ABSTRACT
I've been asked to tell you how Cornell attempted to explain to the rest of the world the establishment here of one of four national centers for advanced computing — the supercomputer facility formally known as the Center for Theory and Simulation in Science and Engineering and nicknamed the Theory Center. This is the center that was founded this spring with part of the 200 million dollars that the National Science Foundation is allocating in the federal government's supercomputer initiative. That's a polite way of saying, we want to “Pearl Harbor” the Japanese before they do it to us in yet another area of technology.
In brief, the Cornell Theory Center will be receiving something in the neighborhood of 30 million dollars from the National Science Foundation and another 30 million from IBM, in equipment and support, over the next three years to build and operate a production supercomputer facility — a sort of jumbo jet of supercomputers — and to conduct research in experimental supercomputer configurations — a program that could be thought of as the X-15 of computing. The Theory Center is still seeking additional industrial support; another $100 million would be a nice round number. Even without the industrial support, this is the largest single research program at Cornell.
I'd like to describe the preparation — the groundwork — that went into this public information effort, as much as two years in advance. We'll go into the gory detail of what went wrong in our announcement, and some of the things that went right, not so many thanks to us. We'll take a look at how the news media covered an event like this — in particular television news — and I'll tell you why the hardest part of our job, as public information practitioners, is still ahead us.
Let's start with who am I and what am I doing at a conference of computer documentation people? In a way, we're in the same business. We're supposed to be explaining computers and computing to people. Your people — your public — can be assumed to be receptive to computing. Or at least they're using it. The general public includes lots of people like myself who are still on the fringes of the computer revolution. They've been involved in a few skirmishes, maybe not even wounded yet, but they're not sure whose side they're on. They know that “user friendly” isn't good enough. They're not ready to learn a new language to speak to a machine. “The damn things are in the U.S. of A. Let them learn to speak American”.
Even the millions of folks who have bought personal computers share a healthy suspicion of computing. Big computers are the ones the IRS uses to lose your tax returns. Big computers are the ones in government weapons labs. Big computers still can't predict the weather. Big computers are the ones, when they make mistakes, you can't argue with.
That has something to do with why the idea of super computing — large scale computing — is not so easy to sell. “Bigger is better” went out with tailfins on cars. Now, if you're going to be bigger, or bigger just to be faster, there had better be a good reason for it.
We didn't realize all of this, however, when ken Wilson won the Nobel Prize for physics in 1982, and immediately began talking about building supercomputers. I guess we were just glad to hear a theoretical physicist talk about something besides “the deep and hitherto unperceived analogies between the phenomena revealed by phase transitions and certain aspects of elementary particle physics.”
The public's first inkling that there would be something called a Theory Center began when Ken Wilson stood up at a press conference that October morning, about four hours after being notified he was a Nobel Prize winner, and said: “I'm working at the national policy level to get people to realize the importance of computers as they become very much more powerful than they are today.” He said that just one field alone—theoretical physics — needed computing support to the tune of $100 million a year. He said, “I hope the prestige of the prize will help me get people — not necessarily to give $100 million — but to look carefully at the problems I've been discussing and to see if we can't get them worked out.”
And from that day on, Cornell began promoting Kenneth G. Wilson — and I'm not ashamed of that word, promotion — and capitalizing on his fame. After all, you're only the reigning Nobel Prize winner for 365 days, and then someone else's phone starts ringing off the hook.
Now Wilson was already serving on government panels to advise on the future of large scale computing and he had been knocking on the doors of executive suites in big business and industry, trying to convince the movers and shakers that American industry needed supercomputers and that the computer industry wasn't going to make very many of them until there was a demonstrated market and the best showroom, if you will, for supercomputers would be the universities where potential customers could come and “kick the tires” of the latest models.
Then suddenly, Nobel Prize winner Wilson was the most prominent member of those government panels. Receptionists would say, “Let me show you right in, Dr. Wilson.” We interviewed Wilson for a Cornell publication a month later and he said, “There is nothing that comes close to providing the kind of forum that the Nobel Prize provides. With the kinds of problems I'm dealing with, with the kinds of barriers I face, anything short of the Nobel Prize doesn't mean very much.”
We took the text of that interview — which talked about new uses for computer modeling and simulation and some schemes for parallel processing — and we sent it to about a dozen key business writers and science writers and editors around the country with a note saying, “Keep your eye on this guy. He knows more about supercomputing than anyone else in the country.”
Now we didn't know whether that was true at the time. We just sort of became convinced of it.
In the meantime, Ken Wilson was stepping up his activities in behalf of supercomputers. He was visiting more industries and getting more involved in advising government policy. When a report or a recommendation came out, if his name wasn't on it, people would ask his opinion. When the Japanese moved a little closer to making some big advance, people would ask Ken Wilson what he thought the U.S. should do. And once you get cited in The New York Times as a “leading expert” then you are one, and everyone else wants to know what you think. He was invited to write lots of articles and give talks on supercomputers and “the Japanese challenge.” He became “Mr. Supercomputer.”
We don't claim all the credit for his fame. A lot, maybe most, of the effort was on Wilson's part. We just did everything we could to keep him in the public eye. When he and IBM and some other industries and the National Bureau of Standards sponsored a conference on large scale computing in Washington, we promoted it, even though it had next to nothing to do with Cornell. He became one of about a dozen almost-celebrity professors at Cornell. The only person more quoted, day in and out, was Carl Sagan.
After Wilson had convinced the Washington establishment and the people holding the purse strings to spend some big money on scientific supercomputing, he had to step back from the role of neutral adviser and apply for some of the money himself. And somewhere along the line, the Theory Center became the Theory and Simulation Center, and it was to be for engineering and not just science. Must be someone figured out that there's a reason why the Fortune 500 doesn't include companies called International Business Theories or General Theoretical Motors. The co-investigators in the proposal to the National Science Foundation were Wilson; Dr. Kenneth King, who is also a physicist by background and is the computer czar for Cornell; and Ravi Sudan, also a physicist and an engineer who runs a lab for plasma fusion studies.
During the time the proposal to the federal government was being reviewed — for months — we couldn't say much about the Theory Center. It's considered bad form to discuss something you're certain you will get. And if you don't get it, you look really silly.
So instead, we concentrated on one little phase of the Theory Center, one that was already going on. This was the so-called GIBBS Project, an attempt by Wilson and some of the computer scientists here to create an entirely new scientific programming language to replace FORTRAN. We asked the public relations firm that represents the College of Engineering, of which computer science is a part, to push the GIBBS Project and they tried. It got some attention in the trade press and in places Like Science magazine, not too bad for something that wasn't hatched yet. The Theory Center, itself, wasn't real for a long time either. The Cornell faculty had given its consent and so had the University Board of Trustees. But Cornell's President, Frank Rhodes, wouldn't allow it to be established until Wilson could show some evidence of funding. They had an office with a name on the door and some furniture and a couple of people, but they didn't exist as far as Cornell University was concerned. We took to calling them the Theoretical Theory Center.
We also started planning how we would announce the center when it was funded, which everyone said it would be, except that was a secret. We started preparing with the National Science Foundation's public information people to make an announcement. They told us they were afraid of a leak, ahead of the official announcement, and it could come from Congress. We thought they meant congressmen from California or Illinois or someplace. Surely, no elected official from New York would engage in something as sleazy as pork barreling, then spill the beans. Remember, I told you something would go wrong…
In our brainstorming sessions, Ken Wilson made a demand that caused some snickers and mumblings of “Boy, is he naive.” He wanted to create the impression in the public mind that all of upstate New York was the next up-and-coming high technology region in the country. That all the isolated high tech areas like Rochester and Schenectady and Syracuse could be working together, rather than in competition, and that they could be linked electronically, by the Digital Thruway. The next Silicon Valley! And he wanted that impression created and established before the Theory Center was established, so that it would seem to be another piece fitting into the high tech picture. So we tossed around some names. Everything new has to have a catchy name. If New York City was the Big Apple, upstate could be the Silicon Apple. Then someone pointed out that gallium arsenide was the next hot semiconductor material and maybe we should be the Gallium Arsenide Apple. But that sounded too much like something the wicked witch would give Snow White. We talked about how the state could become involved. We tried to point out that impressions of prosperity and high tech environments aren't created overnight. Nobody knew they loved New York until millions of dollars worth of jingles and bumper stickers and billboards told them so.
I guess we sensed a few inferiority complexes showing in these men who were about to pull off an astounding achievement — to persuade the federal government and the biggest computer company in the world to risk tens of millions of dollars. I remember Ken King telling of a telephone call he had just received from an acquaintance at another university who said, “Congratulations, but you got the booby prize.” He meant that Cornell — although it hadn't been announced yet — would be the fourth last-minute center funded by the government, and that we had to team up with a company that didn't even make supercomputers to do it.
We tried to point out that Cornell didn't need to apologize for being the odd man in or out or wherever, because we had the element of surprise on our side. Everyone would want to know why the government was designating a private university in the middle of nowhere as a national center. We said that a couple of times, then shut up, We thought we still had three months to prepare for the announcement.
At one point, some thought was given to hiring the same public relations firm that represents the manufacturer of the array processors the Theory Center uses, Floating Point Systems, to represent Cornell as well. They talked a Lot about “building understanding” which is something that p.r. people are big on. “Building understanding” is p.r. shorthand for building understanding of my point of view and convincing you of it. The firm wanted $40,000 to make the announcement of the Theory Center, and that was just to the trade press. It occurred to us that for $40,000 we could parachute Ken Wilson to the roof of every one of the top 100 newspapers in the country to personally hand a news release to the editor. We told them we'd think about it. We thought we still had two months to prepare for an announcement, sometime in the middle of April.
In the meantime, we began preparing background information on the supercomputer center. We did a story saying that supercomputing will benefit American business, that “the advanced power of supercomputing and the research discoveries it makes possible promise to improve the entire corporate production cycle, from conception of a product through manufacturing to distribution.”
We did a story saying that the marriage of supercomputing and three-dimensional, real-time computer graphics would be the greatest advance in communication since cavemen started painting on walls.
We prepared a background piece saying that “Cornell University is a promising Location for a national, advanced scientific computing center because of its experience in operating highly successful interdisciplinary centers for the benefit of the scientific research community.” And we took the opportunity to brag about the Cornell Manufacturing Engineering and Productivity Program (COMEPP) and the Cornell High Energy Synchrotron Source (CHESS) and the Materials Science Center and the National Research and Resource Facility for Submicron Structures (which spells NRRFSS) and the Cornell Biotechnology Institute and the Semiconductor Research Corporation Center of Excellence in Microscience and Technology (which doesn't spell anything).
We did another story saying the research uses of the supercomputer will range from “the study of galaxies to subatomic particles, from the motion of drifting continents to the movement of toxic wastes.”
We wrote a general news release on the announcement, Leaving blanks for the amount of money and the number of years and the actual date of the announcement. We thought we still had a month to get ready.
We solicited statement of congratulations from New York Governor Mario Cuomo and from the congressional delegation from this part of the state and from IBM vice president Jack Keuhler.
When we were writing our news releases, by the way, we had to be careful not to mention IBM in the same breath — or even the same paragraph — with the word supercomputers. That directive came down from on high at IBM. IBM was not in the supercomputer business. Never had been, never will be. We couldn't even say the 3084QX would be a building block of a supercomputer. IBM was just giving us $30 million because they like us.
We prepared biographies of all key personnel involved — all the way from Cornell President Frink Rhodes, who doesn't know anything about computers but who is able, with a little prompting, to speak eloquently on any issue and thank people for giving us money — down through all the vice presidents of: the university and provosts to the principal investigators in the Theory Center grant to the people who will really run the facility.
Then we sat back and waited. Until February 20th, a Wednesday, when the NSF told us the announcement would be made the following Monday morning, in Washington. That didn't bother us. We were ready. We decided to schedule not one, not two, but three simultaneous press conferences. We would send Ken Wilson to the NSF press conference in Washington, along with a couple people from our office to straighten his tie and tuck in his shirt tails. We sent President Rhodes and Vice Provost King and Professor Ravi Sudan to New York for a press conference at Cornell Medical College. And we kept one Cornell vice president and one provost and the executive director of Theory Center, Bill Schrader, and the head of Theorynet, Alison Brown, for a simultaneous press conference in Ithaca.
When we announce a press conference, we are very cagey. We try not to give away very much of the story — just enough to entice people to turn out. There's a reason for this. If you give away the story and if it's worth anything, the news people — being in a very competitive business — will try to run with it, and spoil your announcement. It happens every time. So we said something Like: Cornell University, the National Science Foundation and a major manufacturer of computing equipment will announce the start of a $60 million cooperative research venture at 10:30 a.m. Monday, February 25, at the following locations: …. Then we swore everybody to secrecy, everybody who might get calls Late at night or even be likely to talk in their sleep. There was one exception to that: We lined up an interview with Ken Wilson and The New York Times for Monday morning, just preceding the scheduled announcement. Then we sat back and waited.
Imagine our surprise, on Saturday morning, February 23, to start receiving calls from news media all over New York State: “Could someone comment on the D'Amato announcement?” The D'Amato announcement? They read from a press release: “Senator Alfonse M. D'Amato (R, NY) is pleased to announce that Cornell University will receive at least $30 million and possibly up to $60 million from the National Science Foundation and up to $35 million from International Business Machines Corp. to do fundamental research from America's next generation of supercomputers.” Senator D'Amato said the grants would make Cornell one of the leading institutions in the country.
That surprised us a little. We thought we were already a Leading institution. We are among the top four or five or six research universities in the country, and our academic reputation isn't too shabby, either.
It was obvious what Sen. D'Amato was up to. He serves on one or two committees that some time in the past had reviewed the NSF proposals, probably voted for an appropriation when it seemed that some of the money might go to his home state, and placed a note in his future file: Break this on a slow news day and take credit for it.
We tried to persuade D'Amato and his staff to back off, to correct his misinformation, to join us in a joint announcement on Monday and to shut up in the meantime. No go. He said the NSF had given him the green light to make the announcement. The NSF was furious. They refused to confirm that Cornell would even get a nickel, they started an investigation to determine how the leak occurred, and they blamed Cornell for prompting Sen. D'Amato to jump the gun and spoil the announcement for the other three centers around the country. The congressmen in the House of Representatives who really had gone to bat for Cornell on this one weren't too happy either. But they knew that D'Amato had a reputation for this sort of thing.
Of course the news was in the Sunday papers all around New York the next day. That kind of money, even if the figures are incorrect, gets people's attention. The stories said that Cornell officials refused to comment, except to say that Sen. D'Amato was mistaken, and that an announcement would be made on Monday. It looked like we had something to hide. Or at least that we were caught off guard.
Which we weren't. We had even given Ken Wilson media training. We had put him in front of our own TV cameras. We had dry runs of press conferences. We asked him the toughest, the stupidest, the most repetitive questions we could think of. We tried to teach him to look at the camera, not to fidget or play with parts of his clothing, not to look to the heavens or into his pocket for answers. In short, to give the impression that every question he gets is the most perceptive and original he has ever heard, deserving of a sensitive, profound answer that he just thought of. We weren't trying to make a Carl Sagan of him, just to help him come across as the intelligent person that he really is.
Then we packed everybody off to their respective press conferences. All in all, they went pretty well, considering that practically nobody came to the New York City press conference and that the University of Illinois beat the pants off us in the Washington conference. The problem at the NSF press conference wag Larry Smarr, the Illinois astrophysicist, who was savvy enough to jump up and answer questions that weren't directed to anybody in particular. (We made a note to train Wilson to do that the next time.) Larry Smarr told a little story that any of the supercomputer people could have told but he thought of it first, and it's been quoted everywhere since then. He said that supercomputers are so scarce in this country that in order to do his research he had to go to Germany to find time on a machine — a machine that was made in this country. No big deal, but it was the kind of anecdote that writers love and lots of them used it in their stories.
We knew our homework was paying off, however, when we saw that Ken Wilson was quoted on the front page of The New York Times, even before they mentioned the other places that have supercomputer centers. Larry Smarr had a snappier quote, but it was in the second section of the paper, and besides, IBM was quoted as saying that Cornell's approach to building supercomputers was the only one IBM would consider exploring, not that IBM was particularly interested in making supercomputers, of course.
You have before you copies of some of the newspaper and magazine clips that appeared over the next few days. There have been lots more since. I'd like to play for you some tapes from local television stations that covered the Ithaca press conference. They go from bad to worse. We in the public information business collect these things to try to figure out how the information that we thought was presented so clearly gets so screwed up on the six o'clock news. Later I'll play a segment from a public affairs show on a Local PBS station, and then a piece that one of the national networks did — a really first rate job. Those of you who live in major media centers will be educated on what small town television is like. Those of you from small towns can sympathize.
Adobe Acrobat
QuickTime
Windows Media Player
Real Player
Pdf
C.5