on a Unix
system that stores
(you can probably guess where this is going,
given the node
one's incoming e-mail
In keeping with Unix tradition
mailbox files have a simple format:
The messages are simply concatenated
each one beginning
(that is, separated from its predecessor)
by a "From line
recapitulating the message's sender
and the date it was received.
Everything else in the file
is just the individual messages,
more or less exactly as they were received
over the wire
These days, that means they're RFC822
with additional header lines
followed by the body text
To burst a mailbox file into individual messages,
all you have to do is search for lines
beginning with the five characters "From ",
preceded by a blank line.
Each such line demarcates a new message.
(The software that built the file was careful to
quote any message text lines beginning with "From " somehow,
so that they won't confuse the bursting algorithm.)
Once upon a time, your mailbox was simply a file in this format,
stored on your local system,
usually in /usr/spool/mail or /var/spool/mail.
Your mailreader simply interpreted the file for you
(implicitly breaking it up, as described here,
and presenting it as individual messages),
or if you wanted to,
you could peek at the raw file yourself,
using cat or more or
the text-manipulation utility of your choosing.
More and more, however,
incoming mail is collected on dedicated mail servers,
and is accessed using protocols like POP or IMAP.