A design algorithm of 1986 for minimal order linear functional observers that estimate Kx(t) directly for arbitrarily given K and with arbitrarily given poles, is based on a simplified design formulation that is only a single set of linear equations, and guarantees an observer order upper and lower bounds that are the lowest ever since. Since 1986, it has been claimed that this single set of linear equations is the simplest possible theoretical formulation of the design problem, and that these guaranteed observer order bounds are the lowest possible bounds. This paper formally proves that claim, and further claims and proves that this result of 1986 is the best possible theoretical result of the design problem. This new claim implies that the theoretical part of this problem is solved, that only the computational part of the problem has room for improvement, and this improvement is generically not worthwhile in practice if the algorithm becomes complicated. Only these extremely clear-cut claims can clarify the apparent misconception of the world control community. This misconception is revealed by the following two facts. 1) These claims are too many years after 1986. 2) They are all missed by the research community, because a relatively recent paper only reformulated this design problem to be much more complicated, proposed only the exhaustive numerical search to compute the solution of its complicated reformulation, yet self claimed the only solution to this design problem! If such a clear-cut result of 1986 can be looked down for so long, then any other good result can also have the same fate.