When are we doing too much architecture, and not enough buiding?



Why we all know that architecture is required to create sound systems, however there is a point when there is too much architecture. So, what is the point of diminishing return when it comes to architecture? When is it too much?

When are we doing too much architecture, and not enough buiding?

Good question, David. I do know that not only do we sometimes get caught up in trying to architect everything, but every architecture has to be the best, most sophisticated and elegant architecture that one can imagine. While I do not down play the need to find creative and innovative solutions to problems, I do feel that we sometimes miss the forest for the trees. The architecture effort should match the need - longevity of the solution, level of reusability, etc. I continually push our architects to keep it as simple as possible. In some cases, this may mean SQL and DBLinks if the need is to exchange data between two systems for 6 months at which point one of the systems will be retired.

We need to measure success...

It will be interesting to measure success of approaches, from the very narrow to the very wide. I suspect the answer is somewhere in the middle, but i would love to do a study.