The question I had is: While coffee's roasting, do the stages of roasting, indicated by the casually tossed around term "cracking," correspond to detectable thermal transitions? Temperature profile to the rescue.
I popped 50 grams of green beans to the hopper, dangled my thermocouple into the headspace and let the baby rip for 15 minutes (a time based on some futzing around). Here's the profile:

Really boring temperature profile over roasting period.

The beans are shinier in the image because of my inability to use flash properly
The beans were slightly shiny, and dark roasted and are resting before I grind. I've never been able to hear that definitive cracking - all I hear is the dog whimpering, the cat attacking my feet or the kid running in the hall, so I've relied on time to provide an endpoint for roasting. According to my time profile, there appears to be no thermal events to dictate the perfect roasting endpoint. Just have to stick with plain old trial and error. A fun set of observations though.
... Honey, the popcorn is starting to taste funny.