2 Conventions

2.1 Strings

Strings are always encoded using UTF-8 in the Netfarm systems. No buts, no ifs there. The sequence of code points used to form a string must be preserved in encoding and decoding.

(I believe the Common Lisp babel library preserves code points, as the encoding and decoding functions are quite simple, but some guarantees on how code points are treated would be nice.)

2.2 Integers

Integers are encoded in binary data in big endian format, i.e. such that the most significant octet (or other unit) is written first, and the least significant is written last. This format is most common in networking applications, and is the most convenient for users of human languages that write left-to-right to read.

Octet vectors are written in hexadecimal, with the first octet to be written or read appearing at the leftmost position in a vector, and the last at the right.

For example, 10 01 is the vector consisting of the octets 16 and 1, and can be interpreted as the integer 4097.