Re: specifying an initial guess for non-linear problems with no ebcs