You may have noticed that the folks who have actually constructed keys to
identification are either telling you that the difficulties of the project
exceed its virtues, or are being very very quiet.

There are good reasons for this.

My feeling is that it would be easier, and probably more effective for
those wishing an aid to identification, to post photos and brief
descriptions of species on the Web than to attempt a key to the
Mollusca--the second largest phylum. Keys generally are most useful when
they have only a small number of steps, as in a key to the species of a
genus, or the genera of a family.

Keys are especially difficult to construct for identifying the higher taxa,
because taxa are defined wherever possible to reflect actual phylogenetic
relationships (genealogy), not current similarities in shell form. And the
more time animal lineages have to diverge, the more likely it is that the
overall group will not share any one shell feature. For instance, the
gastropods can be characterized as having a single, expanding helical
shell, but the exceptions to this "rule" are still obviously more like
other gastropods than like anything else. It is very hard to define
"mollusk" at all in morphologic terms that every mollusk must share. (DNA,
of course, is another matter.)

A good key can be great for identifying good material. As a paleontologist,
I find keys to be extraordinarily frustrating because I often deal with
fragments. There are good keys to local floras that accommodate this kind
of problem (keys using all features of a plant; keys to trees using only
leaves; keys to trees in winter), but it's not very practical to ask for
several sets of keys to animals that only a few people study at all (e.g.,
keys using only bivalve hinges; keys using only sculpture of shell
fragments; etc.). I find it more effective by far to use a well-illustrated
guide to a local fauna, or a reference set of actual specimens, than to use
a key.

So keys are harder to write than you may think, and harder to use also.
Yes, I am one of the quiet ones who have actually tried to construct keys.
It's not easy.

If people really want to make a contribution to identification on-line, why
not set up a site for digital photos of well-identified shells? If space is
a factor, the photos can be rotated occasionally, or the shells of one
region can be showcased and then another.

Quietly,

Andrew K. Rindsberg
Geological Survey of Alabama