@ed1conf Incidentally, Kirk McKusick used to tell an anecdote in his BSD Kernel Internals class about the original NFS code.
The NFS implementation came from Rick Macklem at the University of Guelph, who wasn't the best at `ed(1)`. Instead of using `z` or `p` or related commands to display a section of code, he'd just hit Enter over and over to print out lines.
On a 24-line terminal, that meant he could only see about 12 lines of context at once. As a result, he broke all his kernel code up into squillions of little 10-line macros.
Here's an example you can see from #TUHS https://www.tuhs.org/cgi-bin/utree.pl?file=4.3BSD-Reno/src/sys/nfs/nfsm_subs.h
When McKusick set Macklem up at Berkeley to get it integrated into 4.3-Reno, they already had those cool BLIT graphical terminals from Bell Labs, so he asked how many terms he wanted on the screen.
"Oh just one will do."
So he got one 100-line terminal or something, and was delighted that he could now see 50 lines of his code at once!