Tuesday, May 12, 2009

law of last minute standards

i've decided that this has got to be some sort of obscure yet common natural law, much like rule 34 or murphy's law - i don't know if it exists so for now i'll call it the law of last minute standards. essentially, the closer you come to a deadline, the more likely you'll come up with a solution that fits both the original requirements and the amount of time you have left.

case 1: if you are working on a demo for work or a final project, and some feature is not quite working. in the half hour before it's due, you will always come upon a sudden solution that will still make it look like it's working. maybe just less intelligently, but that's all in the background.

example: in my 6.170 class, as our antichess program was due, the ai decision tree kept crashing the software. so we decided that instead of using a decision tree, just to select a random move each time. no more crashes! (the prof noted, "your ai is about as bad as a random move generator.)

case 2: if you are working on a video for some event and decide that watching and cutting clips is too painful, you'll always be able to come up with some way to still give your audience something to watch, without losing too much sleep.

example: in 2002-2003 winter banquet, my last three highlights videos were long cuts of sparring matches with text overlayed, instead of short highlight cuts. this year, i'm probably going to take long cuts of what would have gone into a docudrama, and just put them in the slideshow as extra multimedia.

case 3: a pset.

example: in my lwtc class, i feel like as long as i can submit something to my group, they'll put my name on the final work. even if it's not completely correct. in all my other psets, the solution at 4 am in the morning is to either make simplified physical assumptions that i know are wrong but go for partial credit, or to punt.

No comments: