chuqui@nsc.UUCP (Chuq Von Rospach) (09/06/85)
This is to some degree a followup on the whole discussion of vote fraud
and voting on new newsgroups. For a long time I've felt that voting on new
newsgroups is wrong but we just haven't come up with a better way yet.
One of the real problems with the way we create groups is that there is
absolutely no correlation between the sites that pay the bills for the
net and the sites that coordinate where the net is going. A few major
sites and companies are spending a lot of money subsidizing the net,
and there are a lot of sites that simply don't carry their own weight
on the costs of supporting this network.
I think we need to realize that the net isn't a free lunch, and give
the people that pay the ride a stronger say in what goes on.
Lauren recently suggested a one site, one vote proposal. I think we
need to go one step further -- each site gets a weighted vote based on
the support they feed into the network.
I propose the following rule -- each site can generate a weighted vote
on any subject. The weight of the vote is equal to the number of
OUTGOING news feeds a site supports. This means that a backbone such
as hplabs would have a weighted vote of about 7, nsc would be 3, and a
leaf node that feeds news to nobody gets no vote at all. The weights
can be generated from the mod.map data, giving people a reason to get
their data in their and keep it up to date -- if you aren't on the map,
you don't exist. We exclude feeds between machines in the same building
since that isn't supporting the outside world.
Certain sites that add support to the net outside of normal feeds --
heavily used mail sites, network gateways, or sites with people
involved in the administrative or maintenance of the net (people like
Spaf, Mark, Rick, and Lauren) get increased weights as well.
This means that the backbones have a much stronger say in what happens,
which is how it should be. Sites like decvax, ihnp4, hplabs, and the like
put a LOT of money into the net, and it is time people realize that there
ain't no free lunch. As it currently stands, someone from decvax (which was
putting in 100K a MONTH in phone bills a while back) has no more influence
in the net than the person in Moline with a 20 megabyte TRS16 and a 300
baud modem. This proposal means that the people who pay the bills have a
better chance of deciding where that money goes to, and the people who
don't pay the bills can't force their will on the net.
This makes it quite simple -- like most of the world, you pay with your
wallet. If you don't like the way the net is run, you can put the
investment into it needed to fix it. If you don't, you get what you pay
for.
--
Chuq Von Rospach nsc!chuqui@decwrl.ARPA {decwrl,hplabs,ihnp4}!nsc!chuqui
Son, you're mixing ponderables againlaura@l5.uucp (Laura Creighton) (09/10/85)
Guys, this business about ``one vote per site'' doesn't fix the problem. The people I want to hear from are not necessarily from different sites. Moreover, the usefulness of any group is not a function of how low-volume it is. 4 people giving incorrect answers is no better than 400 people doing the same. The 400 people will cost more,but if cost is now more important than quality or accuracy of information then usenet has changed a lot in the 5 months I've been away. What do we need? Automatic removal of non-active newsgroups. Hey, this has been discussed for years! The ability to remove newsgroups that are no longer wanted. Net.bizarre and net.general have bit it. Now what? The ability to stop certain long winded discussions before we generate another meg that we do not need. I remember a long time ago there used to be a ``what is the meaning/etymology of FOOBAR'' discussion every 3 weeks or so. DO NOT POST, BUT SEND ME MAIL IF THIS IS NEWS TO YOU. *thank you*. What we did was to refuse to answer that question any more except by mail. I think we need a new list of what is overdiscussed, and what is in danger of being discussed again. (I just read net.lang.c. I fear we may go into ``this is my brace style and if you don't use it then your mother was a hamster and your father smelled of elderberries'' again.) Net.announce and net.announce.newusers is not the place to put this. The average new user has no way of knowing that he should read this unless he is told to do so. if he doesn't get told he will go read news about the things he knows he is interested in, and he has no way of knowing that net.announce.newusers was designed for him. By the time he reads mod.newslists (if he ever does) it will be too late. Besides, the document is huge. I am not sure that people actually read it -- and it only takes those that don't to start this all over again. What we need is discussion killers in the newsgroups where we don't want the discussion to start. These would be non-expirable news articles which are short. A new user will get as his first article in net.lang.c ``C does not have a generic null pointer. if you think it does, you are wrong. if you want to argue the point send mail to Guy Harris at sun!guy who has written 700 articles on the subject and will mail you his most recent one''. Similarly the first article in net.columbia: ``this is the name of the newsgroup. Please don't agitate to have it changed to net.shuttle. We have argued this 47 times already. If you want to argue this some more, send mail to research!alb who has argued this many times in the past and he will send you his most convincing argument.'' If we killed some more perennial discussions, and stopped the flow of bad information we would do much better than trying to restrict the amount of news per user or per site. The per site idea is a bad one, by the way. It penalises places like sun or research who use the concept of a gateway machine. Sun has a user population of about 1000 and l5 has a population of 5. Treating them the same way would be unfair. There are also certain people whose stuff is consistently good. I don't care how much they post, and don't want to penalise them over some ``law of averages''. -- Laura Creighton (note new address!) sun!l5!laura (that is ell-five, not fifteen) l5!laura@lll-crg.arpa
mark@cbosgd.UUCP (Mark Horton) (09/11/85)
In article <3215@nsc.UUCP> chuqui@nsc.UUCP (Chuq Von Rospach) writes: >I propose the following rule -- each site can generate a weighted vote >on any subject. The weight of the vote is equal to the number of >OUTGOING news feeds a site supports. This means that a backbone such >as hplabs would have a weighted vote of about 7, nsc would be 3, and a >leaf node that feeds news to nobody gets no vote at all. The weights >can be generated from the mod.map data, giving people a reason to get >their data in their and keep it up to date -- if you aren't on the map, >you don't exist. We exclude feeds between machines in the same building >since that isn't supporting the outside world. I think we need to go further than even this. There should be a transitive closure rule applied - you get credit for everybody downstream from you if all their news goes via you, in addition to your direct neighbors. Thus, the total number of votes the backbone sites get would be roughly the number of hosts on the network. Secondary hubs get smaller numbers - essentially one vote for each site they feed. Leaves get one vote. The fact is that the backbone hosts are carrying virtually the entire load on the network. Their generousity is what has kept the net going as long as it has. If they get too fed up with what they are paying for, they have every right to pull the plug. So they should have the major voice in deciding what new groups to create (or even if old abused groups should be deleted.) For anyone who feels that being a backbone site is some kind of elite fraternity, rest assured that all you have to do to become a backbone host is to volunteer (and to have 3 links to other backbones.) There is nothing to stop people from setting up an alternative backbone with different policies. For example, if the current backbone were to decide to kill (say) net.bizarre, an alternative could be set up for, say, talk.all, and as long as that group is willing to pay to transmit the groups talk.all, everybody is happy. Mark
greg@hwcs.UUCP (Greg Michaelson) (09/13/85)
> One of the real problems with the way we create groups is that there is > absolutely no correlation between the sites that pay the bills for the > net and the sites that coordinate where the net is going. A few major > sites and companies are spending a lot of money subsidizing the net, > and there are a lot of sites that simply don't carry their own weight > on the costs of supporting this network. > > I think we need to realize that the net isn't a free lunch, and give > the people that pay the ride a stronger say in what goes on. The White House computer lives! Reagonomics hits the net! How about taxes on wealthy net users for redistribution to us impecunious sites: 'Netaid' on the lines of 'Medicaid'?
arnold@ucsfcgl.UUCP (Ken Arnold%CGL) (09/16/85)
In article <1471@cbosgd.UUCP> mark@cbosgd.UUCP (Mark Horton) writes: >In article <3215@nsc.UUCP> chuqui@nsc.UUCP (Chuq Von Rospach) writes: >>I propose the following rule -- each site can generate a weighted vote >>on any subject. The weight of the vote is equal to the number of >>OUTGOING news feeds a site supports. > >I think we need to go further than even this. There should be a transitive >closure rule applied - you get credit for everybody downstream from you >if all their news goes via you, in addition to your direct neighbors. Fine, actually. I could live with either of these voting systems, for many reasons already enumerated by others. However, before we all weigh in on this too much, maybe we should have some discussion on how we determine how a site's vote goes. Maybe I have missed something, but I don't remember any definitive discussion on this. Certainly we don't want to make the site administrator the "elector" for a site (i.e., the SA decides for him/herself what the vote shall be). Does the SA become the local vote counter, and the majority rules? Maybe only the highest ranking individuals at a site should vote, which would be consistent with the ideas that the people who pay the bills make the decisions. It is, after all, the managers who decide to pay, or allow the payment of, the usenet bills. Before everyone starts buying into this per-site vote bit, we better decide what we mean by a vote. If we can't agree on that, it is rather pointless to decide that it would be nice to use a per-site vote, since we'll never get one. Ken Arnold
peter@graffiti.UUCP (Peter da Silva) (09/17/85)
> I propose the following rule -- each site can generate a weighted vote > on any subject. The weight of the vote is equal to the number of > OUTGOING news feeds a site supports. This means that a backbone such > as hplabs would have a weighted vote of about 7, nsc would be 3, and a > leaf node that feeds news to nobody gets no vote at all. The weights > can be generated from the mod.map data, giving people a reason to get > their data in their and keep it up to date -- if you aren't on the map, > you don't exist. We exclude feeds between machines in the same building > since that isn't supporting the outside world. YES! As a leaf site administrator, I hereby vote to disenfranchise myself. -- peter@baylor.UUCP
jerry@oliveb.UUCP (Jerry Aguirre) (09/20/85)
> I think we need to go further than even this. There should be a transitive > closure rule applied - you get credit for everybody downstream from you > if all their news goes via you, in addition to your direct neighbors. > From: mark@cbosgd.UUCP (Mark Horton) According to this formula I get >1800 votes because as near as I can figure it every site on the net is "downstream" from me! I see a lot of this kind of thinking on the net and I can't understand whether it represents a mistaken view of the usenet topology or a desire to impose a different topology. The terminology of "upstream" and "downstream" really only has meaning when applied to the path taken by a specific article. Thus for most postings I am "downstream" from hplabs but for any postings on other sites I feed then hplabs is downstream from ME. Granted that I can be considered downstream from hplabs because of volume of transmissions, that is not always an easy decision to make. It is not necessary (or in my oppinion desirable) for the net to follow a star topology. It is perfectly possible for two "leaf" sites to open a full news feed between them. If you are going to base "upstream" and "downstream" on the volume of transmissions then how are you going to collect that data? The only source that I know of is supplied voluntarily by the sites themselves. The weekly news activity reports for our area present a bizzare image of connectivity not readily apparent from the usenet map. While I aggree that the sites that most support the net should have a greater say in the control of the net I am not sure that a weighted voting scheme is the answer. Should a leaf with 40 news readers get less of a vote than a "backbone" with 5? Should the news administrator be the only one to vote? Do I get to count my 4 internal systems as votes? If so and I feed a dozen micros instead of one mega do I get more votes even though they represent the same number of users? I think that the real answer is that the backbone sites already do have a bigger vote. They can just refuse to create or cary a news group that they do not feel is justified. This is fair in that if you want someone else to pay for sending your group around the net then you have to present a good argument for doing so. Otherwise your group remains a local one. Jerry Aguirre @ Olivetti ATC {hplabs|fortune|idi|ihnp4|tolerant|allegra|tymix|olhqma}!oliveb!jerry
mojo@kepler.UUCP (Morris Jones) (09/23/85)
In article <209@graffiti.UUCP> peter@graffiti.UUCP (Peter da Silva) writes: >YES! As a leaf site administrator, I hereby vote to disenfranchise myself. I'm a leaf site too, and I like the scheme. I'm forever grateful to the folks upstream from me. I gladly defer to the judgement of those paying the bills. But the issues raised about newsgroups being decided now solely by usenet administrators is a valid one. I could vote against net.vms just to be mean. Still I think there just might be a higher reasonability percentage among the site administrators. As Lazarus Long says (paraphrase) , "Democracy is based on the premise that a million men are wiser than one man. How's that again? I missed something." (Then he of course points out the equivalent ludicrousness of the opposite viewpoint.) -- Mojo ... Morris Jones, MicroPro Product Development {dual,hplabs,glacier,lll-crg}!well!micropro!kepler!mojo
peter@graffiti.UUCP (Peter da Silva) (09/25/85)
> Before everyone starts buying into this per-site vote bit, we better > decide what we mean by a vote. If we can't agree on that, it is rather > pointless to decide that it would be nice to use a per-site vote, since > we'll never get one. > Ken Arnold How about just letting each site handle the voting as it sees fit?