Tuesday, June 29th, 2004

Time	Event
9:32p	This morning: worked. Worked worked worked. It went pretty well. I'm really getting somewhere, I think. Meeting with Igor tomorrow morning at 11, at the ABP in Harvard Square. (I don't know why we're meeting there. My guess is that he doesn't feel like coming to campus tomorrow, and he lives near Harvard?) sarianna and resolvergirl came into town and we had lunch, then much random wandering around, then eventually dinner. It was pretty good, although I had a rather heavy bag with me that had all my math shit in it; had I had a bit more foresight I would have dropped it off at home before meeting them. I now have new shampoo and conditioner, which might make my hair not be so poofy. I hope. The waitress at Fire and Ice (dinner) took me for a girl when she first approached the table; then she corrected herself. I wanted to correct her again, but I was too chicken to, so we wrote a little note on the back of the check. I've been passing a lot lately. Could it have been the arm hair? Oddly enough, I think that might have been a big part of it. Either that or it's just coincidence. (Comment on this)
9:50p	This doesn't surprise me. (clique meme) I am a member of 1 clique of size 8 madcaptenor, sirlightblade, agoddess, liquidsnake, marissav, unitarymatrix, penniavaswen, nectarfizzics Find the largest clique containing: (Enter your livejournal username here). For those who are not quite as in touch with the LJ universe as I am, all eight of these people live on my hall (or lived there before the summer start). The four individuals madcaptenor, unitarymatrix, sirlightblade, agoddess live in four consecutive rooms. I actually searched this manually at some point a few months ago; I wish I had the results saved. The results were qualitatively like this (there was a large clique, and it was entirely Floor Pi people) but I don't remember if they were quantitatively the same. The algorithm used is a bit imperfect, though, albeit for a large savings in running time. An algorithm that would find all cliques including a certain user would be as follows: input: member. output: set consisting of only that member; this is all the cliques of size 1 containing them. input: all cliques of size n for each clique, get friends of each member. take intersection of those sets of friends. add each of those members to the clique of size n to get a new clique of size n+1. eliminate duplicates (how?); if we don't do this as we go along, a clique of size n will occur n! times. output: all cliques of size n+1 repeat until all cliques have been found. Of course, this algorithm generates a huge tree; if I remember correctly, though, the tree prunes itself pretty quickly, in that most cliques of a certain size (say 4) don't extend to any cliques of size 5. But it has the potential to go crazy pretty fast; comments to the entry above claim it's NP-hard, which certainly sounds possible. (2 Comments | Comment on this)

<< Previous Day

2004/06/29 [Calendar]

Next Day >>