I just posted this to a private discussion list, and thought it was worth repeating here:Chris McMahon wrote:
>For every company whose expensive Six Sigma project yields
>them no benefit at all, there is another company with no
>recognized quality process at all that succeeds wildly.
Have you ever studied Michael Porter's Competitive Strategy
Porter - a Harvard Business Professor - wrote industries go through a transition from wild growth and no standards to maturity and eventual decline.
Companies competing in the growth phase compete by differentiation of /product/. (Think the personal computer market in 1984). In the middle, standardization and consolidation occur, which is happening in the personal computer market right now. At the end, toward the right, you are dealing with commodities like Gasoline or Electricity that have no differentiation at at. Companies living in maturity and decline compete through standardization of /process/ and economies of scale.
Once in a while a disruptive innovation comes along, which can push the entire industry to the left. Consider, for example, book sales in 1993. Borders and Barnes and Nobles were mature businesses. They had defined processes and metrics - and they aimed to turn the corner bookstore into a memory. If you looked at where people spent money, not what they said, nobody cared about the service at the corner bookstore - they wanted variety, comfy chairs, and decap frappechino mochas.
Then came amazon.com with a disruptive business model and a disruptive model of scale - pushing the industry to the left.
That's what lots of software does - It pushes stuff to the left.
And when you are competing on the left - if you are Apple in 1984 or Linux Torvalds in 1991 or Napster in 1998 - you don't need great software development process. You need great ideas.
That's part of what bugs me about the discussion of software maturity. The real innovation and value isn't made in the land of maturity and standards. It's made in the untamed wilderness ... hmm.
I suppose you could call that 'Creative Chaos.'
So that's what I wrote to the discussion list, but this is my question to Creative Chaos readers:
Is what I wrote above the case? And if so, how should that impact the way we test software?
What do you think?