Rules


IEEE Spectrum Magazine, March 2005


Once upon a time, there were rules. They began to give to order to my career from the first day of my beginning electrical engineering course. It was then that I was introduced formally to Ohm’s Law. For a whole year thereafter I dealt with the many possible manifestations of this one great rule. There were countless problems involving such long-forgotten concepts as current loops and Thevenin’s theorem. The rest of my engineering education was similar. In these engineering courses there were relatively simple rules underlying all behavior. Educational life reduced to the endless repetition of the sequence of being given a particular problem, applying the rules, and deriving a solution.

When I first started in industry as an engineer in the Bell System, there was a similar rule-driven paradigm for business behavior. No one who ever worked in that company will forget something called the “GEI” – the giant loose-leaf binder containing the General Executive Instructions. These were the rules of employment. In that thick binder was the rule that would apply to any situation that might be encountered in the business environment. Need to have a paper cleared? What about outside employment, or when it would be proper to accept an honorarium? How should you handle an employee with a drinking problem? Every conceivable event had its own page with the relevant rules.

Those were the quaint days when there were twice-daily deliveries of paper mail. On each of these deliveries my mail would contain new inserts for the GEI. In addition to these new pages, there would be instructions that certain pages should now be removed. The rules kept changing, but mostly they just grew. It seemed like my secretary was kept busy just keeping my GEI binder up to date.

I used to wonder where these rules originated. Who was making them up? I imagined that it was the mail department writing them secretly at night in order to fill up their mail baskets and promote job security.

The omnipresence of rules was so palpable that the absence of a rule was a disturbing event in itself. Most of these disturbing events would be followed by new pages for the GEI. The greatest exception was the lack of clear cut rules for how employees would be evaluated in the annual performance review. Time and again one of my subordinates would complain about this. “How am I to be judged?” they would plead. “There must be rules.” The implication was that without written rules, management could not be trusted, and would undoubtedly be rendered incompetent.

After several decades of my life being governed by rules, one day they seemed to disappear. I don’t know when it was, and perhaps it happened so gradually that I didn’t notice until much later – like now. I surely don’t recall hearing any proclamation like “Henceforth there will be no rules.” However, the inserts for the GEI became less frequent, and then sporadic. The Bell System itself was torn apart and perhaps whatever organization was responsible for creating the rules was shipped off to some doomed offshoot.

My theory is that life was growing so complicated that the number of rules was growing without bound. Before long every employee of the company would need to be occupied with the task of writing and distributing rules, and no one would be left to do the actual work. Rules were becoming both too expensive and too constraining. Someone must have recognized this and decided that life would have to proceed without rules and would become -- well, fuzzy.

I think much the same thing happened to all my cherished rules of engineering. Somewhere, I am sure, Ohm’s Law still applies, but I’m no longer am confident about just where. As the size of circuits shrinks, the life of the electron becomes complex and fuzzy. In the presence of electromigration, parasitic effects, quantum tunneling and other phenomena of the small, the electron may not realize that it has to obey Ohm’s Law. For every rule I used to know, I have to stop and ask myself: What were the assumptions behind this rule? Do they still apply?

Worse yet, there is no Ohm’s Law for software. In the face of its enormous complexity it can’t be depended upon to behave as if it knew about any rules. I imagine a giant GEI containing the rules of behavior for software. Pages would get added every day until someone decided that enough was enough, and henceforth there would be no rules.

Living without rules gives us an uneasy freedom. We make up things as we go along amidst an increasing uncertainty and unpredictability. Life in the business world, and in technology, is fuzzy.

Robert Lucky