Why is this section necessary if we are attempting not to parse messages? If we're not parsing messages, then we can't know whether the message has "CR LF" or "LF" line-endings.

As demonstrated elsewhere in this specification, why not simply require the size of any message/header in octets, and let that be the end of it? After all, from a server's perspective the contents (including line endings) doesn't matter.

Entire messages are not parsed for content, but given the definition of the Internet Message Format in RFC-822/2822/5322 changing CRLF to LF (consistently) is not an intensive task. Also, this will prevent confusion re: what format messages should be in when transmitted via BikINI. This section stays. - dave

Issue done.