Gnomie Filksinger writes:

Leo, I read your short article on the difference between worms, trojans, and
viruses, and found myself wanting to set the record straight. Recently, I have
seen a number of articles that attempt to define the difference between a worm
and a virus, but they seem to have lost the technical definition. The proper
definition of worm and virus, and the difference between the two, has become
rather hazy in recent years, but they used to be clearly defined.

The short version is this: Viruses are files that infect executable files or
bootable drives, and only replicate when these programs or boot sectors are
activated. Worms are standalone programs that spread from one computer to
another on their own.

The more complex version:
A virus is a program that infects either a program or a boot record of a drive
(either a floppy or a hard drive), and are named for their similarity to
biological viruses. It does not function on its own, but only by infecting one
of these two items (as a virus cannot replicate without infecting a cell), and
waiting to activated by the activation of the infected file or drive. A virus
waits until either the file was run or the boot record was used to boot a
computer, whereupon it would infect other files or hard drives on that
computer. A virus does not typically travel from computer to computer on its
own, but must be moved by some other agency.

A worm, on the other hand, is a stand-alone program that finds and transfers
itself from computer to computer. Traditionally, it would autodelete itself
from a previous machine, thus moving from one machine to another, like a worm
traveling from one apple to another. It didn’t infect files or hard drives.
Instead, it would act to find other computers on a network that it could
spread to, and transfer itself to them.

This propagation method is unlike viruses, which only transferred directly from
a file to other files on the same machine, or from a drive to other drives on
the same machine. Traditionally, viruses carried malicious payloads and worms
did not, but that isn’t part of the definition.

Of course, that means that most e-mail “viruses” are worms in that they
propagate themselves across the network, and don’t infect executable files or
bootable drives. Oddly enough, most of these worms, while not viruses, are
also trojans, in that they require a human being to activate them in the e-mail
messages they travel by to infect a new machine.

Thus, an e-mail “virus” is technically an e-mail Trojan worm.