Get Real! (in praise of realtime)

I grew up believing that the Wright Bros. invented the airplane almost by dumb luck. Nothing could be further from the truth. When I'd heard that they were bicylcists, I pictured the singing speed boat racer played by Elvis in a bad 60's movie. But the Wright Bros. were able to invent the airplane in part because they were bicylcists, not despite it. Truth is, they didn't even invent the airplane, but rather refined earlier designs. They understood aerodynamics and they invented the first wind tunnel. (The wind tunnel was smaller than a breadbox, but a modern version at M.I.T. still bears their name.)



So what the hell does this have to do with realtime responsiveness and P2P networks? The wind tunnel revolutionized airfoil development because it shortened the design cycle. It allowed them to test models of airfoils much more quickly and cheaply than their competitors.



Anything that shortens development cycles can adapt quickly to changing needs. This phenomemon drives biological evolution, and it drove the rise of the PC over the mainframe, the popularity of the electronic spreadsheet, and the growth of the Web itself. But the immediate feedback enabled by a wind tunnel or an electronic spreadsheet is of little use unless errors are corrected in future iterations.



Whereas the Wright Bros. optimized their wing design to generate lift, they had to fly into a stiff headwind to generate enough relative airspeed (which is why they chose Kitty Hawk). Gusts of wind would quickly tip the plane, causing it to crash. Their real innovation was error correction! They added wires and pulleys that could make the wings flex--the forerunner of modern flaps and ailerons--allowing them to control the plane's attitude.



There are three markers in a field at Kitty Hawk that demarcate the lengths of the first three successful flights. The first flight was less than a hundred feet, the second nearly twice as far, and the third one nearly twice as far again. There is no fourth marker because the fourth flight was 15 miles long. Realtime feedback and error correction enabled improvements of several orders of magnitude. As the park ranger said, the markers at Kitty Hawk don't show the evolution of the airplane, they show man learning to fly.



I claim that P2P networks are valuable in part because they are relatively responsive. You can request an asset at the network's edge without waiting for a central server. P2P networks have decent error correction insofar as the network reconfigures itself to deliver assets efficiently.



But we're still not taking full advantage of realtime feedback and error correction. In another blog entry, I'll discuss how the Web can take better advantage of these to enhance its value and effectiveness.



### Get your Daily Bruce! ###

Get Real!