Bits and bytes?! Help!

I think you are confused because you have it backwards. You are asking how many bits or bytes a letter represents.

But a letter doesn't represent some number of bits or bytes. Why should it? The letter A is just the letter A, it's been around forever. It has no use for bits or bytes, it's just the letter A.

Look at it the other way around: How many bits or bytes does it take to represent the letter A?

The answer is: It depends? How many other letters do you want to be able to represent? Suppose it's only 2: A and B. Well....you can do that with 1 bit. Let's say 0 means A and 1 means B. As long as you and I agree to that, then we can store A's and B's in 1 bit.

Or maybe you want to represent all the letters from A to Z. There are 26 of them. You can do that with 5 bits (5 bits will give you 32 unique combinations. So you get the 26 letters and have 6 left over to do other things with).

With 7 bits, you can represent 128 different things. That gives you enough to have A-Z and a-z and 0-9 and things like !@#$%^&*() and a few more. That's what ASCII did...it took 7 bits and mapped them to A-Z and a-z and 0-9 and a few other things.

But..you are right...most computers have 8 bit bytes. So there is an extra bit to play with here. That gives you 256 different things to represent. That extra bit got used for all sorts of stuff. IBM used it to add a bunch of line drawing characters, and a few foreign (as in non-American) characters. But nobody ever really agreed with what IBM did, so they came up with their own idea of what letters those 8 bits should represent.

Then the world discovered that there are more than 256 "letters" in the world. I think Japan figured it out first. IBM gave the world 256 different "letters". And someone in Japan said..yeah, so how do I represent 'ローマ字'?

Some convoluted schemes were tried. Google "Shift-JIS" for an example. It was a PITA.

Then Unicode came along. Unicode says...8 bits are not enough. We need 16 (or maybe even more?). 16 bits lets you represent 65,536 different letters. That's almost enough.

Sooo...The letter A can be represented in 1 bit. Or 7 bits. Or 8 bits. Or 16 bits. The letter A doesn't know or care. It's how you and I (and the rest of the world) interpret the bits that matters.

Once you swap the question around it makes sense. Don't ask how many bits a letter represents. Ask how many bits it takes to represent a letter.

A character is either a letter, a number or a special character like asterisks or hyphen. Any key on the keyboard.

Traditionally any given character is represented using 7 or 8 bits, in other words a full byte (the last bit in the case of 7 is simply ignored).

Computer memory works in such a way that only bytes can be addressed, so information in computers are always stored in a multiple of full bytes (8 bits make out one byte).

With ASCII, any character (alphanumeric and special characters) are stored with 1 byte. With the ASCII system, you have 256 (2^8=256) different characters.

However, there are some character sets that allow more characters by using more than one byte to store character data. An example of this is UTF-8. UTF-8 uses one byte to represent all the most common characters, but more exotic characters like Æ, Ø and Å is stored using two bytes. With UTF-8, your computer knows whether to interpret the next byte or the following two bytes as one or two characters by matching the first byte with a certain prefix. If it doesn't match, then the two bytes are two separate characters. If it has a certain bit-prefix (e.g. the first few bits of the byte), then your computer assumes that the two byte together represents one character.