Seriously, who schedules an exam for 8 a.m. on a Monday morning? Being there an hour early as I always try to be, just in case, meant I had to get up at 5 a.m. (oddly I still slept soundly) and burden my wife with getting all three children ready and to their respective destinations. Then again, it was over before 10 a.m. and I am home again, and not feeling quite as disappointed as after intelligent systems last week.
For the exam was quite doable. As promised, the professor had followed closely the Altklausuren of his colleague, who had been teaching computer networks at UAS for many years. Most problems were quite familiar, except he had given us quite a deal more to solve in the same time (90 minutes). Some questions tested basic knowledge (the protocol stack, which protocol maps IP addresses to LAN addresses, that kind on thing). Most were things like filling simple routing or NAT or switch filter tables, or identifying the proper address range (prefix) for some IP addresses from a routing table. That one I redid three or four times though because it was a good deal more nasty than the equivalent problems from the Altklausuren–just before time was up, I realized that the IP addresses differed not only in the usually relevant third octet (group of 8 bits) but also in the first and second! Firewall configuration I had to guess, and calculating the throughput of a two-hop satellite link completely puzzled me for a while, because I had about a dozen formulas on my crib, but none for throughput. I calculated the load instead and offered that a load factor of 1 equaled maximum throughput but was unobtainable in practice. Very late it dawned on me that end-to-end throughput simply meant the minimum bandwidth of all links in the connection, which was given in the assignment and the same for all of them, so I scribbled that in the margin, but left the load calculation as further demonstration of my knowledge (or the completeness of my crib, more likely).
I had replies to all questions, and although I am not entirely sure of all my answers, I expect a good grade if the professor grades reasonably kindly. It’s hard to say though, because unlike all other professors this one puts the maximum possible points for each problem fantastically high–like 140 points for a single (composite) problem. I suppose he simply wants a reserve of granularity so he doesn’t have to split points, but if maximum points for a one-liner are 40 or so, it gives one the bad feeling that the professor might have expected more, or might more easily give partial points. What’s withholding 5 points out of 40, afterall? But in the end, over all the questions, these point deductions will add up. And if there is no point buffer–no bonus points beyond those required for an A+–as is usually the case with the technical CS professors, there goes your grade.
Oh well, never mind. For some odd reason I have lately resigned myself to lowering my expectations. In the end, you know, it doesn’t really matter. Good grades are always primarily a means to an end, and I already have a job.