Every attachment you’ve ever sent, every HTML email with a plain-text fallback, every inline image embedded in a message body — all of it exists because of MIME, and Chapter 8 is where the book gets into the actual mechanics.
It opens with boundary strings, the delimiter text that separates one part of a multipart message from the next, and walks through the single most common MIME parsing failure: a boundary declared in the header that doesn’t match, character for character, the boundary string used in the body. One trailing space is enough to break a strict parser.
Encoding gets equal attention — Content-Transfer-Encoding and the choice between base64 for arbitrary binary data and quoted-printable for mostly-readable text. The chapter also nails down a precision point the book treats as non-negotiable: the literal RFC 2045 token is 7bit, no hyphen, while the adjectival form “7-bit encoding” is hyphenated — both correct, in different grammatical contexts.
The back half builds up to the book’s “Russian Dolls” model: how a real HTML email with an inline image and a PDF attachment is a nested structure of multipart parts inside multipart parts, each with its own headers.