There is base32, there is base64, now there is base-emoji!
How does it differ from oktupol/base-emoji?
- Its written in C#, therefore usable in .NET
- It does not support armor yet.
Currently (as of the latest commit to this repo), yes, if there is no armor.
The intended use for this is as a library. HOWEVER you can build it with bflat as a console application. The default implementation of the main method favors windows, therefore uses files for en- and de-code base-emoji ↔ UTF8
The principle is identical to that of base64. In base64, data bits are rearranged from their original 8-tuple bytes into 6-tuples, of which there are 64, and each of these 6-tuples is then represented with one ascii character.
bytes | 104 = h | 105 = i | 33 = ! | ...
DATA |0 1 1 0 1 0.0 0'0 1 1 0.1 0 0 1'0 0.1 0 0 0 0 1| ...
base64 | 26 = a | 6 = G | 36 = k | 33 = h | ...
Therefore, the base64 representation of hi! is aGkh.
In base-emoji, 1024 different symbols are used for representing 10-tuples.
bytes | 104 = h | 105 = i | 33 = ! | ...
DATA |0 1 1 0 1 0 0 0'0 1.1 0 1 0 0 1'0 0 1 0.0 0 0 1'0 0 0 0 0 0.0 ...
base-emoji | 417 = 🍒 | 658 = 🌒 | 64 = 😟 | ...
Since 10 quite obviously doesn't divide evenly into 8, base-emoji-encoded data
contains a few bits more of information at the end than the original data. In
case of above example, the base-emoji encoded representation of the string
hi! has 6 bits of information overhanging. This is important to know
especially once there are is an overhang of 8 bits, because then it would
otherwise be ambiguous whether the last 8 bits are a byte of the original
information or not.
To indicate the length of the overhang, following symbols are appended to the end of the base-emoji encoded string:
| Padding character | 🕛 | 🕐 | 🕑 | 🕒 | 🕓 | 🕔 | 🕕 | 🕖 | 🕗 | 🕘 |
|---|---|---|---|---|---|---|---|---|---|---|
| Bits of overhang | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
Whereas the padding character for 0 bits of overhang is optional, and the characters for 1, 3, 5, 7 and 9 bits can't realistically occur.
In above example, there are six bits of overhang, meaning the emoji
representation receives the padding character 🕕. Hence, the full base-emoji
representation of hi! is 🍒🌒😟🕕.
All that being said, base-emoji is horribly inefficient at encoding data.
In base64, where every 6-tuple of bits is encoded in one ascii character of one byte, the encoded data size is 4/3 times the original data size, i.e. around 33.3% larger.
In base-emoji, we use 1024 symbols to encode 10-tuples, however, these 1024 symbols are Unicode! An exact number can't be given due to unicode characters being of variable size, but a quick test with 1000 random bytes showed a threefold increase.
head -c 1000 /dev/urandom | base64 | wc -c
==> 1354
head -c 1000 /dev/urandom | base32 | wc -c
==> 1622
head -c 1000 /dev/urandom | base-emoji | wc -c
==> about 3175