Can we discuss the right interpretation of short-term & long-term process capability study, according to AIAG stadards?

You may want to look at some of the previous discussions in the 'Similar Discussion Threads' below (just scroll down).

You might also read:

Statistical process control for precision machining (http://elsmar.com/Forums/blog.php?b=79)

It will illustrate the case where there is no difference between short term and long term capability studies, as capability is a constant value for a uniform distribution: (USL-LSL)/(UCL-LCL).

just fyi
in our company , we do it like this

short term=calculation with last month data
long term=calculation with last 6 months data

then we compare the short Cpk value with the long term Cpk value.