[From Martin Taylor 2009.04.18.11.56]
[From Dag Forssell (2009.04.18 08:40 PDT)]
[Martin Taylor 2009.04.18.01.10]
<snip>
Incidentally, in 1992 and 1993 and maybe later, there are no From: lines in the headers (except for messages from Dag Forssell). I added them by hand for January 1992, but it took 3 hours, and I'm not inclined to spend that kind of time to do all the other months that way. I don't suppose that somewhere you might have a version prior to July 1994 that retains the From: lines? The Kennaway versions, which do have the From: lines, start in July 1994.
Martin, what you have now is all there is.
After reading this, I began rooting around in my pile of ancient backups to CD, and eventually found a nearly complete set of Eudora Mailboxes for 1992-96, which I have zipped and uploaded to my ftp server (much too big for e-mail). They start with 9203 and end with 9611, missing 9204, 9211, 9302, 9604. You (and anyone else on this list) can pick them up at ftp://ftp.mmtaylor.net/CSG login: mmt_ftp pw: Anonymous. They have full headers.
I do not recall in detail how I edited the posts when I assembled them, but do have a vivid memory of removing redundant headers from digests as I went along (perhaps later), long headers saying that this came from the Control Systems Group etc. etc.. Why are you adding back From: headers? From CSGnet or From an individual address? If you can figure out how to add them back, is not that prima facie evidence that they are redundant and unnecessary?
From an individual address. I haven't figured out how to add them back other than by looking at the individual messages. For example, Avery Andrews's posts are identified by "Avery Andrews" at the end of the message, Chuck Tucker's by something a bit more complicated than <<<<<<CHARLES TUCKER>>>>> at the head. Some must be identified tentatively by reading the content and relating "I said" to other material. And yet others I just have to "identify" as "unknown".
I do not propose to worry about it at this point.
Are you undertaking to read CSGnet from the beginning? That is a lot of reading, but can certainly be useful! If so, are you looking for something in particular?
Yes, it might be useful. If I were to try that, I would be doing it in order to collate recurring discussions of the same topic years apart. Even in my small sampling of this early material, I find so much recycling of arguments that it is very discouraging to prospects of future progress in PCT theory by way of CSGnet discussion. I've even mused recently about responding to some messages simply by quoting from 10 and 15-year-old archives!
The reason I got into this is that I would very much like PCT to become a powerful influence in the thinking of both academic psychologists and human factors practitioners. I observe that over the years, competent people join CSGnet, offer ideas, engage in interesting discussions, and then fall silent. Whether they leave CSGnet or not is unknown to me, but clearly they have been discouraged from further participation. This is the opposite of what must happen if PCT is ever to become more than the private preserve of a small number of enthusiasts.
One way I thought of seeking possible reasons for this unfortunate pattern was to produce traces over time of the contribution density of people who have made more than a few (say ten) posts in any one year, and to look at the rise and fall of the proportions of all posts over time made by individual contributors. A sort of Gini index would serve as a kind of indicator of the health of the group: too many posts concentrated in too few names would suggest an unhealthy me-and-thee-against-the-world atmosphere in the list, whereas too flat a distribution would suggest a lack of originality and novelty. From experience in other similar situations, one might expect the distribution to follow Zipf's Law, but with what exponent?
Anyway, to do this, I have been trying to extract the posters' names month by month, and put them into a spreadsheet, but now I think a database approach would be more useful, since entities could then incorporate thread titles and posting dates. This probably would be most easily done with Ruby on Rails, but for that to work requires that the incorporate the texts of individual messages, but I think that would be opening up a Pandora's Box of work, such as developing automatic cross-linking and the like!
For me, the bottom line is that PCT is too important to be kept in a ghetto, and in principle, CSGnet should be the medium by which it primarily expands. The history of CSGnet suggests that there is something about it that prohibits this kind of expansion, and a look into that history may help address the problem. I don't think the problem lies in the unfamiliarity of PCT itself, though I have seen quite a bit of misunderstanding of PCT among the groups locally who are using it (but not participating in CSGnet). I personally have in the past suggested to several people that they should join CSGnet, but more recently I have been suggesting that instead they peruse PCT-related web sites and avoid CSGnet. I don't like doing that, but the historical pattern seems to suggest that this is a more prudent way of getting people to develop a PCT understanding of the world.
Martin
···
From: line be kept in the headings. Ideally, such a database would
