Tuesday, June 28, 2011

Limits of Complexity

I used to say that I was good at what I do (software design) because I made mistakes faster than anyone else that I know.  I was simply following Thomas Edison’s lead and trying as many different options as I could.

Unfortunately, past a certain point of complexity, this is no longer practical.  Testing a particular option requires so much commitment, background work, personnel involvement, data recording, time and expense that only a few options ever get tried.

In many cases, the first successful trial becomes the final product.  “It works.  Ship it!”  Even though better solutions have become obvious, implementing them is just too time consuming or  expensive.

And many concepts never get tried at all, simply because the expense is too great and the outcome is not guaranteed.