It seems that pointers and file I/O are simpler in assembler than in C or C++.
ASM is more direct, C is devious. I don't think reading or writing a file in
ASM would have any problems with ASCII 26 characters, though writing to the
console or printer when there is an ASCII 26 character is tricky (a bug in DOS!)
though there is a way around. I remember Arachne couldn't display the Tiny
Editor help file, which contains a table of all ASCII characters from 0 to 255;
ASCII 26 was the snag. Maybe Arachne inherited that bug from the C (++)
compiler?
I never tried programming the graphics instructions of the newer Pentium chips
in assembler; most of my work in assembler was before that. Now I don't know if
the graphics instructions on the newer AMD chips are the same or different from
the newer Pentiums. But the instructions would be different on the PowerPC,
Compaq (formerly DEC) Alpha, Ultra Sparc, etc.
Ada seems much less tricky than C or C++: strong type checking, and much less of
those sneaky traps for the unwary.
Linux seems to be the most widespread OS in terms of supported CPU
architectures, though NetBSD is not too far behind. Linux now runs on handheld
computers, mainframes (IBM S390), and many architectures in between.
