A few weeks back, Erica Wichers asked if Joel and I would tape an episode of MSDN TV. We, of course, kindly accepted the offer.

Joel was responsible for scheduling the shin-dig. We failed to prepare in the weeks to follow, and so when the date loomed we ran like frightened little squirrels. In other words, we pushed the date back another week. Brad would be quick to note this is a common case of pushing work, delaying the inevitable, allowing the snowball of "stuff to do" to grow, etc.

Well, the date then came again. I suspected it would eventually, but again chose to fiddle with other priorities instead of preparing. Thus we were still totally unprepared. The morning of the taping, we headed out for a latte at Victor's in Redmond (as all such situations call for). We'd tentatively planned on doing a short rendition of our Abbott and Costello act at PDC. But we hadn't prepared anything in the way of demos... What to do?!

Consequently, Joel had recently published a great paper on the performance characteristics of method calls on the CLR. And I had just finished doing a whole lot of playing around with SOS to do with some articles and book content I've been working on. We thought: Let's slam the two together, and talk ad-hoc about the flavors of method calls on the CLR. Sold!

We spent the next hour hacking together sample programs, refreshing our memories, and preparing for the taping. The result is here.

I am truthfully amazed at how well it flows together. I used to watch quite a few MSDN TV episodes back in the day (before joining the firm), and they always seemed so scripted. After seeing how it all works, I bet they all happen in a similar last-minute fashion.