Modeling as a Waste of Time

I actually think that modeling is a hugely useful way to understand the world, but sometimes you do it just to waste time.

Tiffany and I have been e-mailing back and forth about bankruptcy (or complaining about casebook editing... yeah, mostly complaining about casebook editing). I noticed something about the e-mail exchange. Here are the sizes of the e-mails, in kilobytes:

4
4
5
6
8
13
18
25

Notice anything? They're getting bigger exponentially! I think. It's been a while since math class.

Now, unfortunately there's a very boring explanation. We're both just hitting reply and not deleting the old portions of the e-mail, so each message carries all the others with it. Still, that should be linear, right?

Well, almost. Cmail adds > before each line for old material, so if I wrote something 5 e-mails ago, it should now be behind ~10 >'s (5 from her and 5 from me, or whatever). So you can see how that would be exponential: each time there is new text, plus the old text, plus a bunch of new >'s (and more >'s are added every time).

I'm not sure that accounts for all of it, though; that's not so many characters to add to an e-mail. Anyway it would be much more interesting if conversations naturally grew exponentially until entering some kind of supernova phase. I mean, if a conversation starts at 11:00 and doubles in magnitude every minute, and the conversation kills you at noon, when were you half dead? 11:59!

One final point of maximum nerdiness: I wrote this post before replying to the latest e-mail so that I wouldn't get another e-mail that might mess up my data. This is how science fails! Not with the big scams, but with the little data nudges.