Dr. Gene Amdahl is widely regarded as the father of the mainframe because of his contributions to the System 360. In fact, he originated the term MIPS (Million Instructions Per Second) to describe processor speed. He also formulated Amdahl’s Law, which is used for projecting scalability limits of multi-processing systems, and Amdahl’s Rule of Thumb, which describes how processing, memory, and I/O capacity must be balanced to optimize scalability.
Gene left IBM in 1970, raised some venture capital, and spawned the plug-compatible computer industry. Amdahl Corp. soon became Big Blue’s most prominent competitor, succeeding head-to-head against IBM where behemoths GE and RCA had done so and failed, in spite of investing billions of dollars. Such were the accomplishments that made Gene legendary. His achievements were acknowledged by The Times (London), which declared him one of the 1,000 most influential people of the 20th century.
Gene and I have been good friends for more than a decade. We frequently talk and enjoy discussing several delicate topics, including the big three—religion, politics, and Microsoft’s antics. Several years ago, we even traveled to Africa together.
Since Gene has a broad view of our industry and so many experiences to draw from, I asked him to collaborate with me on a few timeless management tips that IT executives in large computing centers might find helpful. Here’s what we came up with for those big iron warriors known as mainframe executives (listed in random order):
ONE Celebrate achievement. You’ll encourage contributions by showing appreciation of those received. Seeing that contributions are actually recognized and welcomed generates an atmosphere that engenders the desire to contribute among all levels of employees.
TWO Be open to new ideas. Everyone approaches things from their own framework of knowledge and experience. When we first discussed the concept of the PC, the year was 1959. I (Gene) made some wrong assumptions about the design of the PC that limited its potential. The logic limitations and limited I/O pins at that time seemed insufficient to make a reasonably robust computer, as viewed by a mainframe designer.
THREE Explore areas of opportunity in both individual and small group sessions. Underlings often are unwilling to voice their opinions in the presence of their boss, so a private, informal drop-in meeting can often be a means of getting alternative views of proposed projects. Getting ideas on the table is the first step to implementing them.
FOUR Evaluate probable potential in the early stages of development. The early stages of defining individual elements of the project often uncover difficult aspects that can make the project more costly than planned, perhaps calling for provision of more resources or even cancellation.
FIVE Invest in software testing. Companies should thoroughly debug developed software before offering it. Microsoft is a big offender of this principle, costing its customers dearly in lost productivity. The cost to the early user may be high enough to engender publicity that can undercut the success of both software developers and users. Costs accrue to the customer and the company. A “bug” costs much more to fix in testing than after it gets to the field. Some companies just eliminate the “fix” later to save money. If they really paid to fix problems in the field, they, too, would invest in testing. Testing also eliminates “feature creep.” Developers can’t just throw in garbage as it goes out the door because it won’t be tested. IBM’s approach was effective. Each delivery of the operating system was like catching a train. Get your package on the train before it left the testing, documentation, and alpha-testing station, and it was on. Miss the date and catch the next train.
SIX Keep tech-savvy people around you who you can trust. Trust is earned by demonstrating an ability to complete IT projects that deliver the full requirements on time and under budget. When projects get into trouble, don’t confuse success with salvage when project requirements are trimmed to meet deadlines.