Changing the cultures of engineering organizations is about as easy as herding cats. More than ever, though, companies must accept and adapt to change, as well as challenge traditional dogma. This is especially important for semiconductor vendors. The era of system-on-a-chip and deep-submicron design makes it a prerequisite for success. I don't mean to get on a soapbox, and if the truth need be said, I don't think that engineers are really opposed to change. Rather, they just want to be sure that proposed changes are for the better. Too often, engineers are yanked in several directions by managers who want it all-shorter development cycles, better-optimized products, lower development costs and designs that are easily reusable.

In benchmarking different chip-design organizations and measuring their productivity, I've found this to be a commonly overlooked problem. For example, engineers and managers often hold different views on the goals and objectives of a project. Yet they don't know that their respective views differ because the manager has not made the hard decisions. What's more, many times there are dramatic differences in understanding even among the engineers themselves. One guy thinks having a certain feature in a product is so important that the product will fail without it, whereas another guy thinks that getting the design taped out is paramount. Another guy thinks that the company's cost-control program is such that he won't even mention the idea of getting a few more simulation licenses to speed up verification. The problem manifests itself in schedule delays, multiple spins and frustration among team members.

I know this must sound obvious, but you would be amazed at how widespread this problem is. How can a chip-design team hope to succeed if everyone isn't marching to the beat of the same drummer? Yes, management wants it all, but good managers know that this is not possible. They tell the team, go faster, make it better, don't spend so much money, so the team runs around like a bunch of chickens with their heads chopped off.

Within the more successful teams, there's a common view on the exact weighting of the various priorities. Faster, better, cheaper are always priorities, but the real question is how much faster vs. how much better vs. how much cheaper? These are the decisions that good engineering managers must make.

— Ronald Collett is president of Collett International Inc. (Santa Clara, Calif.), a management consulting firm. He can be reached at ronc@collett.com.