dcsimg
 

Using the Laney P' Control Chart in Minitab Software Development

When I get to apply tools like the P’ Chart, I get a thrill. What can I say? I’m a huge fan of Minitab® Statistical Software. I won’t claim to be its biggest fan, but let’s just say I’m definitely up there. Probably in the top 50; definitely the top 100.

Admittedly, I’m biased. I have a pretty good gig here. But my history with Minitab started more than a decade before I came to work here. As a quality engineer who spent a lot of time analyzing data, I was a die-hard Minitab user.

I still get excited about each release of Minitab software. That’s partly because I played a small part in creating it, but mainly because I truly enjoy using our software—and because new features we put into our software to help quality professionals help me, too.  

Minitab, like a lot of great companies, uses Minitab Statistical Software to assess and improve its own processes. We gather data, analyze it and look for ways to continually improve our ability to deliver. 

For example, the Laney P’ Chart solved a problem we were experiencing with our continuous integration metrics.

Control Charts and Continuous Integration

Minitab uses a continuous integration system to manage changes to our code base. This means that each time a developer makes a change to our code, thousands of automated tests—actually, tens of thousands—are run against the change to ensure that no problems were introduced. Our goal is to prevent defects from entering our code base, so we don’t have to address them once they have. This results in higher quality, more efficient development.

When someone’s coding change causes one of our automated tests to fail, we call that “build breakage” and we refer to the frequency of build breakage as our “build quality.” We take build quality very seriously at Minitab and we use a P chart to track it. 

How the Laney P’ Chart Helped

A traditional P chart assumes that a process has a constant rate of defectives over time.  But this is not true of our data. The between-subgroup variation was not accounted for, causing the limits of the traditional P chart to contract. As a result, we were seeing “false alarms,” or instances where our process appeared to be out of control when it really wasn’t.  

This meant I was knocking on developers’ and testers’ doors more often than I should have been.  Now we’re able to use the Laney P’ chart, which does not assume a constant rate of defectives, to address that problem. No more false alarms. Everybody’s happy.  

In the control chart below, you can see a sample of our continuous integration metrics. By monitoring the process, we can evaluate results, implement process improvements (identified by the stages) and address any process changes or special situations when they occur.

 Laney P' Chart of Build Breakage

 

So speaking from my role as a dedicated, longtime Minitab user, I want to offer my personal kudos to our development teams—and, of course, to David Laney!

You can find more information on the new control chart options at http://www.minitab.com/Published-Articles/On-the-Charts--A-Conversation-with-David-Laney/ 

Comments

Name: Mohammed A Mohammed • Monday, June 11, 2012

Dear Dawn

I am trying to find an e-mail contact for David Laney.

I co-authored a paper with him in 2006 on overdispersion in healthcare data (Mohammed MA. Laney D. Overdispersion in health care performance data: Laney's approach. Qual. Saf. Health Care 2006;15;383-384), but his original e-mail is not working now.

I would really appreciate your help on this.

with thanks
Mohammed


blog comments powered by Disqus