Q:  If the uncertainty of a typical measurement is
            2 mas, how far away can a star be
            and still have a "good" distance?

  A:  If we adopt the very rough rule that one should
            only use parallax values which are at least
            10 times larger than their uncertainties,

               maximum "good" parallax  ~  10 * 2 mas

                                        ~  20 mas


               maximum "good" distance  ~  50 pc