* 



a nan to laaoolr the shop delivery system; instead of thinlring 

 it more efficient, inoro economical, lio uonld c eminence to 

 Imoeh it. Dut instead of doing that, we installed one man, - 

 one Lie ss one or, an ^ v/0 seloctod some particular little part 

 of the \JQ?~.: arid uo used tlie messenger system in that partic- 

 ulay parw of tlio wor:.: and T73 lot tlie otliei pc.i^ r;o Then we 

 coiaparocl the ooet of doin^: those t~o tilings -ondor as ro arly 

 e::aotly t)ie saiao concLition ac -. -o ;;-ocibly cciild. 7e \."o^lc 



^ 



verify- a raoasuro. This noc.suro is a pur 017 arbitrary one; it 

 is not necessarily the louest one. It io "purely an arbitrary 

 measure; ; - oii nay ta?._o a nore e::pensivo systoa as an arbitrary 

 measnro. 



Ho'/, you gather all tLe olomonts that go in to nalie 

 up the cost. You array those figures on one side; on the 

 other you array the elements and the figures that go to inalce 

 up the oost of -'our shop clo" ivory S3 r stein. lTo\7, you .luct tahc 

 tvro G ets , not the T/hole shop; you tc.ho those "before the raon 

 and say: "Now here, what is there in this that oan he elimi- 

 nated, what is there in the other?" Instec.d of sa3 r ing: "This 

 is the "best", or, "That is the "boot" you cay "Here aie two 

 systems. How, naybo yours is the host, "but hero v;e have an 

 analysis of the cost. 7o i;ill aualyzo them, ho 7 far are thoy 

 concurrent. Oan r ./3 set one over t?\e other and roduce it to 

 the lowest terms. *."Je will eliminate li o factors in "both. 

 You see? ..ft or having eliminated li':o factors in "both you 



have arrived at the things that are incom".ati"ble in "both of 



315 



