A Shout Out to the Fifth Generation
Of course, the assumption is the answer is somewhere in the Cloud, right? We all know the Cloud has all the answers. But what if the Cloud doesn't have the answer?Does the Cloud know what it doesn't know? How long will it take the Cloud to figure out that it doesn't have the answer? How much horsepower should the Cloud use in deciding whether it should even look for the answer?
What if the Cloud determines that there are hundreds of thousands of potentially right answers, how much horsepower should be dedicated to finding which right answer is the right answer? We'll let the Internet be a stand-in for our Cloud for now. And we are not ready to give up on the question we've been talking about, because finding the answer to that question and many like it is saturated with possibilities for concurrency/parallelism and multicore computers. In fact, the skill set needed to exploit the parallelism in finding the answer to our question is precisely the skill set that will be most beneficial in exploiting multicore for application-side problem solving. With the proper paradigm shift in place and koolaid in hand, identifying the potentiality for parallelism in a computer program that could answer that question using the Cloud will be almost trivial.
But we are currently convinced that some of the ghosts of ICOT and the Fifth Generation project know the clues that will lead us to the necessary paradigm shift. How guilty is And/Or Parallelism? Is it one of the culprits? Or is it an old school idea trying to find new application? The voices say something about And/Or graphs and searching but its not quite clear. Yep, Tracey and I both hear voices, voices of the ghosts of ICOT and the Fifth Generation project. Some of loudest voices we hear at this very moment are from Kazuhiro Fuchi, Koichi Furukawa, Tohru Motooka, Ehud Shapiro, Akikazu Takeuchi, Kazunori Ueda, David H.D. Warren, Takashi Chikayama, Leon Sterling, Toshihiko Miyazaki and Vijay A. Sarawat, Robert Kowalski, Gopal Gupta, and Enrico Pontelli. They each give us a clue (some subtle, some not). Those clues bring us closer and closer to the evolution (or revolution?) needed to take true advantage of massive parallelism. It is often said you can't get where you're going if you don't know where you've been. Many of the men and women involved directly (some indirectly) with ICOT and the Fifth Generation project had much to say on parallelism and computers. Much of it was well before its time. Only now does the cost and accessibility of multicore computers put their research and work easily within the reach of anyone concerned. We are at a definite advantage here. Not only do the ghosts of ICOT tell us where we've been, they provide a substantial road map telling us where we are going.