Pascal was the first "real" language I learned (BASIC doesn't count) with a compiler that ran on the IBM mainframe at work. From then it was an easy transition to a Pascal compiler from Motorola that produced code for the 68000. Tellingly, the Motorola compiler had a bunch of extensions that made programming a lot easier and mirrored the features Wirth introduced in Modula-2. In some ways, learning C was a step backwards, although in terms of my marketability it was absolutely necessary.
The 68000 was a wonderful CPU, and the last one that I genuinely enjoyed writing assembler for. A sane number of registers, a very orthogonal instruction set, good performance, non-segmented RAM, and high-level. If we had to have a 1980s CISC chip be The One True ISA, I really wish that 68000, not x86, could've been it.
Stack based instruction set designed from scratch for programmers (and implementing high level languages). tagged architecture. Single cycle context switching. It was also a secure architecture directly on hardware, tasks couldn't even generate addresses to munge another tasks data if they wanted to.
Yes, yes, yes! C was a HUGE step backward. And also how I paid the rent in the 90s. ANSI C had the advantage of being portable, though. Even if C, and especially C++, tended to be somewhat deadly.
Turbo Pascal was about 90% of the speed of C with 30 to 50% of the bugs. And yes, I could use pointers and pointer arithmetic in TP if I wanted to bypass array bounds checking, but I had the choice when I needed to do so. Memory management is a lot easier with Pascal functions (and I assume Module/Oberon) than C as well, since the caller allocates the return value. Alternately, you can pass (always non null) values by reference instead of having to use (maybe null) pointers.
