| Internet-Draft | HCTTP | November 2026 |
| Sakai | Expires 20 May 2027 | [Page] |
- Workgroup:
- Network Working Group
- Internet-Draft:
- draft-sakai-hcttp-00
- Published:
- Intended Status:
- Informational
- Expires:
Human Conversational Turn-Taking Protocol (HCTTP)
Abstract
This document specifies the Human Conversational Turn-Taking Protocol (HCTTP), an integrated session-layer communication protocol for multi-node human spoken interaction. HCTTP formalizes connection establishment (Three-Way Handshake), token arbitration, turn transitions, real-time backchannel signaling (ACK/NACK), dynamic rate adaptation, conversational padding (filler frames), collision recovery, and standardized conversational status codes over shared acoustic channels.¶
Status of This Memo
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 20 May 2027.¶
Copyright Notice
Copyright (c) 2026 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document.¶
1. Introduction
Human spoken dialogue over shared verbal and acoustic channels frequently suffers from coordination anomalies, including premature interruption, unhandled transmission silence, buffer starvation, context mismatch, and packet collisions (overlapped speech). While sociological Conversation Analysis (CA) has observed empirical turn-taking mechanics, practical dialogue systems and human participants often lack a formal, enforceable protocol specification.¶
This memo standardizes the Human Conversational Turn-Taking Protocol (HCTTP), bridging empirical Conversation Analysis models with rigorous network transport specifications.¶
2. Conventions and Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174].¶
Additionally, in recognition of habitual human communication failure modes, key words from [RFC6919] are incorporated: "MUST (BUT WE KNOW YOU WON'T)", "SHOULD CONSIDER", and "OUGHT TO".¶
3. Session Management and Handshake
3.1. Three-Way Handshake (Session Establishment)
Prior to transmitting conversational payload, nodes MUST establish a stateful session via a standard Three-Way Handshake:¶
Initiator (Node A) Responder (Node B)
| |
|-------- SYN ("Got a second?").---------->|
| |
|<------- SYN-ACK ("Sure, what's up?") ----|
| |
|-------- ACK ("Great, so...") ----------->|
| |
| [Session Established: Payload Streaming] |
¶
3.2. Graceful Teardown and Reset
To terminate a conversational session gracefully, the initiating node MUST transmit a FIN frame (e.g., "Anyway, I'll let you get back to it"). The session is formally closed upon receipt of a FIN-ACK (e.g., "Sounds good, talk soon").¶
In the event of unexpected physical interrupts (e.g., incoming phone call, urgent alarm), an immediate RST flag (e.g., "Gotta run!") MAY be broadcast to tear down the connection unilaterally.¶
4. Architecture and Frame Formats
4.1. Turn Constructional Unit (TCU)
The primary atomic data packet emitted by a transmitting speaker node is defined as a Turn Constructional Unit (TCU):¶
+-------------------------------------------------------------------+ | Payload (Linguistic Data) | TRP Header (Acoustic / Syntactic) | +-------------------------------------------------------------------+¶
4.2. Feedback Frame (Backchannel Signals)
A listening node MAY transmit asynchronous control frames over an out-of-band backchannel without interrupting the primary TCU stream:¶
+----------------------------------------------------------------+ | Frame Type (ACK / NACK / WAIT) | Bitrate / Pitch Modulator | +----------------------------------------------------------------+¶
- ACK (Standard Acknowledge): e.g., "Yeah", "Uh-huh", "Right". Confirms successful packet decoding and authorizes current transmission bitrate.¶
- DUP_ACK (Duplicate ACK / Fast-Forward): e.g., "Yeah-yeah-yeah", "Right-right". Signals payload pre-processing and requests immediate payload compression towards terminal conclusions.¶
- NACK (Negative Acknowledge): e.g., "Huh?", "What?", head-tilt modulation. Signals decode failure or context misalignment, demanding retransmission with alternate linguistic encoding.¶
- EMOTIONAL_ACK (QoS Amplifier): e.g., "No way!", "Seriously!?". Signals high engagement and amplifies conversational bandwidth.¶
4.3. Filler Frames (TOKEN_HOLD Padding / NOP)
When a transmitting node requires computational latency for lexical retrieval without relinquishing the speaking token, it MAY transmit zero-entropy conversational padding frames (TOKEN_HOLD), commonly realized as "Um...", "Uh...", or "Like...".¶
Nodes MUST NOT (BUT WE KNOW YOU WILL) emit continuous TOKEN_HOLD padding exceeding 3.0 seconds. Exceeding this threshold triggers an automatic 408 (Request Timeout) on peer listening nodes.¶
5. Turn-Taking Algorithm
Upon detection of a TRP delimiter in an incoming TCU, all participating nodes MUST execute arbitration according to the following deterministic precedence rules:¶
[Transmitting Node: TRP Detected]
|
+---> (Rule 1) Explicit Next-Speaker Routing?
| |-- YES: Token transferred to designated node.
| +-- NO : Proceed to Rule 2.
|
+---> (Rule 2) Self-Selection Contention (CSMA/CD)
| |-- First node sensing carrier gains Token.
| +-- No transmission initiated: Proceed to Rule 3.
|
+---> (Rule 3) Speaker Continuation
|-- Transmitting node retains Token for next TCU.
¶
- Rule 1 (Explicit Routing): Target address explicitly designated via vocal addressing (name) or directional beamforming (gaze).¶
- Rule 2 (CSMA/CD Contention): Receiving nodes contend for the acoustic medium on a first-come, first-served basis.¶
- Rule 3 (Speaker Continuation): Transmitting node retains the session and initiates subsequent TCU transmission.¶
Receiving nodes MUST (BUT WE KNOW YOU WON'T) refrain from executing Rule 2 contention prior to explicit TRP delimiter verification.¶
6. Flow Control and Rate Adaptation
Transmitting nodes MUST dynamically adapt payload delivery based on feedback frames:¶
- ACK Received: Maintain bitrate; continue current TCU stream.¶
- DUP_ACK Received: Transmitting node SHOULD CONSIDER compressing payload and advancing directly to the conclusion.¶
- NACK Received: Halt stream; retransmit previous TCU at a lower baud rate with simplified semantic constructs.¶
- NO_ACK Timeout: If no ACK is received within 2.0 seconds, enter Keep-Alive state by pausing transmission and issuing a PING ("Are you still there?"). Resume upon receipt of PONG ("Sorry, yeah, listening").¶
7. Collision Detection and Recovery
When multiple nodes emit sound carriers simultaneously, a channel collision (Overlapped Speech) occurs:¶
Node A: ---[ TCU Payload A ]-------------------------->
Node B: |---[ TCU Payload B ]---> (COLLISION!)
^
Collision Detected
¶
Upon collision detection:¶
8. Conversational Status Codes
HCTTP adopts standard three-digit status codes to represent session feedback and cognitive states:¶
| Code | Name | Description |
|---|---|---|
| 100 | Continue | Listener acknowledges receipt and authorizes continued narration. |
| 200 | OK | Linguistic payload fully decoded and accepted. |
| 400 | Bad Context | Utterance violates conversational context or logical cohesion. |
| 403 | Forbidden Topic | Topic violates social boundaries or taboo constraints. |
| 408 | Request Timeout | Speaker or listener stalled in prolonged silence (> 3.0s). |
| 429 | Too Many Words | Speaker exceeds listener buffer capacity (machine-gun talk). |
| 503 | Service Unavailable | Node cognitive processing exhausted (fatigue or intoxication). |
9. Fault Definitions and Protocol Errors
| Fault Code | Role | Description and System Impact |
|---|---|---|
| UDP Mode | Transmitter | Omits TRP headers and ignores ACK frames. Causes listener buffer overflow. |
| NACK Invalidation | Transmitter | Ignores NACK frames; fails to retransmit, causing session corruption. |
| ACK Storm | Receiver | Emits rapid ACK frames without decoding. Results in semantic divergence. |
| Silent Drop | Receiver | Omits ACK during deep internal processing, inducing unnecessary PINGs. |
| Interrupt Bug | Receiver | Forces Rule 2 before TRP detection, inducing excessive collisions. |
10. IANA Considerations
This document requests that IANA assign a well-known port number for conversational token mediation under the Service Name and Transport Protocol Port Number Registry:¶
Service Name: hcttp Transport Protocol: tcp Port Number: 7110 Description: Human Conversational Turn-Taking Protocol Assignee: IESG <iesg@ietf.org> Contact: Conversation Analysis Working Group¶
Additionally, IANA is requested to establish a registry for HCTTP Conversational Status Codes as defined in Section 7.¶
11. Security Considerations
Nodes MUST implement appropriate authentication and emotional encryption (Politeness Protocol Layer) to prevent unauthorized channel intrusion, malicious payload injection, or Denial-of-Service (DoS) via excessive filibustering or interruptions.¶
Contributors
The following non-human entity contributed significantly to the protocol specification, state machine formalization, and drafting of this document:¶
- Gemini (Google)¶
Acknowledgments
The author would like to thank Gemini (Google) for collaborative protocol design and drafting, as well as the Conversation Analysis research community for empirically reverse-engineering human turn-taking state machines.¶
References
Normative References
- [RFC2119]
- Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, , <https://www.rfc-editor.org/info/rfc2119>.
- [RFC8174]
- Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, , <https://www.rfc-editor.org/info/rfc8174>.
Informative References
- [RFC6919]
- Barnes, R., Kent, S., and E. Rescorla, "Further Key Words for Use in RFCs to Indicate Requirement Levels", RFC 6919, , <https://www.rfc-editor.org/info/rfc6919>.