* Comments in code (ca. 1987, while learning BASIC I was like, but REM doesn't do anything, what's the point?)
* Type systems (ca. 1996, when I first learned Perl)
* Intuitionistic logic (ca. 1999, my final paper in Philosophy of Math was a diatribe against intuitionism)
* Sequent calculus (ca. 2001, it just seemed so backwards; now I can hardly think in natural deduction anymore.)
* Display logic (ca. 2008. It seemed like a wash of funny symbols until I connected it very recently to the hybrid labels in my thesis work --- I'm impressed at Frank's prescience in suggesting Rajeev for my thesis committee)
I am particularly embarrassed I took so long to really understand display logic. It's not something I completely subscribe to, but I'm astonished how early Belnap got very close to some really important issues.