@Setanta,
His basic thesis has absolutely been borne out, it's one of the fundamental ideas of modern biology.
If I remember right, the thinking goes along the line that his data fit very nicely with models of inheritance based on a small number of alleles resulting in a small number of traits -- but that the traits described also are subtly influenced by multiple genes, and that Mendel's data fits his simple model very nicely, but is quite improbable if the inheritance patterns of actual pea plants are taken into account.
Or something. I think it was brought up as a side note in a genetics lecture several years back.
At any rate, you can definitely have a good idea and do bad science. You can have a terrible idea and do very good science, too.