



Network Working Group                                             Rayner
Internet-Draft                                               Independent
Intended status: Informational                            20 August 2025
Expires: 21 February 2026


     Proquints: Readable, Spellable, and Pronounceable Identifiers
                        draft-rayner-proquint-05

Abstract

   This document specifies "proquints" (PRO-nounceable QUINT-uplets), a
   human-friendly encoding that maps binary data to pronounceable
   identifiers using fixed consonant-vowel patterns.  The concept was
   originally described by Daniel Shawcross Wilkerson in 2009.  This
   document formalizes the format for archival and reference.

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 21 February 2026.

Copyright Notice

   Copyright (c) 2025 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.







Rayner                  Expires 21 February 2026                [Page 1]

Internet-Draft                  Proquint                     August 2025


Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
   2.  Requirements Language . . . . . . . . . . . . . . . . . . . .   3
   3.  Format  . . . . . . . . . . . . . . . . . . . . . . . . . . .   3
   4.  Encoding  . . . . . . . . . . . . . . . . . . . . . . . . . .   3
   5.  Decoding  . . . . . . . . . . . . . . . . . . . . . . . . . .   4
   6.  Encoding and Decoding Specification . . . . . . . . . . . . .   4
     6.1.  Letter Tables and Indices . . . . . . . . . . . . . . . .   4
     6.2.  Bit Layout  . . . . . . . . . . . . . . . . . . . . . . .   5
     6.3.  Encoding Algorithm (Pseudocode) . . . . . . . . . . . . .   5
     6.4.  Decoding Algorithm (Pseudocode) . . . . . . . . . . . . .   6
     6.5.  Normalization . . . . . . . . . . . . . . . . . . . . . .   8
     6.6.  Test Vectors  . . . . . . . . . . . . . . . . . . . . . .   8
     6.7.  Error Handling  . . . . . . . . . . . . . . . . . . . . .  10
     6.8.  Backward Compatibility  . . . . . . . . . . . . . . . . .  10
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .  10
   8.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .  10
   9.  Acknowledgments . . . . . . . . . . . . . . . . . . . . . . .  10
   10. References  . . . . . . . . . . . . . . . . . . . . . . . . .  10
     10.1.  Normative References . . . . . . . . . . . . . . . . . .  10
     10.2.  Informative References . . . . . . . . . . . . . . . . .  11
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . .  11

1.  Introduction

   Proquints encode binary data as alternating consonant-vowel letters
   grouped into five-letter syllables, yielding identifiers that are
   readable, spellable, and pronounceable.  The idea and specific letter
   tables were first described by Daniel Shawcross Wilkerson in 2009
   ([WILKERSON2009]).  This document does not claim originality for the
   concept; it reformulates and formalizes the description for archival
   purposes.

   While multiple schemes exist for encoding network addresses and other
   binary data, Proquints aim to provide a unique blend of human-
   reabability, accessibility, and long-term usability.  They reduce
   transcription errors, are friendlier for non-technical users, and
   offer mnemonic qualities that can help in educational or operational
   contexts.  Although they may not replace all existing
   representations, Proquints can serve as a complementary format that
   improves clarity in documentation, user interfaces, and spoken
   communication, particularly where accuracy and inclusivity matter.








Rayner                  Expires 21 February 2026                [Page 2]

Internet-Draft                  Proquint                     August 2025


2.  Requirements Language

   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] when, and only when, they appear in all
   capitals, as shown here.

3.  Format

   A proquint encodes data in 16-bit blocks.  Each block maps to a five-
   letter syllable of the form CVCVC (Consonant-Vowel-Consonant-Vowel-
   Consonant).

   The mapping tables are fixed:

   Consonants (indices 0..15):

   *  b d f g h j k l m n p r s t v z

   Vowels (indices 0..3):

   *  a i o u

4.  Encoding

   *  Encoders MUST process the input as an ordered sequence of 16-bit
      words formed from the octet string in network byte order (big-
      endian).

   *  If the input contains an odd number of octets, encoders MUST
      append a single zero octet (0x00) to complete the final 16-bit
      word and MUST signal this padding by appending a single trailing
      hyphen (U+002D HYPHEN-MINUS) to the end of the proquint string.
      Encoders MUST NOT append a trailing hyphen when the input length
      is even.

   *  Hyphens between syllables remain optional for readability;
      decoders MUST ignore interior hyphens.  Only a single trailing
      hyphen has special meaning as a padding signal; multiple trailing
      hyphens are invalid.

   *  For each 16-bit word, map bits 15-12 to the first consonant, bits
      11-10 to the first vowel, bits 9-6 to the second consonant, bits
      5-4 to the second vowel, and bits 3-0 to the final consonant.

   *  Concatenate syllables.  Hyphens MAY be inserted between syllables
      for readability; decoders MUST ignore interior hyphens.



Rayner                  Expires 21 February 2026                [Page 3]

Internet-Draft                  Proquint                     August 2025


5.  Decoding

   *  Decoders MUST reverse the mapping in Section 4.

   *  Decoders MUST accept upper- or lower-case input and MUST ignore
      interior hyphens.  If the input ends in a single trailing hyphen,
      the decoder MUST: (1) decode the syllables to octets; (2) verify
      that the final octet is 0x00; and (3) remove that final octet.  If
      a trailing hyphen is present and the final octet is not 0x00, the
      decoder MUST treat the input as invalid.

   *  If no trailing hyphen is present, the decoder MUST NOT remove any
      trailing octet, even if it is 0x00.

   *  Inputs with multiple trailing hyphens, a trailing hyphen without
      any syllables, or a length not divisible by five letters (after
      removing hyphens) MUST be rejected.

6.  Encoding and Decoding Specification

6.1.  Letter Tables and Indices

   Proquint encodes each 16-bit word as five letters in the pattern
   CVCVC (Consonant–Vowel–Consonant–Vowel–Consonant).  The mapping
   tables and indices are fixed and normative.

   Consonant table (index 0..15):

   Index  Hex  Bits  Consonant
   -----  ---  ----  ---------
     0     0   0000     b
     1     1   0001     d
     2     2   0010     f
     3     3   0011     g
     4     4   0100     h
     5     5   0101     j
     6     6   0110     k
     7     7   0111     l
     8     8   1000     m
     9     9   1001     n
    10     A   1010     p
    11     B   1011     r
    12     C   1100     s
    13     D   1101     t
    14     E   1110     v
    15     F   1111     z

   Vowel table (index 0..3):



Rayner                  Expires 21 February 2026                [Page 4]

Internet-Draft                  Proquint                     August 2025


   Index  Bits  Vowel
   -----  ----  -----
     0    00      a
     1    01      i
     2    10      o
     3    11      u

6.2.  Bit Layout

   Each 16-bit input value (bits 15..0, most significant bit first) MUST
   be mapped to letters in this order:

   bits 15..12 -> first consonant (C1)
   bits 11..10 -> first vowel     (V1)
   bits  9.. 6 -> second consonant(C2)
   bits  5.. 4 -> second vowel    (V2)
   bits  3.. 0 -> third consonant (C3)

   Encoders MUST process input as an ordered sequence of 16-bit words
   formed from the input octet string in network byte order (big-
   endian): octet[i] contributes bits 15..8 and octet[i+1] contributes
   bits 7..0 of the word.  If the input contains an odd number of
   octets, encoders MAY pad a single zero octet to complete the final
   16-bit word; applications using padding MUST specify how the original
   length is recovered.

   Encoders MAY insert ASCII hyphens (0x2D) between syllables for
   readability.  Decoders MUST ignore interior hyphens, but not trailing
   hyphens which indicate padding.

6.3.  Encoding Algorithm (Pseudocode)




















Rayner                  Expires 21 February 2026                [Page 5]

Internet-Draft                  Proquint                     August 2025


  Input: bytes[]  // octet string
  Output: string  // proquint

  consonants = "bdfghjklmnprstvz"
  vowels     = "aiou"

  function encode(bytes):
    out = ""
    i = 0
    pad = false

    while i < len(bytes):
      hi = bytes[i]; i += 1
      if i < len(bytes):
        lo = bytes[i]; i += 1
      else:
        lo = 0x00
        pad = true

      w  = (hi << 8) | lo
      c1 = consonants[(w >> 12) & 0xF]
      v1 = vowels    [(w >> 10) & 0x3]
      c2 = consonants[(w >>  6) & 0xF]
      v2 = vowels    [(w >>  4) & 0x3]
      c3 = consonants[(w      ) & 0xF]
      out += c1 + v1 + c2 + v2 + c3
      // optional: insert interior '-' between syllables for readability

    if pad and len(out) > 0:
      out += '-'   // trailing hyphen signals padding was added

    return out

6.4.  Decoding Algorithm (Pseudocode)

   Input: string pq  // CVCVC syllables, hyphens optional
                     // (except final hyphen is meaningful)
   Output: bytes[]   // octet string

   consonants = "bdfghjklmnprstvz"
   vowels     = "aiou"

   function indexOf(ch, table):
     pos = table.find(ch)
     if pos < 0: error("invalid character")
     return pos

   function decode(pq):



Rayner                  Expires 21 February 2026                [Page 6]

Internet-Draft                  Proquint                     August 2025


     pad = false
     if len(pq) > 0 and pq[-1] == '-':
       pad = true
       if len(pq) >= 2 and pq[-2] == '-':
         error("multiple trailing hyphens")
       pq = pq[0:-1]

     if len(pq) > 0 and pq[0] == '-':
       error("leading hyphen not allowed")
     if "--" in pq:
       error("consecutive interior hyphens not allowed")

     s = toLowercase(pq)
     parts = (s == "" ? [] : split(s, '-'))

     if len(parts) == 0:
       error("empty input not allowed")
     for part in parts:
       if len(part) != 5:
         error("syllable length must be 5")
       c1,v1,c2,v2,c3 = part[0],part[1],part[2],part[3],part[4]
       if indexOf(c1, consonants) < 0
           or indexOf(c2, consonants) < 0
           or indexOf(c3, consonants) < 0:
         error("invalid consonant")
       if indexOf(v1, vowels) < 0 or indexOf(v2, vowels) < 0:
         error("invalid vowel")

     out = []
     for part in parts:
       c1,v1,c2,v2,c3 = part[0],part[1],part[2],part[3],part[4]
       w = (indexOf(c1,consonants) << 12) |
           (indexOf(v1,vowels)     << 10) |
           (indexOf(c2,consonants) <<  6) |
           (indexOf(v2,vowels)     <<  4) |
           (indexOf(c3,consonants)      )
       out.append((w >> 8) & 0xFF)
       out.append( w       & 0xFF)

     if pad:
       if len(out) == 0 or out[-1] != 0x00:
         error("trailing hyphen requires a final 0x00 padding byte")
       out.pop()  // remove the padding byte

     return out






Rayner                  Expires 21 February 2026                [Page 7]

Internet-Draft                  Proquint                     August 2025


   Decoders MUST accept input in either case (upper/lower) and MUST
   reject any character not in the defined consonant/vowel sets (after
   stripping hyphens).  If applications use padding on encode, they MUST
   specify how to remove any trailing zero octet introduced solely for
   padding.

6.5.  Normalization

   Encoders SHOULD produce lowercase output.  Encoders MUST append a
   single trailing hyphen only when signaling padding (odd input
   length).  Decoders MUST treat input as case-insensitive, MUST ignore
   interior hyphens, and MUST apply the trailing-hyphen padding rule
   defined in this document.

   Encoders and decoders MUST use the tables and ordering defined in
   Section 6.1 and Section 6.2.  Substituting letters or re-ordering
   bits is not Proquint and will not interoperate.

6.6.  Test Vectors

   The following vectors are derived directly from this specification
   and can be used to verify independent implementations.





























Rayner                  Expires 21 February 2026                [Page 8]

Internet-Draft                  Proquint                     August 2025


# Single-word (16-bit) values:
0x0000 -> babab
0xFFFF -> zuzuz
0x1234 -> damuh
0xF00D -> zabat
0xBEEF -> ruroz

# Two words (32-bit), big-endian byte order:
bytes:  0x12 0x34 0xF0 0x0D
words:  0x1234, 0xF00D
pq:     damuh-zabat      (with hyphen)  or  damuhzabat (without)

# Raw ASCII example ("F3r41OutL4w"),
# UTF-8 bytes, zero-padded to even length:
ASCII:  46 33 72 34 31 4F 75 74 4C 34 77
Length: 11 bytes
Pad:                                      00
Words:  0x4633 0x7234 0x314F 0x7574 0x4C34 0x7700
PQ:     himug-lamuh-gajaz-lijuh-hubuh-lisab- (interior hyphens optional)

# Padding examples
# Even-length input (no padding, no trailing hyphen):
bytes:  01 02 03 00
words:  0x0102, 0x0300
pq:     bahaf-basab           (or "bahafbasab" without interior hyphen)
out:    01 02 03 00

# Odd-length input with padding signaled by trailing hyphen:
bytes:  01 02 03
encoder pads:                -> add 00 to form final word 0x0300
pq:     bahaf-basab-          (trailing hyphen REQUIRED)
decoder: decodes to 01 02 03 00, verifies last octet 00, then removes it
out:    01 02 03

# Invalid (trailing hyphen but last octet != 00):
pq:     bahaf-basad-
-> decode last word to ... 01 (not 00) => ERROR

# Invalid (multiple trailing hyphens):
pq:     bahaf-basab--         => ERROR

   Implementations MUST reproduce these outputs exactly.









Rayner                  Expires 21 February 2026                [Page 9]

Internet-Draft                  Proquint                     August 2025


6.7.  Error Handling

   Decoders MUST fail input that: (1) contains characters outside the
   defined tables (after interior hyphen removal); (2) has length not
   divisible by 5 letters; or (3) violates the CVCVC pattern.  Error
   signaling is application-specific but MUST reject invalid input
   rather than attempt to guess.

   A trailing hyphen MUST only be used to signal removal of a single
   trailing 0x00 octet; any other usage is invalid.

6.8.  Backward Compatibility

   Implementations that predate this specification’s padding
   specification may ignore a trailing hyphen and therefore retain the
   trailing 0x00 octet.  To interoperate with such decoders, producers
   SHOULD avoid relying on padding removal when communicating with
   unknown peers.

7.  Security Considerations

   Proquint is a presentation encoding.  It provides no confidentiality,
   integrity, or authentication services.  It does not add or remove
   entropy, and it MUST NOT be used as a cryptographic transform.

   Use of a trailing hyphen reveals the parity of the original octet
   length.  This leaks at most one bit of information (even vs. odd
   length), which is unlikely to be security-relevant for typical uses
   of Proquints.

8.  IANA Considerations

   This document has no IANA actions.

9.  Acknowledgments

   The author thanks Daniel Shawcross Wilkerson for originating the
   proquint concept and publishing the initial specification in 2009
   ([WILKERSON2009]).

10.  References

10.1.  Normative References

   [BCP14]    Best Current Practice 14,
              <https://www.rfc-editor.org/info/bcp14>.
              At the time of writing, this BCP comprises the following:




Rayner                  Expires 21 February 2026               [Page 10]

Internet-Draft                  Proquint                     August 2025


              Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

              Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

10.2.  Informative References

   [WILKERSON2009]
              Wilkerson, D.S., "Proquints: Identifiers that are
              Readable, Spellable, and Pronounceable", arXiv 0901.4016,
              January 2009, <https://arxiv.org/html/0901.4016>.

Author's Address

   Thomas Rayner
   Independent
   Email: thmsrynr@outlook.com






























Rayner                  Expires 21 February 2026               [Page 11]
