Ah, sweet whiff of youth! History Flow is pretty good, well chosen examples. They reminded me of that other, once-berpromising vaporware tech, Ted Nelson’s Xanadu, or Xanalogical Storage system, that’d have provided always up-to-date credit to any part or span of interlinked docuverse (down to granular character level!). Who wrote what; who quoted whom; where and when did a transclusion originate, and so on. Looked like it’d be happening for a while, well before the WWW [info.cern.ch] made its first apearance and until its once-White Knight Autodesk Inc. pulled the plug on it around 1992 or so. Lingers in Internet limbo since then, aka Bithell. But two years before that, after meeting the inventor eye-to-eye, I wrote this account, complete with a leading declaration of my own word-authorship of it (=71%), and attribution for the rest, 29%, to my oft-quoted subject, the promoter of the concept Ted H. Nelson. So, for the historical declarative-authorship record, I give you this:
That history flow is cool. Where can one fine a tool to help generate such a thing in the open source arena. That would be really cool to try to apply to a code base on high change rate files and such.
Wiki stats are enormously misleading. I could write a bot which formats dates into the WikiApproved™ fashion, and become a top contributor, even though I really didn’t contribute anything of note.
I rewrote a major section of the article on Tae Kwon Do once, and that was… one edit.
I think a better statistic would be to color each word to see where it came from, and look at the authorship of THAT. Should be quite fascinating.
“I expected to find something like an 80-20 rule: 80% of the work being done by 20% of the users, just because that seems to come up a lot. But it’s actually much, much tighter than that: it turns out over 50% of all the edits are done by just .7% of the users … 524 people.”
Under the 80-20 rule, 51.2% of the edits would be done by .8% of the users. More than 50% of the edits getting done by .7% of the users isn’t too far off.
You should check out our project called WikiDashboard (http://wikidashboard.parc.com) that attributes the work of editors in Wikipedia to the articles they’re heavily involved in.
Back up a second: a question is a different beast from a snippet of encyclopedic knowlege. I’d expect an analysis that gets closer to know your customer with respect to questions from a blog subtitled human factors.
Cleartype tuned properly is bearable, but it’s still mangling the characters.
Very often we read blogs because of the author, for their particular style and way of thinking. I know when I read a news article or opinion piece I like to know who’s written it, especially if I’m not familiar with their work.
Wikipedia doesn’t attribute authorship directly to the user (and I don’t think I’ve ever looked at a revisions page on wikipedia.com) probably because that’s the way nearly every other encyclopaedia does it, print or online. I admit that knowing who the author is in an about.com article isn’t always important to me, but I’d like to see authors given more credit in Wikipedia. Anonymous might well stay anonymous for whatever reason, but seeing a name and a profile can give added authority because the author is attaching his (or her) reputation to the content.
Plus, the more I read stuff from wikipedia, the more I ask myself ‘who wrote this crap?’ Not because I disagree with the content, but because the style can be disjointed and sometimes just plain unreadable. Then again, if all the real contributors are outsiders, it would be no benefit to look out for certain authors or mentally block others because those authors may only have contributed a couple of articles. Then again, if I knew the author (or at least their reputation) from outside wikipedia it would add substantially to the utility of the article (or futility, depending on the author).