How will you identify if a COBOL program is having double precision computation or not ?
there are 2 ways to search for floating point fields...
1) the application specificatios and the record layout
( if no floating point is cited in the specifications then there shouldn' t be any in the programs )
2) look at the program/copybooks sources to find the comp-1/comp-2 constructs
( use the ISPF search for utility )