I disagree with your hypothesis of Diffusion being Incomplete Dominant. For Diffusion to be Incomplete Dominant, the heterozygous would need to be
consistently a different phenotype and it isn't.
Sometimes heterozygous Diffusion express characteristics of the homozygous, but not always. The same goes for Caramel...
Sometimes heterozygous Caramel express a more yellowed color. It is however, not a sure-fire way to identify hets for either mutation.
It has been suggested before to call cases like Diffusion and Caramel "Variable Dominant," though I have not heard that term in quite some time.
Also, Motley x Stripe produces
het Motley Stripes (Motley-Stripes or Motley het Stripes, however you personally would like to work it). The offspring produced always have one copy of each gene (Motley & Stripe). If you pair a Motley with a het Motley Stripe, then you would get a mix of Motleys and het Motley-Stripes.