Throughout our lives, we make decisions based on the facts at hand and the likely outcome of our actions. A classic moment in the film The Empire Strikes Back comes to mind. As they approach an asteroid field, C-3PO tells Han Solo, “The possibility of successfully navigating an asteroid field is approximately 3,720 to 1.” Solo replies, “Never tell me the odds!”
This type of exchange is repeated in another Star Wars film, Rogue One, with K-2SO telling Cassian Andor and Bodhi Rook the chances of failure and once again, Andor doesn’t want to hear it.
Han Solo and Cassian Andor don’t want to know the facts because they’re not relevant to their decisions. Each knows that their chances at success are nearly impossible, but they will proceed anyway. Is this how we approach our software development projects? Do we proceed without facts? There is something to be said for reckless abandon in a movie, but perhaps not in a development project. As we approach our own “asteroid fields”—the bottlenecks and other anomalies that slow development—do we want to exhibit boldness, but this time, informed by relevant information?
If yes, then what information is helpful to that decision and where do we get it? One good source of data around mainframe software development is BMC Compuware zAdviser, a service provided free with current maintenance, which continuously measures your processes and uses machine learning to help improve them.
Here are some examples of information zAdviser can provide:
- Learn exactly how long it takes on average to go from checking out code until test, or even to production. You can use this metric to identify high-performing teams and learn from them.
- Identify programs that are taking significantly longer to work with. These may represent areas where refactoring can make a difference.
- What does “Good” look like? zAdviser aggregates all data allowing you to compare your KPIs to all other customers (data is kept anonymous).
- Look at percent of Abends escaping into production compared to your overall abends. These are the abends that your users are going to experience.
- A Development Manager saw that it was taking 3 weeks to test the code the developers took 4 days to write. He commented that this is what his developers have been telling him for a while but there was no concrete method of proving or showing it to senior management. If you are looking to implement automated testing, these are the metrics you can use for your baseline and for measuring progress. You can read more about it in “Improving Quality and Shift-Left Testing with BMC Compuware zAdviser” by Spencer Hallman.
zAdviser provides visibility into successful practices and helps detect problem areas, delivering actionable insights so you can address them and proceed boldly but safely, avoiding the asteroid fields that can decrease the quality, velocity, and efficiency of your mainframe software delivery.
Learn more about improving your DevOps processes with actionable data in the e-book, “Charting Your Mainframe DevOps KPIs for Continuous Improvement.”