Hi. You've reached Nigel.in
What you have landed on, possibly inadvertently, is the website of a certain curious character a.k.a Nigel Fernandes
The aim of this page was to serve as a landing point for those of you who want to know a little more about me. The links on the right and left will take you deeper my world, or possibly on to different exciting things.
Its a big web out there and this is just one more street for you to saunter down. I hope you like it.
Close this.
About Me
- Name: Nigel Fernandes
- Location: Panjim, Goa, India
Crazy, dancing, programming, Goan.. I'm a computer geek and proud to be one. I program in Java, Ruby and .Net, PHP, Javascript. A lot of my recent work has been about CSS and UI design practices for large scale websites and Agile teams. I still while away hours dreaming up a web startup.
Smelling Bugs
Thursday, August 13, 2009
A few months ago, we learnt that our team had the highest bug count per point delivered as compared to any other team. Since the program of work is composed of several teams all working on the same code base, this was a point of obvious concern.At first we tried throwing more QA's and more QA related process at it. Seemed logical but did not help much. In fact intense QA and QA process meant that story cycle time went up.
Then we started analyzing the bugs, and found fairly distinct trends based on the nature of the bug report.
We had a higher percentage of bugs caused by :
* Implementation defects.
* High levels of customer feedback coming in as bugs from UAT.
* The customer expecting more than the story scope.
Our actions:
* Return to core XP practices. Much more focus on pairing and pair rotation.
* Spend more time on story analysis. Get the programmers to pair with BA's on story definition.
* Tightened story acceptance criteria.
The results are trickling in now. Looks like we got our actions right.
Lesson: Pay attention to the nature of your bugs, not just the bug count.
Labels: bugs, projects, software
Sometimes people are just too focussed on the numbers and don't want to see what they mean.
My take is that the exact phases through which 'a' bug slipped through is not very relevant.
I do agree there is value in analyzing the trends in the phase of leak through.
Subscribe to Post Comments [Atom]
<< Home