Home | Login | Recent Changes | Search | All Pages | Help

OneChallengeOfCulturalChange

Currently the most visibile aspect of my work is cultural change. The reactions to what I am doing range form enthusiasm to cynical skepticism. On the cynical side is a culture where people have seen various good or reasonable initiatives fail because the culture couldn't figure out the economic model that provided support for the change to continue. I have been working in an enterprise area where we want to identify opportunities for components and define the business requirements for those components. Statements from management reveal a desire to obtain those benefits, and the day-to-day projects are still under the gun for schedule and $ commitments.

I am interested if any of you have tackled this problem and what successes and failures and what insights you might offer to me.

- BeckyWinant 2003.09.08


A little self-promoting, but why don't you look at the chapters in Quality Software Management, volume 2, for some methods of producing First-Order measurements you can provide these execs on the value of your various changes. - JerryWeinberg 2003.09.09
Jerry, I've already incorporated ideas from your book. :)

My measures so far include Usage measures: (a) stats related to tracing the models used to gather requirements for the components to the components and the systems they are used in, (b) Stats on the architecture documents that the models are used in. Satisfaction based on surveys: Architects and Component Builders.

I'm struggling with the coverage measures and the quality measures. I can look at the number of things (classes, associations, states etc) that are built - breadth - and compare them to the number defined - depth. I thought about getting an estimate of my probable target by looking at sizes of data models and adding in for behavior and then looking at what we have compared to that number. I have some others that I am not remembering ... and will add in later.

Part of the struggle on coverage and quality is what to publish. Another challenge is that the model will take time for requirements gathering and verification and then going through the component build cycle to get to an application.

What early statistics might you suggest for such a situation? - BeckyWinant 2003.09.11

- BeckyWinant 2003.09.11


Becky - As I read your description the measures I was wondering about how they relate to the bottom line. Is it a given that if architects and component builders are more satisfied with coverage and quality, that the things they build will make (or save) more money for the company? If not, are there things that can be measured that will close the loop? BobKing 2003.09.11


Updated: Thursday, September 11, 2003