First binary encoding of the alphabet?

Tony Duell ard at
Sun Nov 28 13:11:24 CST 2010

> I've been reading William Blair's 1807 article on cryptography, 
> called "Ciper": 
> Around PDF page 16, near the bottom of the first column, he proposes 
> a code that represents letters of the alphabet, using the symbols a 
> and b.  What's interesting is that this is nothing more than a binary 
> code, with A being aaaaa, H being aabbb and so on to Z being babbb.

I think that iether hte letters are not encoded in order, or some are 
missed out, then.

If 'a' = 0 and 'b' = 1, then A=00000 = 0 (OK), H = 00111 = 7 (again, OK 
if the letters are encoded in order), but Z = 10111 = 23

> Is this the first mention of using a binary code to represent letters 
> of the (Roman) alphabet?

I doubt it. Since a telegraph line has 2 states (on/off, mark/space, etc) 
you could claim that Morse code is a binary code with '.' being 
equivalent to the bit pattern '10' and '-' being '1110'. So for example 
'S' = ... = 101010, 'P' = .--. = 101110111010 and so on.

I don't know where the idea of binary encoding for various things
originated. I do have an old book on electrical wiring (about 90 years
old) that suggests putting 15 lights in a 5*2 matrix on the ceiling of a
large hall and then wiring them to 4 swtichse (cntorllingg 1, 2, 4, 8
lamps respectively) so you get 16 levels of illumination (from all off to
all on), the lights being arranged to get a fairly evcen distribution of
light in all cases (for exampole, the '1' switch controls the lamp in the 
very centre).

Incidnetally, the idea of assigning a 5 bit code to eack letter is 
commonly used to look for hidden messages in religious texts, the works 
of Shakespear, and so on. Basically, you pick a 5 bit encoding, then 
decide on how to split up the letters in the original manuscript into 0's 
and 1's (this may depend on how they are written, or whatever), and see 
if you van get a readable message. Since you hxve so much choice in 
encoding, etc, it's hard _not_ to find soemthing, but...


More information about the cctech mailing list