>The problem was predicting the running time of tape sorts on main frames (surprise - it's not dependent on the number of records.)

I was lucky enough to start on VT100, even emulated when dialing in via C64 over 1200/75 Bd.

>It was a logarithmic problem and I got it to run on the first shot.

Nice work. Similar experience shoved me into coding
"God is Real, unless..." :-))

>It remains one of my favorite languages.

I made nice money on the side with student assingments, which gave me enough confidence to build stats and time series later on other peoples basic routines. But Pascal having Records and Pointers was later on much easier to work with.