For the following analysis, fifty programs were selected at
random from among those currently being developed or modified
by the programming staff at the authors' place of employment.
These programs varied in size from a few hundred lines of code
to over seven thousand lines. There were forty-one different
authors involved in writing these programs. Thirty-one of the
fifty (62%) were written within the past year, fifteen were from
one to five years old (30%), and four (8%) were over five years
old. All were written since 1980, and all were structured with
no GO TO statements (except to the EXIT of a PERFORM THRU construct).
Each of these fifty programs was compared to the twenty rules
developed earlier in this paper and it was noted whether the program
followed the rule, did not follow the rule, or whether the rule
was not applicable to this program (e.g. most programs did not
do sequential file matching, so the file matching rule was not
applicable). The exception was rule nine, which recommends not
to use any particular USAGE mode. For this rule, the type of
USAGE actually used was simply noted. Appendix B contains a summary
of this analysis. Chapters 19 thru 23 list each of the rules
and note how the programs followed or did not follow them.
Finally, seven of these fifty programs were chosen at random
and were modified as necessary to adhere to these twenty rules.
Chapter 24 contains the results of this program modification.