MyGratuitousARPAmbiguityNotes

Your notes on the ambiguity are in two topics in public.md, with related evidence in a third topic in private.md.

The main notes: ARPAnnouncement (public.md:11759)

This is where the old GratuitousARPAmbiguousTerminology paragraphs went in Phase 2. The core one is at line 11768:

RFC5227 introduced the new name partly because gratuitous ARP had become ambiguous, since people used it for both requests and replies. @ref

The paragraphs around it build on that: - 11764: an announcement and a gratuitous ARP request are identical packets, and only the context differs. - 11766: the term “ARP Announcement” comes from RFC 5227, the conflict-detection RFC. - 11770: the real difference is that an announcement is one step in a defined procedure. - 11772: “Gratuitous ARP request” is the older, more generic term, for a packet sent for any reason. - 11774: every announcement is a gratuitous request, but “announcement” signals RFC 5227.

The pointer: GratuitousARPMessages (public.md:2113)

Known as gratuitous ARP, a term used for both requests and replies, often ambiguously. See ARPAnnouncement for the name RFC5227 gave the request @ref

The evidence: private GratuitousARPMessages (private.md:100001–100002)

These two lines don’t say “ambiguous”, but they show the problem in a real source: - 20261006T200503 defines gratuitous ARP as replies only: “an ARP Response that was not prompted by an ARP Request”. - In a comment, the same author says “Gratuitous ARP can be Opcode 1 or 2”.

Two other passages touch on it without using the word: - ARPProbe (public.md, around line 11846): calls a probe “gratuitous” for being unsolicited, then says it isn’t a gratuitous request in the strict sense. - My last answer (form vs “unsolicited”) adds a second kind of ambiguity: requests vs replies, and also by form vs by purpose. It isn’t in your notes yet.