A condensed version of recent events

[logomancer]

Don't worry, I'm still here. But I've been busy, as usual, and have had no time for LJ. To sum up, this past week was the last week to finish my AVL tree project -- it had been extended a week. Between then and now, my hard drive decided to die on me, and I had to get a new one. I spent last night (and some of early this morning) banging on the AVL tree until it worked, all while the spec changed constantly due to last-minute questions on the CS forum.The result: at 3:30 this morning I have a tree that works on my set of test data, and I called it a day there. Submitted, signed, sealed delivered.

I don't know why McQuain dropped the ball with this spec -- it seems like it was poorly written. Given the fact that he's been teaching for a long time, and taught this course fairly recently, the only conclusion I can draw was this was done intentionally to screw with our minds, another project designed to weed out the ones who can't take it. Come to think of it, it seems that way. The original project called for all of this to be done by March 22nd, giving people two weeks from when it was assigned to do it. It's a week later and most people still haven't managed to finish it. Nitpicking over the spec, impossibly short turnaround...this stinks of a weed-out project to me. Given all the whining and stupid questions on the forum, I can't say that that was truly a bad thing.

From this project, I have learned many things:

1. Deadlines are not to be toyed with.
2. The absolute ass pain of implementing (and subsequently debugging) an AVL tree is not worth the marginal increase in speed. BSTs are much easier to maintain.
3. Many (if not all) of your coworkers will be morons who won't know their ass from a hole in the ground.
4. Even McQuain can fuck up.

There are many other object lessons, but I can't remember them right now. Anyway, the next project has been assigned -- binary file I/O. Not hard, but I will need a day or two. Better get busy.

Edit: Apparently, Capra wrote the project spec. This makes more sense. I stand corrected.

Comments

[nius.livejournal.com]

We had 4 days for our AVL tree.. taking DS2 over the summer was nice, no time for BS in the projects, we just inserted int's into our structures to prove they worked, then went on with the next project.

Of course, my AVL tree was actually broken below 4 branches deep, but oh well. BST is definately easier to maintain, but you still can wind up with one reaaaaally long branch that turns into a linked list, and all your effort is wasted. I've heard that a Red/Black tree is easier than AVL, almost as fast, and fixes that BST problem.. dunno.

Binary I/O is easy.. and kinda fun too.

[robertliguori.livejournal.com]

Our trees need to parse a few books. 17 hard-coded levels would probably handle it, but I'd prefer not to risk it.