Consonants-only Base 32

This brief document specifies a consonants-only alphabet for base 32 encoding.

Overview

The proposed alphabet aims to prevent the occurrence of words in base 32 encoded strings, reducing the risk of inadvertently generating negative or offensive language.

Additionally, the absence of digits in the alphabet eliminates the risk of parsing the encoded strings as numbers.

Some use cases can be:

• Embed word-safe IDs into a URL slug, e.g. www.example.com/this-is-a-slug-xhPhSqz;
• Represent 64-bit IDs as strings in languages such as Javascript, which hold 53-bit integers in memory;
• Have an alternative string format for UUIDs, which are disliked by some people for its length and dashes.

NOTE: Obviously the idea of ​​using only consonants is not original, as has been mentioned elsewhere. This document only proposes a reasonable alphabet, out of many possible ones, that is ready to use.

Description

The proposed alphabet is a list of sixteen uppercase and lowercase letters, resulting in a case-sensitive list of 32 symbols. It is made up of a series of letters from the Latin script, except the letters A, E, I, O, U, J, V, W, Y, and L. The excluded letters are vowels or those historically evolved from vowels. Additionally, the letter L is excluded due to its lowercase resemblance to the letter I.

This alphabet can be used to both encode integers and binary data. The integers should be encoded by the regular algorithm which uses remainder operation. The binary data should be encoded following the RFC 4648 recipe, except for the alphabet.

We do not recommend using separators such as dashes. We also do not recommend using padding such as equals signs. In both cases, it must be expressed in documentation when necessary for a particular application.

The following table maps each symbol to its respective value.

Value	Symbol	Value	Symbol	Value	Symbol	Value	Symbol
0	B	8	N	16	b	24	n
1	C	9	P	17	c	25	p
2	D	10	Q	18	d	26	q
3	F	11	R	19	f	27	r
4	G	12	S	20	g	28	s
5	H	13	T	21	h	29	t
6	K	14	X	22	k	30	x
7	M	15	Z	23	m	31	z

The following ABNF provides the formal definition of the alphabet.

consonants-only-alphabet   = upper-case-consonants lower-case-consonants
upper-case-consonants      = %s"BCDFGHKMNPQRSTXZ"
lower-case-consonants      = %s"bcdfghkmnpqrstxz"

Integers encoding example

Integers should be encoded using remainder operation. See the following example in Python language.

#!/usr/bin/python3
alphabet = "BCDFGHKMNPQRSTXZbcdfghkmnpqrstxz"
def encode(number, alphabet):
    output = ""
    base = len(alphabet)
    while number:
        output += alphabet[number % base]
        number //= base
    return output[::-1] or alphabet[0]

print(encode(0x000000000000, alphabet));
print(encode(0x000000000ac9, alphabet));
print(encode(0x000000b1cd82, alphabet));
print(encode(0x00057e088d9b, alphabet));
print(encode(0x7fffffffffff, alphabet));

The following table is a list of base 32 encoded integers generated by our code example.

Positive integers	Base 32
0x000000000000	B
0x000000000ac9	DkP
0x000000b1cd82	RFfSD
0x00057e088d9b	hzBcFSr
0x7fffffffffff	Fzzzzzzzzz

Binary data encoding example

Binary data should be encoded with the algorithm described in RFC 4648. See the following example in Python language.

#!/usr/bin/python3
from uuid import UUID;
from base64 import b32encode;
alphabet = "BCDFGHKMNPQRSTXZbcdfghkmnpqrstxz"
def encode(value, alphabet): 
    standard = "ABCDEFGHIJKLMNOPQRSTUVWXYZ234567"
    translator = str.maketrans(standard, alphabet)
    return b32encode(value.bytes).decode("ascii").translate(translator).replace("=", "")

print(encode(UUID("00000000-0000-0000-0000-000000000000"), alphabet))
print(encode(UUID("222e11dd-91d1-400d-994c-dadd71d28015"), alphabet))
print(encode(UUID("c460fffc-c456-4df7-a28e-ebaffcf78e3a"), alphabet))
print(encode(UUID("75a847d7-0c6f-4119-a096-0e69536f925d"), alphabet))
print(encode(UUID("ffffffff-ffff-ffff-ffff-ffffffffffff"), alphabet))

The following table is a list of base 32 encoded UUIDs generated by our code example.

UUIDs version 4	Base 32
00000000-0000-0000-0000-000000000000	BBBBBBBBBBBBBBBBBBBBBBBBBB
222e11dd-91d1-400d-994c-dadd71d28015	GNmCFmScqHBBrKQSrRXmFggBDg
c460fffc-c456-4df7-a28e-ebaffcf78e3a	ncbZzzKGQpKzZNgXtXmzptsXMN
75a847d7-0c6f-4119-a096-0e69536f925d	XkgGZhnSTtBcfNGkCpghKrsdRg
ffffffff-ffff-ffff-ffff-ffffffffffff	zzzzzzzzzzzzzzzzzzzzzzzzzs

License

This document is in public domain, i.e. with no rights reserved, under the terms of CC0 license.