Visualizing Quality

We are exploring ways to shift our quality mindset from "fix it later" to "fix it now".? In my last post, I discussed a specific strategy to encourage a quality mindset, Transparent Communication, and claimed that it needs to be more than a PowerPoint presentation at a lunch-n-learn.? If we are going to change our quality mindset, quality needs to be shown.? To show quality, we need to find some ways to measure it.

You can always go to Google or ChatGPT and ask about quality metrics.? If you do that, you will see things like Code Complexity, Code Test Coverage and Documentation Quality.? These are static code analysis metrics normally evaluated in the DevOps pipeline using a tool like SonarQube.? These metrics are fine, but they don't tell you about problem areas your already have.

Happily, there are some metrics you can create using data you already have.? Here are some alternatives that will help you see where quality can be improved today.

• Platform Currency - Evaluating the currency of operating system, DBMS and OEM software versions.? While it may not be necessary to be on the most current versions, making sure you are using versions that are still in support makes good sense.
• Incident Counts - Problem areas can be identified by the incident counts for an application.? While these counts can be aligned with the size of the application, a large, monolithic application is problematic, too.
• Known Vulnerabilities - A count of known cyber vulnerabilities in your operating platforms and OEM solutions.? (There is a relationship between platform currency and known vulnerabilities - as your platform becomes more current, you will have fewer vulnerabilities.)
• Adherence to Principles and Patterns - As solutions are being designed and developed, an assessment / review of how closely the solution follows your principles and patterns provides quality information.

As you implement these metrics and create methods to score them, it is important to realize that the truth isn't in the individual scores.? It is in the comparative scores.? For example, if AppA has a platform currency score of 75, that number by itself doesn't mean anything.? Compare that to AppB with a platform currency score of 131.? In a scoring system where larger values indicate less currency, you can say focus should turn to AppB.

By implementing these measures, you can gain a clearer understanding of your system quality and encourage proactive steps to "fix it now".