we begin with a setting in which all projects are finely divisible, i.e., may be increased or decreased by small amounts. as will be noted later,this is not a very realistic assumption, but it permits us to bring out the basic rationale of project selection