The pain of models

Here's a piece from Revkin's Dot Earth blog at the NYT with Gavin Schmidt discussing climate models, and how they never really work perfectly but are vital and useful all the same.

This is interesting to me right now because I'm ruminating, as is my general wont during the summer vacation. I use my long hours of manual labor in gardening, animal husbandry, mechanics, construction, and renewable energy fieldwork to regurgitate the semester's highs and lows and think up new ways to teach useful stuff.

And let's face it, for almost all the students I taught models to last semester, in GL4003 Global Change and PS 3003 Sustainable Energy, how many of you were dragged kicking and screaming to the table of college-level, multivariate, dynamic, predictive mathematics?

Nearly all.

It was like pulling teeth! Or more like jumping into a cold pool. Nearly all of you needed one-on-one encouragement before you even cracked open the software.

This is partly because of the unique and own-to-earth nature of the average Unity College student. We are practical people, after all. If my job were primarily to teach practical engine rebuilding or basic construction, I'd get joyful participation our of 95% of my students.

But models are useful too. In fact, without design models or building systems models, there wouldn't be engines or houses to build or rebuild.

It's all in the translation from physical reality to mathematical abstraction. You have to be able to imagine and even visualize the process.