From Steve Huston's review:

Part 1, sect 2.1: "two's-complement" and "signed" seem to be used
interchangeably but they're not equivalent. Using both terms leaves open
the question of why, and casts doubt about what "signed" means (does it
mean 1's-complement?) Recommend specifying signed integer types as
"signed" and specify in the last sentence of section 2 (before section
2.1) that signed integer values are transmitted using two's-complement
representation, in network byte order.