Review of "The Supermen: The Story of Seymour Cray and the Technical Wizards behind the Supercomputer"

The Rise and Fall of Midwest Computing Sans PLATO

This is a great book. It conveys much of the flavor of what it was like to be in the Midwest's computing culture in its heyday of the 60's through the 70's. What it failed to do was tell the real story of the Midwest's demise as computing leader of the world -- which isn't the story of Seymour's obsession with packaging rather than on-chip integration, as implied by this book. Rather it is the story of the failure to deploy the network revolution, now embodied in the Internet, to the mass-market 20 years early on Seymour's matured hardware via the PLATO networking project at Control Data Corporation. PLATO was a $1 billion (1970s) 'bet the company' investment by Bill Norris, the farmer/CEO of CDC who put a windmill pump from his Nebraska farm in front of CDC's corporate towers to remind people where they came from. That is the story of epic proportions only grazed on by this book. PLATO was ready to go to mass market, but Wall Street combined with classic middle mismanagement killed the mass market version of PLATO before it could even be test marketed -- for which it was ready. Had it gone otherwise, Seymour probably would never have left the Midwest, and his supercomputer architecture would have focused more on the directions taken by Sun and Hewlett Packard -- except with Seymour's inimitable qualities.

I personally worked with the PLATO project and tested a version of it that would have leased a network computer with Macintosh-like interface, including network service, for a flat rate of $40/month with capital payback in 3 years. It had everything -- email, conferencing, user-programmable electronic commerce, multiuser real-time graphics games not to mention thousands of hours of computer based education courseware for which the PLATO system was originally designed. We could get this performance because the culture surrounding the land grant colleges of the Midwest, such as the University of Illinois where PLATO originated, combined with Seymour's astounding performance levels created the right tradeoffs between hardware/software. Some of us were looking forward to incorporating Seymour's newly marketed Cray-1 as the foundation for the next generation of mass-market PLATO system. Initial benchmarks looked to provide an outstanding bang for the buck as an information utility hub -- even without some of the more obvious architectural optimizations that would help in this new kind of application of his systems. This would have shielded Seymour from the vagaries of the government-dominated supercomputer market and driven his architectures into higher levels of silicon integration faster. This, in turn possibly providing the kind of capital in the kind of organization that could have delivered on gallium arsenide's potential, unlike the disaster that occurred when Seymour left his farm and went cheek-to-cheek with the military in Colorado Springs, CO.

The Internet Explorer web browser was based on the NCSA Mosaic web browser developed at the University of Illinois -- right across the street from where PLATO was invented. This was no fluke. PLATO had a profound impact on the culture of the University of Illinois particularly its young students who wanted to push the envelope in networking. The NCSA also gave rise the most widely used web server, Apache, and the the founders of Netscape. The loss of possibly 20 years of 'new economy' is incalculable, but suffice to say, comparable losses have been suffered as the result of open war.

There are a lot of anecdotes this book doesn't tell that will probably die with the people who lived the tale. Just one, to capture a bit of what will be lost to history:

People looking for Cray Research's facility in the fields of Wisconsin could drive up to a farm house and ask where 'Cray Research' was located and a friendly neighbor would say, 'Oh, you mean Seymour's place...' and then give directions to an area surrounded by an almost invisible network of intelligence agency surveillance equipment -- protecting what was seen as a national treasure from potential espionage. In a speech to one of these agencies, Seymour told them they could come out and protect his folks but only if they never got in the way, and that meant not even letting anyone know they were around. Well, you could tell they were around, but at least they didn't get in the way!