This is my review of draft-ietf-httpbis-binary-message-04 as part of 
ART Last Call review.

Overall this is a concisely written document, and although the specification
makes it clear that its format is distinct from existing framing defined within
HTTP/2 and HTTP/3 (in addition to borrowing elements from QUIC) readers who are
not already familiar with these may be at a considerable disadvantage trying to
understand this document.

Editorial Comments:

Section 3 is quite difficult to grok, and I agree with the suggestions made by
David Schinazi in his Gen-ART review. However as one additional minor suggestion
is that references to section 4 within section 3 appear repetitiously, and as
section 4 doesn't provide specific guidance for each of the invalid message
scenarios but generic text I'd suggest you include "invalid message" as a
definition in section 2 that states all usage of the word is applicable to
section 4.

Section 6 - As a suggestion I would propose the normative text around
pseudo-fields instead be placed within appropriate sub-sections of section 3,
keeping section 6 exclusively informative.

Asides from this there are no further nits or major issues to mention.