Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!rpi!tale From: tale@cs.rpi.edu (David C Lawrence) Newsgroups: news.misc Subject: Re: Articles most often cited in comp.all Message-ID: Date: 28 Dec 89 19:57:32 GMT References: <6872@tank.uchicago.edu> <4510@hydra.gatech.EDU> Organization: Rensselaer Polytechnic Institute, Troy NY Lines: 119 In <6872@tank.uchicago.edu> jill@tank.uchicago.edu (jill holly hansen): > could someone please tell me what the value of this BIWEEKLY posting is? In article <4510@hydra.gatech.EDU> gt1342a@prism.gatech.EDU (LEDUC) writes: > Hey, _I_ like it. You're one of few; I think this is the first public posting I've seen in support of this. A lot of other people seem to think that the information which it communicates is pretty useless in a few regards. Lets look at the last one posted: } Here are ranks, counts, and Message-IDs (and subjects, authors, } and newsgroups if available) of the articles that appeared So why bother if you (not you, LEDUC. news@elsie.UUCP) don't provide that other information? Item 10 on the list is an excellent example. Only three of them made this list; previous postings have had many more than this. Apparently the referenced article expired so the other information was not available. Ok. So what use is the fact that it was #10 on the list? Very, very few sites keep full articles around for more than fourteen days such that someone could retreive that article if they wanted to because they saw it in this list. } most often in comp.all References: lines in articles received here } over the last 14 days. What is this telling us though? Is it an attempt to say that they were good articles? Or that they were bad articles which got flamed a lot? Or something else? The fact that it is saying, "These got cited a lot" doesn't seem to be a meaningful reason to post biweekly statistics about it. Often times it is the case, too, that the article being cited really has little bearing on the current state of the thread. For example, the number one article on the list was cited 132 times. I am not going to check, but I would be very surprised if greater than 50% of those 132 articles made direct reference to that article. Speaking of References: I note that though you responded directly to Jill you did not include her article's id in a References: line in your article. The instability of the References: line, which is commonly screwed up somehow, is yet another factor that makes the accuracy and significance of the statistics questionable. In the last posting, of the top eleven articles (don't forget the #10 spot with only a Message-ID:), ten of them were as such: } Subject: Re: Xerox sues Apple!!! The groups varied a little, certainly, as did the people and the ids. It still seems to be quite an odd use of statistics, though. > Takes up lots less bandwidth than other useless crap floating > around the net, Well, that is true too. So what? We're not discussuing the other useless crap. We're discussing this useless crap. > and (oops) it's sort of fun to see which flame wars reached > ridiculous proportions. From what I saw of the Xerox sues Apple!!! thread it really wasn't that colossol a flame war, yet it swamped this list. Most of it didn't seem like a flame war, just a lot of talk. > At the moment, I'm wondering if this very subject will become a > flame war... It has potential, but only for a minor one. Won't make that stats anyway unless it gets moved to comp.all. > I love to have the opportunity to say "told ya so" while being > faux-nice enough to not do so while mentioning that I could. HA. Er, "told ya so" what? Say this thread just died right here, as it might. What would it have been that you would have told me (and Jill and Tom Neff and others that have asked news@elsie to just not do this)? > NOW what's a waste of bandwidth? Lots of things are. Most of USENET is from a very strict point of view. This article right now is wasting my time and the time of everyone who is reading it. We could be out saving the world. Instead we sit around USENET and read and respond, sometimes helping people (that's the non-waste part) to get more work done or just enjoy some R & R, sometimes flaming wildly (sort of fits in the R & R for a lot of people too) and sometimes just not getting anything beneficial accomplished. USENET is a humongous cramp to productivity just as much as it is a humongous boon to that very same productivity. This biweekly list, though, seems to fall more into the former category than the latter. Bandwidth is both a physical thing and a conceptual thing -- it is not only the resources an article consumes in order to get from place to place but how other people consume the article itself -- scanning, reading, replying, skipping, et al. To many people this article that I am now composing, with a lot of bytes written about a pretty petty topic, will be a waste of their personal bandwidth; nearly every article that passes through USENET is a waste of most people's bandwidth somehow. Then again, so are most of the articles when we read a large newspaper. > Possesseth thou NOT an 'n' key? The 'n' key is NOT a cure-all! It still takes time to look at an article and say "Oh wow, how useless. n." > Where WOULD you put such an article, comp.sys.mac? alt.dev.null. > Take a nembutal and check the news in the morning. Thanks for the prescription, Doc. My cheque is in the mail. Dave -- (setq mail '("tale@cs.rpi.edu" "tale@ai.mit.edu" "tale@rpitsmts.bitnet"))