At 16.54 -0700 01-08-03, Bill Barowy scrobe:
>According to Eva's data, both the people who post the most and those who post
>the least are in violation of ziph's law.
Thanks for doing the math, Bill!
But if the list isn't following Zipf's Law, that means that the system has
drifted away from mechanistic behavior in the gross quantitative aspect,
doesn't it? Or does it just mean there's a screw loose in the mechanism?
Now, joking apart, I should not imply that there is one single "mechanism"
behind the mailflow on this list (or any other I know of). Whatever the
social order is that produces this mailstream, it is not a monolith (as
Eric seems to suggest). The list has more than one function and people post
for more than one reason... no, make that: people post for more than one
single SET of reasons (see what I mean?)
>Both groups of people are not
>posting as much as they should. At the top end, this includes:
>
>Mike Cole
>Paul Dillon
>Diane Hodges
>Bill Barowy
>Eric Ramberg
>Judy Diamondstone
You know, I've been there, in the "top six" of the xmca. And one thing I
know is that it takes an awful amount of time. Writing time and time spent
attending to THIS activity as opposed to all other things one might do. So,
well, I don't really think I want to get absorbed into it again - I should
stop responding, I suppose. But it's hard to let go, once I have started.
>At the bottom end it is everyone from Alberto Valeria who are not posting
>enough.
But IF the one-message contributors posted more, they would not BE in the
bottom end. What you are saying is that if this mailinglist system were to
follow Zipf's law, there should be a greater number of transients (one-time
contributors).
>Of course, it could be the complement of people in the middle who are
>the problem. It could be that everyone from Nate Schmolze to Tina Birkeby are
>posting a little too frequently.
Well, too frequently for a perfect Zipf, anyway. But what you report from
the realms of calculation could be rephrased into saying that this
mailstream is produced by a participantry that is broader and more stable
than Zipf would predict, right? Now, is that a good thing or a bad thing?
And of course these quantitative speculations don't BEGIN to address the
way participants relate to each other by electronic mediation (or the way
later postings relate to previous ones by mediation of list participants
;-) - which I take it was Eric's concern. But they may be useful as a
background?
Eva
This archive was generated by hypermail 2b29 : Mon Oct 01 2001 - 01:01:57 PDT