Base64 and Uuencoding fundamentals 

Both Uuencoding and Base64 encoding use a similar technique to convert non-printable binary values in the range 0-255 to a set of "safe" ASCII characters that can be easily handled by most computer systems.

This example shows how a small binary data stream is converted into printable characters by the Uuencoding mechanism.

1. The data stream consists of three bytes containing the values 157, 97 and 226 (i.e. 10011101, 00110001 and 11100010)
2. Only one of these bytes (97) represents a printable ASCII character (lowercase "a"). The other two are eight bits long and would not transmit over a 7-bit link in any case.
3. The first step is to split the three 8-bit bytes into four 6-bit words (i.e. 100111, 010011, 000111 and 100010). This will give us a number range of 0 to 63.
4. In order to map the data to the printable ASCII characters, 32 is added to the value to give a value in the range 32 to 95 (i.e. ASCII characters from SPACE to UNDERSCORE).
5. In the example above this gives us 71, 51, 39 and 66 (0100111, 00110011, 01000111 and 01000010) which maps to G, 3, <backquote> and B. These are all "safe" to include in the body text of an email message.
6. Finally, a "begin" and "end" line are added and some additional formatting is included to indicate line length and data checksums.

Why is it necessary?

The first 32 values are not printable, indeed some such as XOFF would have catastrophic effects during data transfer as they are are used to control the flow of data from the host (XOFF turns the remote transmission off). This reduces it to 96 characters, but then DEL is not printable either. Worse still, some computer systems do not support lowercase characters, so Uuencoding concentrates on just 64 characters that are widely compatible. 

Even then, some computer systems (such as those running EBCDIC instead of ASCII) cannot properly substitute some of the characters, so a variation of Uuencoding, called XXencoding exists for this purpose.