Our Quality Manager has decided that we need a more formal means of measuring software / code quality. She wants us to produce some kind of metric - bugs per lines of code, for instance - but to our Development teams any such number is meaningless (what defines a bug? how do you decide the seriousness of a bug? etc).
What other metrics do your organisations use to measure Quality? Is there a way of producing some "number" that enables comparisons across projects, while still retaining some sense of what actually happened in the project?
Thanks! Saving Changes...
Sort By:
Michael WoodProject Manager / Business Analyst / Business Process Improvement Guru| Independent ContractorGig Harbor, Wa, United States
Alison,
Certainly the quality of applications is important. However the metrics to measure that quality should start from the outside in. The Quality Assurance approach is inside out. I suggest that you start with metrics that measure the application’s conformance with design. Functionality trumps bugs per lines of code every time.
Imagine a metric on your new home that was based on number of bent nails vs the functionality of the home or it conformance to your design.
Metrics that micromanage misdirect and can actually detract from quality.
Put more focus on design and less on construction. A good design will be easier to program and maintain.
So base quality metrics on Stakeholder needs and expectations. That is where the real quality will shine.
Good Luck
Saving Changes...
Michael CunninghamPresident/CEO| Harvard Computing Group Inc.Devens, Ma, United States
I agree with Michael's post. In addition, I always found that with very complex projects automated test procedures and having the team members fix their own code (as opposed to a separate maintenance team) made a huge difference. Often the timelines that are measures are incorrect, such as time to beta, or delivery of final code. Final code is only represented when in successful production, or when users are paying good money and satisfied with the system.
We use a High, Medium, Low categorization for bugs, and then an escalation procedure to ensure that serious issues are dealt with quickly. Saving Changes...
