The difficulty is that in the days when C was being designed, computers were much more irregular than they are today. There were one's complement machines, machines that didn't have power-of-two word sizes, no standardization of character sets, IEEE floating point hadn't been invented yet. The irregular machines weren't fringe stuff, they were the dominant architectures (IBM 360/370, DEC PDP-10, Pr1me, just a big collection of weird stuff). And compiler technology was much less advanced. So C was a messy compromise.
High level system programming languages are about 10 years older than C, which in its early days only cared to target the PDP-11 model used for the first UNIX rewrite.
Authors just chose to ignore what was already out there and do their own thing instead.
Well I guess it is lesson for both: Those who think technology is better because it is so successful and who think technology will be successful because it is so much better than other things.
Iirc, C was also designed by committee, and a lot of industry players got some of their grubby hands on the spec... (I could be wrong) but I belive the utter mess that is short/long/char sizes arises from hardware manufacturers wanting their code to be "trivially portable" across platforms with different machine words.
