A simple key or guide to the families of marine shells sounds like a good
thing, but writing one would be far from a simple project.  The way an
identification key works of course, is by gradually excluding more and more
groups, through the selection of increasingly similar characteristics, until
all but the correct group have been eliminated.  The problem is, some
families are so large and diverse that there are virtually no conchological
characteristics common to all members of the family.  In Trochidae for
example, there are tall slender species that look like auger shells, and
flat ones that look like sundial shells; smooth, glossy ones and coarsely
sculptured ones; thin and translucent vs. thick and heavy; open umbilicus
vs. closed; toothed apertures vs. toothless; nacreous vs. non-nacreous (both
inside and out); etc.   About the only shell features that all trochids have
in common are the presence of an apex and an aperture, which is hardly
sufficient to separate them, en masse, from other families.  Therefore, to
write a key that would simply allow a person to identify an unknown shell as
a Trochid would mean writing at least dozens, if not a hundred or more
separate branches, or sub-keys, taking into consideration all the diversity
of the family.  And even if a person took the time and did the research to
accomplish that, it would not be foolproof.  Nora pointed out the difficulty
in separating some members of Trochidae from Turbinidae.  Indeed, some of
them are so similar that a key based on shell structure alone could not
separate them.  The only way I can separate Astralium haematragum
(Turbinidae) from Trochus saccellum (Trochidae) is by the operculum.  If the
specimen doesn't have an operc, it could be either species, and therefore
either family.  Some other families, like Modulidae (and even some land
snail families!) include species that are very similar to Trochids.
One thing we need to keep in mind as shellers is that a shell is just a
piece of something, in fact a relatively simple component of a much more
complex organism.  Even though we commonly speak of "new shells" being
described and named, we know that it is really new species of animals that
are being discussed, not just their exoskeletons.  Often the assignment of a
species to a family is based largely on non-conchological characteristics.
And, we know that classifications based on the shell alone sometimes have to
be revised when the whole animal is eventually studied.  The best-known
example of this is probably Marginella (Afrivoluta) pringlei, the world's
largest margin shell, which was classified for many years as a volute, based
on the shell.  A key to families would actually be much easier to write if
ALL morphological and anatomical characteristics were considered.  The
problem is, once written, such a key would be impossible for the average
person to use, since we typically don't have access to the radula,
periostracum, tentacles, foot, eyes, reproductive organs, etc. of our
specimens.
Years ago, when I was teaching a high school zoology course, I put together
a box of about fifty assorted shells, and wrote a dichotomous key that would
identify, to species level,  just that particular group of shells.  This was
simply to let the kids get a feel for how a key works.  Even in that modest
venture, I found that some specimens had to be identified through more than
one branch of the key.  This was because a characteristic which would
clearly separate most of the shells into two subgroups might be ambiguous
for a few specimens.  Therefore, in order to ensure the correct
identification of those specimens, they had to be included in both branches.
It took me some hours to come up with a workable version.  (I know this must
be confusing for anyone who isn't familiar with such keys and how they work
- sorry).
Paul M.