our quest this semester was to read conference papers to find out what's new and interesting in the world of network protocols, so a partner and I teamed up and decided to explore the effects of bandwidth hogging nodes that set their contention windows shorter than they're supposed to in CSMA/CA ad hoc networks. we have to run a simulation to back up our conclusions, so we're using glomosim (NS-2 is a more popular, but from what I hear more time consuming, choice). you have to know some C but my partner is a CS major so I was like, hey, no problem.
it turns out that glomosim runs best on linux, and my partner revealed to me that she's not all that great with C and doesn't really want to touch this simulation software with a ten foot pole, so I'm like fine, whatever, I'll run it. she's doing the writing. since when am I ever the code one?
sadly, this simulation thing sucks... it's not supporting any conclusion. I have bandwidth-hogging nodes exploiting the hell out of FTP services on a massive network and net throughput for the network goes up, and throughput for my cheating nodes goes wherever it wants to with no obvious pattern, and I don't know C, I just hack at it (coming from PHP, I'm like, "what? declare variables? screw you!").
Our teacher has implied that he'd like us to discuss issues we ran into with simulation, which I hope means that he's not going to be real picky about this. I mean, we're not phd candidates here, we're not submitting these papers to an IEEE conference, we're grad students... tired, cold, hungry, and not nearly as smart as we should be. this is a tough class. there are smart people in it. they make me jealous.
it's a painful bi-weekly reminder that I am not a smart girl, I just play one on the internet. I really hope I can swing a C in here. Then I think I might drop out and take philosophy or something.