rcj@burl.UUCP (Curtis Jackson) (03/03/86)
I was going to send the following to the person who just posted the request for etiquette in creating net.books.sf, but I thought it might be of general interest. See if the stuff below appears sound to you, mail me comments/criticisms/suggestions, and I'll try to work it into something that might be included in net.announce.newusers or posted to this newsgroup with a long expiration time. If there is already something like this in net.announce.newusers that I've missed just drop me mail to that effect and accept my apologies in advance: a) Post a request for votes in groups where there might be substantial interest; make sure that one of those groups is net.news.group. [Here that would probably be: net.news.group,net.books,net.sf-lovers] b) Stress VERY STRONGLY in your posting that you want votes via mail ONLY! c) Ask the 'yes' voters to indicate whether they would read the group, post to the group, or both. d) Put a line in your article header that reads: Followup-To: net.news.group so any further discussion generated by the 'f' key will go to net.news.group only. e) Note in your posting that followups should only go to net.news.group (protects against ancient software that doesn't process the Followup-To line). f) During the time you are collecting replies, subscribe to net.news.group because some votes will invariably be posted there (usually, and hopefully, as afterthoughts in articles discussing issues involved in the creation of the group in question) and you'll want to gather them, too. g) Keep all replies (in case any suspicious person asks for verification). When they have trickled off to next to nothing (usually about 2 weeks), summarize the results to net.news.group. h) If the results are positive enough (both # of responses and percentage of 'yes' votes), talk your news admin into posting an announcement (in the same posting as your summary is fine) saying that the group will be created on date X unless some people bitch NOW. Since the decision on how many responses and what percentage of 'yes' votes is required to create a given group is highly subjective (at least at this point in USENET evolution), this is a good way to get the opinion of concerned net citizens on whether your estimation that there *is* enough interest to create the group is valid or not in their eyes. i) Assuming you are still alive and relatively unscorched after the results of (h) above, on date X, have your news admin create the group. Let me hear from you, -- The MAD Programmer -- 919-228-3313 (Cornet 291) alias: Curtis Jackson ...![ ihnp4 ulysses cbosgd mgnetp ]!burl!rcj ...![ ihnp4 cbosgd akgua masscomp ]!clyde!rcj
lauren@vortex.UUCP (Lauren Weinstein) (03/03/86)
Greetings. I'm not going to dig up all the details again, but as I've pointed out in the past: 1) Voting on Usenet is flawed. If nothing else, we should be counting ONE VOTE PER SITE, not one vote per person. And any form of voting in this environment is of only minimal value in any case. If sites are going to be counted, then the RATIO of sites that vote "yes" to the TOTAL NUMBER OF SITES on the network should be the issue, not the absolute number of sites. The point is that it should be necessary for a SIGNIFICANT fraction of the TOTAL network to be interested in a topic before newgroup creation is even considered. 2) Human nature is such that you'll almost always receive more yes votes than no votes. The people interested respond "yes," while the people who don't care about such a group simply ignore the whole issue and never bother voting "no." Thus, the fact that you get plenty of yes votes may still often mean that many more people thought the group was worthless, but didn't have the time nor inclination to send in a vote. Of course, the people who WANT the group are highly motivated to vote yes. 3) Simply because some number of people want a group doesn't make it appropriate for Usenet newsgroups. You'll recall my scenario where I suggested that even the most bizarre topics could achieve a threshold of yes votes, but that that alone didn't make the creation of a newsgroup appropriate. 4) The issues of mailing lists vs. newsgroups are still far from fully understood. I personally believe that most of the statistics posted on this issue (thresholds for newsgroups vs. mailing lists) are currently misleading and cannot be relied upon until more complete research is done into this issue. We just don't have enough information about this yet. --Lauren--
roy@phri.UUCP (Roy Smith) (03/04/86)
In article <886@vortex.UUCP> lauren@vortex.UUCP (Lauren Weinstein) writes: > >1) Voting on Usenet is flawed. If nothing else, we should be counting > ONE VOTE PER SITE, not one vote per person. [...] The point is that it > should be necessary for a SIGNIFICANT fraction of the TOTAL network to > be interested in a topic before newgroup creation is even considered. If you demand that you get a quorum, nothing will ever get done. A quick count of the votes in my "delete net.general" campaign (results to be announced any day now) shows only about 150 people voting. This is what, maybe a couple percent of the people on the net? Even if you defined "significant" as 10% of the users, I doubt you could get that many to vote. Why should a single-user site have as big a vote as a site with 100 users? This is a non-issue anyway. Until you can get lots more people to vote, it works out about the same either way. Most sites don't vote, and most sites that do vote only have one person voting. -- Roy Smith, {allegra,philabs}!phri!roy System Administrator, Public Health Research Institute 455 First Avenue, New York, NY 10016
greenber@phri.UUCP (Ross Greenberg) (03/04/86)
I tend to agree with my SA (Hi roy!)...Voting is the best mechanism we got, and the flaws Lauren mentioned (I feel) are not valid. Roy says he got about 150 votes! This is pretty astounding -- usual surveys get only about 25-50 responses. Roy: Howzabout you do a little stat work: how many votes per site, how many votes per backbone site, how many SA's voting, etc. -- ------ ross m. greenberg ihnp4!allegra!phri!sysdes!greenber [phri rarely makes a guest-account user a spokesperson. Especially not me.]
campbell@maynard.UUCP (Larry Campbell) (03/05/86)
> 1) Voting on Usenet is flawed. If nothing else, we should be counting > ONE VOTE PER SITE, not one vote per person... This sounds reasonable at first. However, a perhaps unintended side effect is to give excessive influence to the owners of "fuzzball hosts" (like vortex and maynard!). Should someone with three IBM PCs in his basement really count more than 100 users of ihnp4? > 4) The issues of mailing lists vs. newsgroups are still far from fully > understood. I personally believe that most of the statistics > posted on this issue (thresholds for newsgroups vs. mailing lists) > are currently misleading and cannot be relied upon until more > complete research is done into this issue. We just don't have > enough information about this yet. Agreed. But what should we do? Never create newsgroups until the issues are undersood? How about a proposal for a way to research this question? (Hmm, if I had more time... all the necessary information is in the uucp map, it just needs to be crunched...) -- Larry Campbell The Boston Software Works, Inc. ARPA: maynard.UUCP:campbell@harvard.ARPA 120 Fulton Street UUCP: {harvard,cbosgd}!wjh12!maynard!campbell Boston MA 02109
lauren@vortex.UUCP (Lauren Weinstein) (03/06/86)
It would appear to me that the number of users on a site is largely irrelevant to the issue of "voting." That is, I see little if any reason why a site with 30 users should actually get more voting weight than a site with 1 or 2 users. The key is traffic topology. Only one copy of any given netnews need be sent to any given site regardless of their netnews-reading population. And it's that traffic that is the main issue under discussion here. To put it another way, simply because a site has 20 people who might want to see one group, doesn't mean that 20 copies of that group need be sent to that site! They need exactly as many copies (one) as a small site. The issue isn't really whether or not lots of people want to discuss a given topic. The issue is whether or not it makes sense to use newsgroups to handle the traffic involved in discussing such a topic. An extreme example: Large site foo, with 30 users who actually read netnews, and large site bar, with another 30, see a vote request to discuss styrofoam in a proposed new group net.foam. As it turns out, foo and bar's people have a thing about stryofoam, and send in 60 yes votes. Not too many other people care one way or another so there aren't many no votes. Is this an appropriate situation for creating a new newsgroup? I think most people would agree that it is not. The few sites interested could pass the material around via mail much more efficiently, and save the entire network a great deal of expense (and possibly a lot of disk space as well). Of course this is an extreme case, but it serves to point out that the number of users on a given site isn't as important as the ratio of interested sites when it comes to the dynamics and topology involved in newsgroup creation and traffic flow. --Lauren--
smith@ncoast.UUCP (Phil Smith) (03/07/86)
> Article <886@vortex.UUCP> > From: lauren@vortex.UUCP (Lauren Weinstein) > Greetings. I'm not going to dig up all the details again, but as I've > pointed out in the past: > > 1) Voting on Usenet is flawed. <plus comments on why> > > 2) Human nature is such that you'll almost always receive more yes > votes than no votes. The people interested respond "yes," while > the people who don't care about such a group simply ignore the > whole issue and never bother voting "no." Thus, the fact that you > get plenty of yes votes may still often mean that many more people > thought the group was worthless, but didn't have the time nor > inclination to send in a vote. Of course, the people who WANT the > group are highly motivated to vote yes. > Why not use a method based on the way smith@phri conducted the net.general vote. But set it up so if you simply reply your vote will be NO. That way it will be easy for us lazy people to vote NO. -- ncoast!smith@Case.CSNet (ncoast!smith%Case.CSNet@CSNet-Relay.ARPA) ..decvax!cwruecmp!ncoast!smith ncoast is dead, long live ncoast!
mc68020@gilbbs.UUCP (03/07/86)
In article <258@maynard.UUCP>, campbell@maynard.UUCP (Larry Campbell) writes: > > 1) Voting on Usenet is flawed. If nothing else, we should be counting > > ONE VOTE PER SITE, not one vote per person... > > This sounds reasonable at first. However, a perhaps unintended side > effect is to give excessive influence to the owners of "fuzzball hosts" > (like vortex and maynard!). Should someone with three IBM PCs in his > basement really count more than 100 users of ihnp4? No, not *MORE* important. Equally important, yes. But I do not agree with Lauren on his one vote per site stance. I gather that you (Larry) and Lauren run essentially single user sites in your homes. While I am not a single user site (2.5, maybe? (-: ), I do run my system in the closet of my bedroom (does this mean I'm 'coming out of the closet"? (-: ). I do most vehemently object to being identified as a "fuzzball host" simply because my computer doesn't represent hundreds of thousands of dollars worth of hardware, and I don't have a big, impressive sounding name for my organization. -- ==================================== Disclaimer: I hereby disclaim any and all responsibility for disclaimers. tom keller {ihnp4, dual}!ptsfa!gilbbs!mc68020 (* we may not be big, but we're small! *)
lauren@vortex.UUCP (Lauren Weinstein) (03/07/86)
If even a significant fraction of the total Usenet readership actually DID respond to vote requests, the amount of traffic generated, all pointed at the vote counter (be it a human or an automated system) would be truly immense. Right now, so few people (relatively) respond to vote requests that this issue hasn't appeared, but if 10's of 1000's of people DID start to vote... well... --Lauren--
laura@hoptoad.uucp (Laura Creighton) (03/09/86)
In article <886@vortex.UUCP> lauren@vortex.UUCP (Lauren Weinstein) writes: >Greetings. I'm not going to dig up all the details again, but as I've >pointed out in the past: > >1) Voting on Usenet is flawed. If nothing else, we should be counting > ONE VOTE PER SITE, not one vote per person. And any form of voting > in this environment is of only minimal value in any case. If sites > are going to be counted, then the RATIO of sites that vote "yes" > to the TOTAL NUMBER OF SITES on the network should be the issue, > not the absolute number of sites. The point is that it should be > necessary for a SIGNIFICANT fraction of the TOTAL network to be > interested in a topic before newgroup creation is even considered. This would be a valid argument if the popostion ``I am going to create net.frozzbozz and then traffic about frozzbozzes will increase dramatically''. But this doesn't happen. What happens is ``there is a lot of frozzbozz traffic in net.arch ... if we don't get it out of here then a lot of people will unsubscribe. But people want to discuss frozzbozzes -- so give them their own group.'' I am one of the people who unsubscribed to net.sf-lovers because of the high non-book content. The last time there was a proposal for net.sf-lovers.movies (or net.sf-lovers.starwars) I lost out -- I wanted the movie goers to go their own way. As a result, I don't read net.sf-lovers any more. The greatest problem with voting is that people use it to express their approval of the disirability of *anybody* discussing *anything* about the topic. A while ago, there was a discussion in net.micro.68k about os9. Hoptoad is a 68K system (Sun 3). I can't remember whether it was John Gilmore or I who finally got pissed enough to suggest that the os9'ers move to their own newsgroup. But there was a ton of followups which said ``OS9 is useless!! Don't give it its own nesgroup!!!'' This is the wrong attitude to take. The attitude is ``OS9 is useless. Get it the hell out of net.micro.68k!! Since the OS9'ers show no sign of shutting up, put them somewhere!!!'' [Note -- I don't know beans about OS9. I don't think that it is useless. There are pleanty of useful things which I am not interested in. It is just fresh in my memory. > >2) Human nature is such that you'll almost always receive more yes > votes than no votes. The people interested respond "yes," while > the people who don't care about such a group simply ignore the > whole issue and never bother voting "no." Thus, the fact that you > get plenty of yes votes may still often mean that many more people > thought the group was worthless, but didn't have the time nor > inclination to send in a vote. Of course, the people who WANT the > group are highly motivated to vote yes. So what? If I were the *only* person who didn't care about OS9 and nobody minded having 2 groups, why shouldn't I get what I want? It is easy to read both net.micro.68K and net.micro.OS9 right after. Nobody loses under this scheme, and even if I were the only person who won -- would that be such a bad thing? -- Laura Creighton ihnp4!hoptoad!laura utzoo!hoptoad!laura sun!hoptoad!laura toad@lll-crg.arpa
laura@hoptoad.uucp (Laura Creighton) (03/09/86)
In article <891@vortex.UUCP> lauren@vortex.UUCP (Lauren Weinstein) writes: >An extreme example: > >Large site foo, with 30 users who actually read netnews, and large site >bar, with another 30, see a vote request to discuss styrofoam in a proposed >new group net.foam. As it turns out, foo and bar's people have a thing >about stryofoam, and send in 60 yes votes. Not too many other people >care one way or another so there aren't many no votes. Is this an >appropriate situation for creating a new newsgroup? I think most people >would agree that it is not. Nope. I think that most people would agree that it is. What really would happen is that the people on foo and bar would start sending messages about styrofoam to net.singles (styrofoam couches as a great way to conclude a date) and net.periphs (styrofoam packaging as noise sheilding) and net.sports and net.rec (styrofoam surf boards rewconsidered as a result of lacquer 123) and net.consumers (400 uses for a piece of styrofoam). I mean, they are *really interested*!!. What options do you have? Maybe you can convince the foo'ers and bar'ers to set up a mailing list. This can be tried, irrespective of setting up net.foam. If not, your choices are -- put up with them in your favourite newsgroup, unsubscribe to that newsgroup, or give them their own place to post and unsubscribe from that. A mailing list would be the best solution here -- but unsubscribing to net.singles and net.periphs and net.sports and net.rec and net.consumers simply because you don't like the volume of styrofoam articles would be the worst. -- Laura Creighton ihnp4!hoptoad!laura utzoo!hoptoad!laura sun!hoptoad!laura toad@lll-crg.arpa
jay@ethos.UUCP (Jay Denebeim) (03/16/86)
Now, I was wondering about something. What percentage of the USENET comunity is using RN? The Kill feature works great in this situation, you can get rid of all the stuff you don't like, by subject, entire header, or entire article. Nice feature... -- Jay Denebeim "One world, one egg, one basket." {seismo,decvax,ihnp4}!mcnc!rti-sel!ethos!jay Deep Thought, ZNode #42 300/1200/2400 919-471-6436