Re: Shewhart's Maximum Control

[SOPKrules@aol.com]

Danswart recently responded to RNM@y12.doe.gov asking him/her about what was 
meant by "modeling."  RNM@y12 previously posted a note to the group talking 
about the importance of  'matching the exponential distribution when modeling 
cycle time.'  RNM also added that it is important to use the control chart to 
look for special and common causes of variations. 

My assessment about the thread is that you should use the histogram, control 
chart and design tolerance in a single equation to solve for Process 
Capability symbolized by any of the Cp index numbers.  Use them all to make 
the assessment about your processes' capability of meeting the specified 
design limits.  The tolerance range consists of design specification limits 
and are preferably two sides with an in between target.    

Report the capability of the process and base your assessment on the 
statistical control and your reasonable estimate of the what the process 
should deliver.  And I don't think that many processes show statistical 
control.  I have seen some though.   They just appeared after a lot of work 
and coordination.  

By the way, are any of you familiar with Henry Mintzberg's work on the cost 
of coordination?

Bob
==============================================================================