inconsistency in coercing to support functions

[RobinHankin]

In man/rrace.Rd, for example, I use the functions rankvec_likelihood() and preftable_to_supp() to do the same thing: coerce an observation to a likelihood function. This is confusing. Not least because I use "likelihood" to mean the same as "log-likelihood".

What I should do is use, consistently, [S3 generic] suppfun() as per R/suppfun.R and use it for the three different types of table. It might be desirable to call rankvec_likelihood() or even ordervec2supp3() if given a character vector or named numeric vector.

There is a note at the bottom of man/ordervec2supp3.Rd which I should implement in this connection.

But....do all this in one go.