One of the commonly requested features for OWL is some form of key
support, generally phrased as allowing inverse-functional datatype
properties. For a variety of technical reasons, these were not included in
OWL DL (though they were available in OWL Full). OWL 2---a revision to OWL
which is under development by the W3C---introduces a form of key
representation, so-called ``Easy Keys'', which avoids much of the
implementational pain of general inverse-functional datatype properties,
``dumbs down'' nicely to weaker languages (like RDFS), and can be
specified (or even implemented) in terms of \dlsafe rules. The Easy Key
approach also allows for a form of compound key. In this paper we
explicate the design and rationale of Easy Keys, sketch a decision
procedure for them, and compare them with alternative possibilities,
including general inverse-functional datatype properties.