar.spar.slb.com spdcc.com spectrix speedy.mcnc.org sphinx.uchicago.edu splut spool.cs.wisc.edu spp3 spt.entity.com sq.com sri-spam.istc.sri.com starfish step.dwr.com stsci studsys.mu.edu sunybcs super.upenn.edu sw1e swan.ulowell.edu swatsun swlabs t9103 tahoe.unr.edu tandem td2cad.intel.com teemc tekcae.tek.com tekchips.tek.com tekcrl.tek.com tekecs.tek.com tekfdi.tek.com tekgen.tek.com tekgvs.tek.com tekig.tek.com tekig4.tek.com tekig5.tek.com tekigm2.tek.com tekirl.tek.com teklds.tek.com tekred.tek.com teksce.sce.tek.com tektronix.tek.com telesoft temvax terminus thelink thorin.cs.unc.edu ti-csl.csnet titan.arc.nasa.gov titan.arpa tmsoft toccata.rutgers.edu topaz.rutgers.edu trillium trwrb.dsd.trw.com trwrc tslanpar ttidca.tti.com tub turtlevax tut.cis.ohio-state.edu twitch tymix ucbarpa.berkeley.edu ucbvax.berkeley.edu ucdavis.ucdavis.edu ucqais ucsd.edu udccvax1.acs.udel.edu uflorida.cis.ufl.edu ufqtp ukecc.engr.uky.edu umbc3.umd.edu umd5.umd.edu umix.cc.umich.edu umn-cs.cs.umn.edu umn-cs.cs.umn.edu umn-d-ub.d.umn.edu uni2.bcm.tmc.edu unido unisoft uokmax uoregon uport usceast ut-emx ut-sally utah-gr uthelios uthub.toronto.edu utmanitou.toronto.edu uunet.uu.net uvaarpa.virginia.edu uvacs.cs.virginia.edu uvaee.ee.virginia.edu v.sei.cmu.edu valid van-bc vaxine vice.tek.com videovax.tek.com violet.berkeley.edu viper.lynx.mn.org virginia.acc.virginia.edu vixie voder.nsc.com vu-vlsi wa3wbu watale.waterloo.edu watcgl.waterloo.edu watdcsu.waterloo.edu watdragon.waterloo.edu water.waterloo.edu watmath.waterloo.edu watt.acc.virginia.edu watvlsi.waterloo.edu watyew wb3ffv westmark whuts wiley wp3b01 wucs1 wucs2 xanth.cs.odu.edu xicom xray xroads yetti yunexus zeus.hf.intel.com zeus.tek.com zorch.uu.net zycad
matt@oddjob.UChicago.EDU (Mr. nEtural) (03/01/88)
Jeff Siegal writes:
) Sendsys is easily turned off so that news administrators
) can keep ... informaiton private.
It's been a long time since I read the instructions. Does the
installation guide no longer say, "The contents of your sys file is
public"?
Mattper@erix.UUCP (Per Hedeland) (03/01/88)
In article <1939@epimass.EPI.COM> jbuck@epimass.EPI.COM (Joe Buck) writes: >In article <5077@swan.ulowell.edu> page@swan.ulowell.edu (Bob Page) writes: >>About a month ago there was some discussion on alt.config about how >>big the alt 'network' was. I sent out a sendsys message to find out. >> Since I only got 550, I'd say rather than 66% of the network >>sites carrying alt.sources, it might be closer to 6%. >Bob, the flaw in your analysis (in getting your 6% figure) is >assuming that your sendsys reached all the sites on the altnet >and that all of them successfully sent a reply to you. Actually, there's another flaw: "the size of the alt 'network'" (by some definition) isn't necessarily the same as "the percentage of the network getting alt.sources". We (and I believe many other in Europe) belong to the latter group, but maybe not the former (depending on the definition). And no, we didn't get the sendsys message, and no, I don't know the definition (i.e. what group the sendsys was sent in - alt.config.ctl?). --Per Hedeland per@erix.ericsson.se or ...{mcvax,uunet}!enea!erix!per
page@swan.ulowell.edu (Bob Page) (03/02/88)
Yesterday I posted another list, based on an analysis of the Path:
lines. Well, the script I used also counted usernames as host names
(ye old fencepost error strikes again). The real count is 477 unique
sites that have passed an alt group towards ulowell in the past six
weeks, not the 852 that I originally came up with.
This is 73 less sites than were reported in the sendsys sample.
However, 284 of those hosts did not respond to my sendsys message,
making a combined total of 834 hosts.
In addition, 116 hosts did not respond to sendsys AND did not pass an
article, BUT filed an arbitron report claiming somebody there reads
at least one alt group.
Total hosts: 950 sites, plus or minus a few (due to duplicate host
names that I didn't catch) that get alt. This does not include
those sites that:
1. did not post an article in the past six weeks,
2. did not pass an article towards ulowell in the past six weeks,
3. did not respond to my sendsys control message, and
4. did not file an arbitron report.
How many sites fall into that category? Nobody can tell, although
I suspect there's quite a few.
I'd still like to see a better way of gathering network statistics.
..Bob
PS Here's how I got the list of hosts based on Path: lines.
#!/bin/sh
cd /usr/spool/news/alt
egrep "^Path: " */* |
cut -d\ -f2 |
awk -F! '{ for (i = 1; i < NF; i++) print $i }' $* |
tr A-Z a-z |
sort |
uniq
--
Bob Page, U of Lowell CS Dept. page@swan.ulowell.edu ulowell!page
"I don't know such stuff. I just do eyes." -- from 'Blade Runner'rimbold@dino.cpe.ulowell.edu (Rob Rimbold) (03/02/88)
In article <303@bacchus.DEC.COM> reid@decwrl.UUCP (Brian Reid) writes:
+
+ As always, the commercial plug: can I convince any of the 389 sites listed in
+ the second column to send in arbitron reports?
+
+ +Host shown in arbitron data only
+ |
+ | +Host shown in sendsys data only
+ | |
+ | | +present in both
+ | | |
+ | | V
+ | V
+ V
+ dino.ulowell.edu
Hmm.. we *do* run arbitron. At least, we haven't yet been listed as expired.
Please check and let me know if you did indeed receive our arbitron data.
--
'Rob
Usenet: rimbold@dino.ulowell.edu
AT&T : (617) 452-5000 x2693
University of Lowell Productivity Center, Lowell MA.emv@fleetwood.cc.umich.edu (Edward Vielmetti) (03/02/88)
In article <5184@swan.ulowell.edu> page@swan.ulowell.edu (Bob Page) writes: >..Bob > >PS Here's how I got the list of hosts based on Path: lines. > >#!/bin/sh >cd /usr/spool/news/alt >egrep "^Path: " */* | > cut -d\ -f2 | > awk -F! '{ for (i = 1; i < NF; i++) print $i }' $* | > tr A-Z a-z | > sort | > uniq >-- If you have MH handy, it's relatively easy to generate the path info for a newsgroup. In my MH directory (~/Mail) I've linked Alt to /usr/spool/news/alt, Comp to /usr/spool/news/comp, etc. Then a line like this: scan +Alt/sources last:20 -format "%(size)\t%{Path}" generates size and path information for further munging by awk. --Ed Edward Vielmetti, U of Michigan mail group.