Refreshed BIP324

In a discussion on Bitcoin-dev, Pieter Wuille suggested increasing the space for message IDs by treating the first 1 or 2 bytes as a simple variable-length integer. This would result in having two separate variable-length encodings, but instead, they could merge and have a single variable-length command structure that can be used for both long-alphabetic and short-binary commands. Each byte of command encoding will indicate whether another byte follows, with no special meaning attached to the top bit of byte 11. This will provide a uniform space for assigning commands. With this approach, every alphabetic command of L characters becomes L bytes, 102 non-alphabetic 1-byte commands can be assigned, and 15708 non-alphabetic 2-byte commands can be assigned. New commands could also just be introduced as short ones only (even in v1).