Question common wisdom
Jon Titus, Editorial Director- September 1, 2002
Journalists routinely put "Moore's Law" into descriptions of semiconductor technology. These writers talk about microprocessor speeds doubling every x months, the size of transistors shrinking by half every year, or some other simplistic measure of progress. Perhaps these journalists feel they'll awe people by quoting a law that promises a simplistic doubling or a halving—concepts most people can visualize. In almost every case, though, journalists repeat anecdotal information and misrepresent what Gordon Moore actually said in 1965.
If you read Moore's original article (www.intel.com/research/silicon/moorespaper.pdf ), you'll find he described a logarithmic relationship between time and "the number of components per integrated function." Moore's logarithmic plot shows five points between 1959 and 1965, connected with a rough straight line. After the final data point for 1965, Moore wisely drew a dotted line. The word "law" never occurs in the article.
In natural science, laws arise from careful research and the evaluation of many hypotheses. Only after the test of much empirical data does a hypothesis become a theory, which undergoes even more scrutiny before it becomes a law. Laws must stand the test of time and must hold up for all conditions and variables. So, if writers wish to call Moore's work a theory or a hypothesis, so be it, but let's not call it a law. Even Moore himself has started to doubt that the straight line he drew in 1965 will continue along the same inflexible path that a law demands.
Casually attaching the word law, rule, or regulation to an idea can stifle creativity. People get the idea that if we call something a law, it's unchangeable. A law is a law, so why fight it and try to do something new? In the early 20th century, physicists had to battle against laws that could not explain physical behavior at atomic levels. The community of physicists split into those who rigidly tried to fit new data into old laws and those who decided that at some point, the old laws no longer held true. Thus, quantum mechanics was born by developing new theories and ignoring the older laws of Newtonian mechanics.
Just decreeing something a law doesn't make it one. And even valid laws may not apply in every case. So, the next time someone says you can't do something because it violates a "law" or because "that never worked before," ask for the details, and then keep going. Who knows, one day someone might name a new theory after you.