As I explained to unrequitedthai, even though OTHER people will read my blog in their own styles, *and* REFUSE to read it if it is not on live-journal, *I* will be reading and working with my blog in its own style! So ;phtpbhpthbpthbpthpthtbptbthy!

I think I worked in that same research group later, as an Undergrad. I worked with Eugene.

Yes, it was the KBS group. It became, during the period I was there, an amazing and diverse team that built software that worked on deadline. I was there Spring-Summer 2001.

Janet was still alive then. Michael Hamman was there and more or less in charge of software engineering, Misha Voloshin was gone, but I knew him.

Essentially, 9/11 killed most of the funding for that project.

So, of those, I worked with (in order of your listing) Eugene Grois (who now works for ALG to handle reasoning and text analysis after Duane Searsmith moved to Riverglass to become my boss), Zhijin Sheng, Ron Carbonari, J.C. LeMentec, and Michael Hamman. J.C. left while I was there (and some sure didn't mind, although when I used CogNet it was extremely good at some things; my conflicts with him were more social than technical). Brent's younger brother Kent Spillman was there while I was there. He handled graphics internals there, and was just a very competent developer. He was the first person I knew to start practicing test-driven development.

I knew Misha Voloshin outside of KBS when I was a freshman. I had heard stories about how intimidating Vadim Bulitko would be during interviewing, and how scattered Mike Bobak was. Guoming Shou was favorably commented on by the amazingly competent Mark Hoemmen, an obsessive-compulsive Catholic numerical analyst who compared long coding spells to hours of chanting liturgical music, and who had to wear gloves to keep him from hurting himself while writing the tightest numerical loops you ever saw.

Virtually everybody who was there while I was there was viciously talented, and for many projects I'd still just reassemble that group. Michael Hamman was a widely read post-modernist mucisian turned Agile development guru (since turned Agile managerial organization guru). I was in intelligent reasoning, handling all non-DCA agents. Also with me was Greg Dhuse, a mild-mannered Bayes-net expert.

In the next cube over was Kent Spillman and the quad-threat Braden Kowitz (programmer, interaction designer, graphic designer, industrial designer) who I really wanted to snag before he went to Google. Braden wrote the graphics frontend rendering and ship layout view, and was known for wearing tropical shirts and leis, playing tropical music, having a beach as the background on his two side-by-side screens, and playing Hawiian music.

In the cube in front of that was the nearly excessively tall Dave King, who wrote VB in the dead of night and who reigned court even just coming off his sophmore year in programming languages and graph layout, and probably was the second most mathematically competent person in the group after Mark. Also in that cube was Karl Shultz, the laid back natural language programmer, who's now on the research staff at Stanford and who would not move back from Berkeley for any reasonable price. Also there was Mike Pilot, who was a Windows programming internals enthusiast and is still around, at Wolfram, although I haven't seen him since those days, and not often even then.

In the furtherest cube away was Michael Collins, who is unquestionably the best release/build/configuration management person I've ever met. Even on source safe, he could give you every working build, ever. You could get the rawest, most bleeding edge, or something that would build clean, no errors, no warnings, or with a known working version of any component. He currently does configuration admin for jet aircraft at Boeing, and we could really use him at RG.

This list was deliberately not comprehensive, ignoring the sysadmins, a grad student or two, the librarians (and the book cutting madness), and anybody else who I just didn't feel like mentioning.

Yeah, I never worked with Ron on anything, but he was a great guy in general. It looks like the university is giving him trouble with only half-time appointments. I think he'd be good to have at RG, in a "guy who gets good code done fast and plus is reasonable to work with" position

In terms of scheduling, we really got on board with Michael H. You can appreciate that software scheduling for a bunch of undergrads is a strange beast, when compared with hardware, but he helped us all whip up a bunch of UML goodness and managed to start doing honest-to-goodness statistical schedule-tracking.

Misha got really burnt out. Essentially, he wrote some Bayes net papers there, and more or less didn't get any credit for them, and everybody knew it. Eventually, there was an understanding: he would just play video games and answer questions, and still be paid.

I think that's as it should be. In experimental particle physics, they will have papers with very large numbers of coauthors, because those were the people that participated in the production of the results: not all equally, not all in comparable ways, and not all together. If it's legitimate in physical sciences, it should be allowed for information science systems of a similar scale.

It's never high standards that bother me, it's aggressiveness. I very much prefer a feeling of "we're only taking high performers, but we'll take every measure to make sure that we are seeing you not under some distorted stress situation, but more like in the course of your day-to-day behaviors, and we'll error on the side of giving you a too-friendly situation, just to see your best instead of your worst"

Or wasn't, in the time you were there. I had an interview with Eugene and Michael Hamman that totally worked that way. They were all like "Do you know C++?" and I said that I could do what I wanted with it, given a reference, but was not an expert. And then they said: "Here is a recent bug on the whiteboard here; what is wrong with this code?" And I said, "I don't know, I'd have to look this up, but this looks wrong. I think it could be fixed this way, but look at this, this might also be legal. Is this legal C++?" And they said, "We don't know. That's a very novel approach." And, I more or less was in from there forward, I think.

I think everybody had done enough last-minute coding bursts that almost worked, that the group as a whole was just sick of things as they stood. People at KBS, by the time I got there, were ready to take the step from coders to developers.