Bibi-binary | Hacker News Reader

Bibi-binary

en.wikipedia.org

128 pointspndy47 comments

It looks very similar to the "Logic Alphabet".

http://www.logic-alphabet.net/images/flipstick_2347_2.jpg

Both are systems of symbols meant to encode 4 bits of information. In both cases, pointy bits in the corners of the symbols represent 1 or true, while round bits are 0 or false.

A difference is that in the case of bibi-binary the intent is to encode numbers, while the intent of the logic alphabet is to encode truth tables of binary logic operators. But fundamentally they encode the same information.

The other difference is that the inventor of bibi-binary was a comedian who obviously did it as a tongue-in-cheek novelty, while the logic alphabet's creator appears to have been quite convinced of the importance of his invention...

That's actually a really practical set of operators.

Can you elaborate? TIA

The shape of the operators visually shows how the function works. It also has some nice visual properties as

A and B -> A d B

not A and B -> flip horizontally: A b B

A and not B -> flip vertically: A q B

not A and not B -> rotate: A p B

etc...

> A difference is that in the case of bibi-binary the intent is to encode numbers

bibi-binary is intended to encode raw binary data, not specifically numbers (at least not beyond 15). We already have a script and words which work mostly fine for numbers after all.

> The other difference is that the inventor of bibi-binary was a comedian who obviously did it as a tongue-in-cheek novelty

He was also a mathematician — his studies stopped because of world war II forced labour (from which he escaped) and he did patent bibi-binary, it was not just a joke.

And calling him a comedian is a bit reductive, comedianship was his outlet for a passion for wordplay, something he first tried as a songwriter (and poet), but he could find no singer, so ultimately ended up doing it himself.

More explanations here: https://en.wikipedia.org/wiki/Logic_alphabet

In the same order of idea, Proquints are identifiers that are readable and pronounceable. In "Pronounceable Identifiers", Daniel S. Wilkerson proposes encoding a 16-bit string as a pronounceable quintuplets of alternating consonants and vowels as follows. Four-bits for consonants, and two-bits for vowels:

127.0.0.1 lusab-babad

63.84.220.193 gutih-tugad

63.118.7.35 gutuk-bisog

140.98.193.141 mudof-sakat

64.255.6.200 haguz-biram

128.30.52.45 mabiv-gibot

An ascii string: "Many hands make light work. "

hujad kunun fadim kajiv kidog fadit kajir kihab kudin kitim libab litiz lanir funir

https://wiki.xxiivv.com/site/proquints.html

Too many similarly pronounced consonants. A variant that uses only two or three bits for the consonants would be interesting. Maybe some 5-bit subset of the Japanese Gojūon.

That first example's b-b is an immediate blatant problem, especially since h is a valid terminal. With unfamiliar words, there will also be problems with j/g/k. b/v/f, t/d, etc.

The one thing this does better than bibi is that it avoids the "E" which confuses English speakers.

It's a fun idea, but I'm confused as to why the inventor decided that decoding would rely on reading the bits in order of top-to-bottom, left-to-right. To make it intuitive to readers of French, you'd expect the order to be left-to-right, top to bottom. To make it intuitive to mathematicians, you might expect it to start in the upper right and proceed counterclockwise, as per the numbered quadrants of a Euclidean plane. The 16 symbols themselves will remain the same regardless of the order in which you visit the bits, but the sequential position of each symbol will change.

I agree. I also think that the alphabet leans too far in the direction of making the symbols distinct, and not enough in the direction of making them transparently interpret-able as binary - which is the unique aspect of a special binary notation like this and should be emphasized.

I sketched some alternatives here:

https://i.imgur.com/7MZs4ji.png

The bibi-symbols are quite bad for those of us with dyslexia, because mirror variants are so common. Yours are actually worse for that; every rotation of every symbol is present! Serifs could help.

Yeah, that makes sense. Given your account, it sounds like the 2D paradigm of the bibi-symbols (which I'm following) is fundamentally bad for interpretability with dyslexia, because it results in a lot of symmetries.

Even when considering the population of people without dyslexia, I can't think of examples of vernacular human alphabets that have many pairs of distinct symbols that are identical under some symmetry. So maybe symmetry is an alphabet design antipattern.

I was confused that some of the combinations of single-bit shapes don’t match the shapes of the corresponding two-bit combinations. For example, the shapes for 0001 (╮) and 0010 (╯) combined result in the shape for 1100 (⊃), not in the shape for 0011 (⊂).

Of course, if that shape-combining logic was followed consistently, the circle would have to be 1111, not 0000.

The bibi binary notation encodes shift of bits very interesting ways. Take 1, 2, or 3 (only some of the last two bits set) and shift by two bits and you get a left-right mirror image; take 4, 5, shift by one and you get upside-down mirror. I think the notation is awesome.

For the non-French people here: Bobby Lapointe is also a quite famous singer who loved plays on words, quite difficult to translate!

I wasn't sure if it was the same Boby Lapointe.

But reading the "pronunciation" line in the table made me smile and confirmed that it was.

From now on, I'll be singing the hexadecimal pointers when debugging C++ code.

HIHA !

Yes, and another reading on the pun not explained in the page, on top of 'bibi' meaning 'me', is that 'bibine' (present in the prononciation of 'bibinaire') is French slang for alcohol.

It reminds me superficially of marching squares https://en.m.wikipedia.org/wiki/Marching_squares

Non-mobile link: https://en.wikipedia.org/wiki/Marching_squares

This part in the introduction says it doesn't have a citation, does anyone have more information?

I'm very curious to learn more:

   "It found some use in a variety of unforeseen applications: stochastic poetry, stochastic art, colour classification, aleatory music, architectural symbolism, etc.[citation needed]"

This is slightly similar to a ternary encoding recently released by Devine Lu Linvega[1]. Very interesting.

[1]: http://wiki.xxiivv.com/site/heptavintimal

That section on negative numbers seems incomplete. It doesn't indicate how you write negative numbers instead talks about how normal binary computers use twos complement.

Edit: I think I was assuming you'd only use 4 bit or 8 bit values with this notation.

As far as I can see the section on negative numbers was completely made up, or at least it is not sourced in the system's patent. The encoding was mostly defined as a way to transmit (and write down) binary data in nibbles (groups of 4 bits), it does not really concern itself with what that data encodes.

The one thing it does support which the wiki does not note, is a run-length encoding: if a syllable ends in R, then the segment until the next R is interpreted as a positive integer and a repetition count so HORBAR is 05 (5 repetitions of 0), and DIRHAHOR is 1516 (16 repetitions of a fully set nibble).

And of course I forgot about HN's shitty markup and didn't notice before the edition timer ran out.

HORBAR is repeat(0x0, count=5)

DIRHAHOR is repeat(0xF, count=16)

Also in script form this is written as a strange parenthetical. Strange because the repeated nibble gets an open paren "stuck" to it, like a tail. So repeated nibbles don't look much like their base form.

Indeed the section on negative numbers is bullshit. One's or two's complement representation only works with fixed-length numbers. Arbitrary-length numbers would need an infinite number of leading 1 digits.

I'm annoyed by the use of one's complement instead of the typical two's complement. Why use a system with two representations of zero?

Obligatory to also mention the historical Cistercian digits, which used an in-digit encoding to create a base 10000 number system:

https://commons.wikimedia.org/wiki/File:Cistercian_digits_(v...

I can see novice monks scraping 9900 on the walls.

Doesn’t feel entirely fair to call that base-10000. Practically, it’s just base-10 with an idiosyncratic and mandatory ligation scheme involving mirroring, and only defined for 0001–9999 and not 0000 for some reason (though it’s obvious how zero should be represented).

Still, the graphical complexity per "digit" is not much higher than of a latin letter.

Not in Unicode. (But neither is Bibi-Binary.)

> Using four consonants (HBKD) and four vowels (OAEI), one obtains sixteen combinations:

Why not ordered?

bcdf

aeio

The patent explains the reasoning, though I can't say I understand all of it. This is all in the french phonology

- open vowels for even, closed for odd (no idea why)

- back vowel for the second bit being unset, front for set (according to the wiki e is ə and central and mid rather than back and open but...), this kind-of matches the reasoning of the consonants positionality where the front sets the bit.

- H is the first consonant because it's mostly silent

- The patent justifies B, K, and D as labial, guttural, and dental. I guess physically indicating the bits set, with the low bit being the front of the mouth: labial is the low bit, guttural is the high bit, and dental being inbetween is both.

You can see the same physicality in the script, where "sharp edges" point to set bits when the nibble is written in pairs of two, big endian, columnar (so each block is read top to bottom, left to right, rather that left to right top to bottom).

Hence U stands for

    11
    00

which encodes 1010, or 0xA.

c encodes

    01
    01

0011, 3 while ɔ encodes

    10
    10

1100, 0xC (somewhat ironically).

not a fan of phonetics, but an interesting complication nonetheless, thanks for the info

Bobby Lapointe, the creator of the system, was a comedic singer who did a lot of wordplay, so building the system off of phonetics is entirely in line.

Here's one of his most famous songs: https://youtu.be/Brt3f-jlrao

Two other wordplay-based songs: https://youtu.be/MXmYP43AS98 https://youtu.be/BdWzG3tFVmM (no lyrics on those so good luck with em, especially the second one, the entire thing is homophones and double entendres).

Even did some english, though I expect the standard is not very high for english speakers since it was targeted at french audiences: https://youtu.be/FnFnms--ifk

Kaktovik numerals are quite nice for hex as well.

They are 5 based starting from one so there are nice symbols for 5, 10 and 15 digit. Not binary though.

> "7D0" in conventionally-written hexadecimal, would in Bibi-binary be spoken aloud as "BIDAHO".

I see what they did there.

What did they do ?

I'd gander that it's pronounced as "Be da ho' !"

I still don't get it?

Are these in Unicode?

No, this is something which doesn't exist right now. And I guess for greater good.

Nope, even finding similar glyphs is difficult for many.

The Logic Alphabet (linked to above at http://www.logic-alphabet.net/images/flipstick_2347_2.jpg), which has a similar concept, seems to have better support: it can represent twelve of the symbols using Latin letters (o, p, b, q, d, c, u, s, z, n, h, and x, although z, u and n seem a bit forced) and the others seem to have decent Unicode equivalents (ɔ or ⊃ (superset), μ (Greek letter mu), ɥ or ч, and maybe ʎ).

Unfortunate name considering the situation in the middle east

Definitely something that occupies the mind.

Reply on news.ycombinator.com