December 7th, 2003


(no subject)

So here I am back from Connecticut. (I didn't win - I didn't get to the finals, again. But I did pretty well overall, actually - I beat all the people I was worried about.)

(Yes, this means that the guy who beat me was someone nobody expected. So it goes :P)

Now here's the tricky bit. See, I got a new PDA before leaving. And I played with it . . . a lot . . .

And you know how you have a great idea for an LJ entry, but then you forget it before you write it down?

That doesn't happen anymore for me. The forgetting part, not the idea part.

So here's the problem: as I write this, I have a list of no less than 23 entries I plan to post. Some of them are already written. Some of them aren't. But 23 is a lot.

So I'll be staggering them out over the next few weeks, occasionally dated from when I wrote them, occasionally not. Of course I will accumulate more as I go.

I just hope I can keep ahead of my stream of consciousness without flooding you all too badly.
  • Current Mood

delays and planning, written dec 7 while waiting on the plane to go home

The plane was supposed to leave at 8:00. They said we would board at 7:30.

At 8:00 we started boarding. They said we'd leave in 10 minutes.

At 8:20 they said we were just waiting for a few last people. Their plane would arrive in 10 minutes.

At 8:45 they said the plane had arrived, and the passenger would be here in 5 minutes at most.

At 9:30 he arrived and we left the gate. We needed to de-ice the plane though, and that would take 15 minutes or so.

Now it's 10:30, and we just finished de-icing. I hope we're going to take off soon.

Why are people so incapable of making good estimates? Do they never pay attention to how long things actually take? This is practically a cliche in compsci, and I have *never* understood why. I've heard dozens of excuses - "you can't plan for the unexpected", "you don't know what could go wrong" . . . to be honest, those are *terrible* excuses.

"Why is my house not built yet?"
"Well, I forgot I needed to sleep. You can't predict things like this!"

But the fact is you *can*. You can add some constant factor for "things that will inevitably screw up". Sure, you can't predict the details, but you don't need to - it's a time estimate, not an itinerary!

When I wrote the routefinder algorithm, I knew I could have it done in a day. Then I added a day, because there was no way my first design could possibly be the optimal solution. Then I added another day for the parts that I'd underestimated the complexity of. Then I added another two days for debugging.

"It'll be done in a week."

It took a full week.

We left at 11:00. Apparently we're only an hour behind our planned arrival time. I'm not sure how this works either.