[Mtgvenue] Exploration of a "posting" metric - draft-elkins-mtgvenue-participation-metrics

["Fred Baker (fred)" <[e-mail]>]

In your draft and in your talk, you mention the possibility of measuring a person's participation or contribution in terms of email postings. You also mention a number of other possible models, such as whether their name winds up be "acknowledged" or listed as a "contributor" in a draft or RFC.

I took a crack at extracting information about email postings from the mail archive. You can find a very preliminary result at https://www.dropbox.com/sh/2t1d3m96fi4brbf/AACqZ8-fQ08bbtFwJA9ZQczQa?dl=0

The point of that was to identify low-hanging fruit. We can and do measure a person's registration for or attendance at face-to-face meetings. We can also determine whether their email address shows up in an internet draft (I find the email address easier to think about than their name, as names can have several formats, wrap around line breaks, and so on). In this prototype, I rsync'd the mailman archive, and for each list and each month's box capture, parsed the file to find "From:" lines from emails.

The output is in the CSV file in that dropbox link. There's nothing magic in there - we could sort things in a different way or etc easily enough. What I tracked in this version was a sender's email address, the lists they posted to, for each such list the total number of messages they sent, and then per month how many messages they sent in that month.

My thought in that is that it's pretty easy to associate an email address with a mailing list, and determine an individual's transmission rate. If one knew that a given person used a given email address, one could determine a person's transmission rate on a monthly basis. What isn't readily found is anything about topic, whether the contribution is positive or negative, useful or otherwise, etc. Semantic analysis, and perhaps the distinction between "participation" and "contribution", is not, from my perspective, "low-hanging fruit".

I'm wondering here what you want in terms of a participation metric.