Some edge cases with respect to parameters

I'd like to remove p:parameter from p:declare-step. Now that pipelines
can compute parameter names at run-time, I don't think we're getting
enough value from the ability to limit the QNames of parameters, or
declare default values for them, to warrant the complexity of allowing
it.

A global default for the value of a parameter on every instance of a
step of a particular type doesn't seem like it would have much value.

Does the presence or absense of a p:parameter-input port have any
impact on the ability of a pipeline author to put literal p:parameter
elements on a step? I could live with the constraint that parameters
are not allowed on any step that does not have a parameter-input port,
but I don't see a lot of value in that, except perhaps consistency.
Why allow literal parameters but not computed ones?

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | A philosophical contempt of life is no
http://nwalsh.com/            | guarantee of courage in the face of
                              | death.--Gustave Vapereau