Data analysis using program reports
- One use for the program reports is to track program usage over time as
an indirect gauge of the growth of the load on the system.
Every Enter does not represent an identical load on the system. Different entry points may have different instruction path lengths. Such differences tend to cancel one another out when aggregates of programs are considered instead of individual programs.
To assist the study of aggregates of programs, the OPTIONS file allows definition of packages of programs through program naming patterns.
- Another use of the program reports is to compare packages of programs.
The purpose of this comparative study is to roughly gauge the relative share
of packages on system load. Alternatively, differences in the design and implementation
of the respective packages can be roughly gauged.
For example, the z/TPF analyst may want to investigate the reasons underlying an observation that a package of programs has a comparatively high mean nesting level but a comparatively low percentage of ENTNCs and ENTDCs in favor of ENTRCs. One explanation may be that classical structured programming guidelines were followed too literally for the z/TPF environment. A high mean nesting level but a low percentage of ENTNCs or ENTDCs can mean that the package is holding on to more nested system resources than necessary and that the package is placing an unnecessarily large load on the system.
The z/TPF analyst can use the program reports to look for significant differences between packages and between subsystems, and then try to account for those differences.