Optimized response

We had a review session in Algorithms yesterday ahead of Monday's midterm. There's no one thing that has me worried, but I am a little concerned about time. The test is only 75 minutes and from the way the prof framed it, we'll have probably 15-20 minutes of "objective" questions followed by some problem solving. Algorithms problems take a while to get right. Producing a correct solution in 10-15 minutes is no small thing.

I was thinking about why I was uncomfortable with this. We work under serious deadline pressure all the time at work. There are a couple things that make it different. The biggest is the immediacy of the deadline. Having 50 hours of work to do in a week is a lot different than having an hour and a half of work to do in 75 minutes. In the former case, you figure you can probably cut out some fat from the rest of your day and if you have to stay late a couple evenings, big deal. In the latter, you either deliver or you don't.

And, that's why businesses generally don't operate that way. It's setting your staff up to fail. At work, you give the "right answer" as quickly as possible ("right answer" here means delivering something of acceptable quality, not perfect). On a test, you give the best you can within the time. While that distinction may be subtle, it is also fundamental. Staffed properly, the first strategy yields consistently good results. The second will at best yield very expensive results (by setting the due dates conservatively) but more often just causes delivery of crap to production.

So, I have to change my mindset a bit. I would never knowingly give my boss a wrong answer. If I wasn't reasonably sure, I'd just say I'd get back to him (and, I'd make sure I did that quickly). But, that response won't fly on the test. I'll just have to come up with the best I can and try not to wince when I write my name at the top of the paper.