conspire 1. to agree together, esp. secretly, to do something wrong, evil, or illegal. 2. to act or work together toward the same result or goal. 3. to plot (something wrong, evil, or illegal).
Random House College Dictionary
Was a web of conspirators involved in the JFK assassination? How did Marilyn Monroe really die? What did FDR know prior to the Japanese attack on Pearl Harbor? How did Howard Hughes waste away in the splendor of his penthouse? We may never know the "truth" in these mysteries, but Americans tend to lean toward the conspiracy theory. I'll introduce my own conspiracy theory -- a conspiracy that may never be the subject of a movie but, nonetheless, affects each of us.
An agreement may be implicit or explicit. A conspiracy can take place through the use of winks or nods as opposed to an explicit agreement, either orally or in writing. Picture the medieval inn with some wayward travelers staggering in from a storm. As they sit at a table and complain to the innkeeper about the long road ahead and the need to be on the lookout for highway bandits (they carry a gift for the king), three unkempt rowdies sitting at a nearby table listen to their tales, nod to one another, and slip out into the night. What do we expect? The listeners have an implied conspiracy to rob the travelers without having spoken a word.
There are conspiracies of omission as well as commission. An unpopular supervisor is raked over the coals by upper management for failing to meet production guidelines. The staff is aware that the goals were unrealistic (and have been active in ensuring they would not be met), but when asked for comment, the staff fails to mention the relevant details that could have saved the supervisor's job. The result is a new supervisor and a successful office conspiracy of silence.
Now I'll tell my tale on intrigue. Each year, the U.S. government, private industry, and educational organizations spend hundreds of millions of dollars on important applications requiring the use of supercomputers. These subjects include climate and weather modeling, aerodynamics, materials modeling, cosmology, the human genome project, combustion engineering, ecosystem modeling, and many more. The eternal cry of the research establishment is that they are always just three orders of number-crunching magnitude away from the ability to model the real world. First, we needed gigaflops (10^9 floating-point operations per second or FLOPS); now we need teraflops (10^12 FLOPS). The hardware industry fans these flames of desire by pushing technology to produce expensive equipment to meet these perceived needs. We salivate as each new hardware release promises more power. Congress responds enthusiastically with a deluge of funds to find technological solutions to the nation's woes.
Unfortunately, we may be living a lie. Does an unwritten conspiracy of omission exist among the agencies, industries, and universities demanding this new technology? Examine the evidence and draw your own conclusions.
In the early days of scientific computing, we attempted to solve relatively simple problems, perhaps a million total floating-point operations. Granted, they were problems that could not be solved readily without a computer. However, they were tractable, and, in theory, we could solve them with a pencil and paper.
Times have changed. The problems we are tackling now cannot be solved by hand in the lifetime of a normal person, nor in the period of recent history. (See the sidebar "How Long Does It Take?") The complexity of the problems we are attacking leads to another potentially more serious problem. How do we know we are solving the proper problem and attaining the correct answer?
