Okay, here's a bit more. Following up on a question this morning, I took at look at who replies to the *beginnings* of threads, to look for recurring pairings. This only looked at the first four responses, chronologically (according to Yahoo, anyway) to the first message in a thread. A thread beginning was a message that didn't have "Re: " at the start of the subject, which was followed by at least one message with "Re: " and the same subject. Naturally, that creates some errors, for reasons that will be obvious if you look at the subjects. Some non-beginnings were included and some beginnings were missed if their first message had "Re: " at the start. I'll do real threading later. I've done it before... Ilana Halupovich shows up more often than she really should, because her mailer apparently doesn't add "Re: ".
The links between people were trivially weighted by position -- first reply was worth four points, second was three, third was two and fourth was one. A more meaningful score could be developed if we had a good sense of the relationship between message order and connections between the authors. Ground truth would be tedious to develop, but I think it's safe to assume that the first four messages in a thread are typically responses to the initial post. In short, this is a list of people who tend to post close to each other within threads. This covers messages starting 1/1/2001. Now, onto the data! I came up with 2,249 pairings. The top ones and their somewhat meaningless scores: Darryl Shannon-Alberto Monteiro 238 Halupovich Ilana-Julia Thompson 208 John D. Giorgis-John D. Giorgis 170 Halupovich Ilana-John D. Giorgis 157 Halupovich Ilana-Gary L. Nunn 153 John D. Giorgis-Julia Thompson 136 Darryl Shannon-Julia Thompson 94 Alberto Monteiro-Alberto Monteiro 93 Darryl Shannon-John D. Giorgis 89 John D. Giorgis-Alberto Monteiro 89 Marvin Long-Marvin Long 84 Halupovich Ilana-Jeroen van Baardwijk 84 Darryl Shannon-Charlie Bell 83 John D. Giorgis-Jeroen van Baardwijk 77 Darryl Shannon-Jeroen van Baardwijk 76 Gary L. Nunn-Ronn Blankenship 75 Halupovich Ilana-Alberto Monteiro 73 Darryl Shannon-Marvin Long 70 John D. Giorgis-Doug Pensinger 67 Julia Thompson-Julia Thompson 66 Darryl Shannon-Gord Sellar 65 Darryl Shannon-Ronn Blankenship 61 John D. Giorgis-Charlie Bell 59 Charlie Bell-Charlie Bell 58 John D. Giorgis-Joshua Bell 58 John D. Giorgis-Steve Sloan 54 Darryl Shannon-Erik Reuter 54 If anyone wants to draw lines and circles, a links diagram can now be created. I'm not sure that I have anything handy to do that. One thing jumps out -- JDG is the person most likely to respond quickly to the threads he starts. Marvin is next. Feel free to suggest other ways to explore this kind of data. I'm in serendipity mode. The processing time was fast enough that I can do this kind of link analysis for all messages, I believe. However, beyond the start of a thread, it seems less safe to assume that messages are replies to those that preceded them by 1-4 messages, as above. Nick Arnett Phone/fax: 408-904-7198
