The fact, is there is no real definition of ``genus'', any more than there is of ``family'', ``order'', etc. (See ``What do terms like phylum, order and family mean?'' .) There's no scientific way to decide whether a given difference in anatomy is sufficent that two species differ ``at the genus level'' because the idea of a genus is so artificial: designating certain groupings as genera is really just an arbitrary but convenient way for us to partition up what would otherwise be an uncomfortably large set of taxon names.
As Thomas Holtz says, ``If you are more interested in book-keeping than science, then `genus' might be of great importance; otherwise, however, it doesn't have any particular biological significance.''
This is just as true in classifying extant animals as it is with dinosaurs. For example, the blue whale is variously known as Balaenoptera musculus and Sibbaldus musculus, according to the beliefs and preference of individual workers. Similarly, the American black bear may be Ursus americanus or Euarctos americanus, and the lion Felis leo or Panthera leo or even its own genus, Leo leo.
The only hard and fast rule here is that of precendence: that is, when someone considers that two or more genera are actually the same, then the older of the genus names must be used. For example, if Saurornitholestes langstoni and Velociraptor mongoliensis turn out to be species of the same genus, then that genus would be called Velociraptor (established in 1924) rather than Saurornitholestes (established in 1978); and the species would accordingly be called V. langstoni and V. mongoliensis rather than S. langstoni and S. mongoliensis.
In summary, there is no consensus even within dinosaur palaeontology - let alone across biological disciplines - on the what ``genus'' should mean and how inclusive or exclusive it should be in terms of morphological, behavioral or molecular diversity.
A someone once wrote, ``Linnean taxonomy is as much an art form as a science.''
Who wrote that, exactly?