Thursday, September 7, 2017

Summer 1981, When I advised NASA on their computing infrastructure

IMGP4383
Something I painted when I was, oh, nine or ten. See that meteor streaking at the upper right? My teacher, Glenn Brougher painted that. How do I know? The brush strokes. He had a deft hand. I didn't, not at that age.
It wasn't just me, of course. There was a bunch of us. It happened like this...

Sometime in the spring of 1981 I learned that NASA was offering 10-week summer fellowships for college and university faculty. I was at RPI, in Troy, New York, at the time. Most of the fellowships were to pursue research in some area of interest to NASA; you’d travel to of NASA’s centers and work with their people and their facilities. As I studied literature and cognition those fellowships weren’t relevant to me. NASA did not have a center for the study of extraterrestrial myths, tales, and epics.

But one fellowship was different. It was sponsored by the Goddard Spaceflight Center in conjunction with the University of Maryland and had an intriguing title: Computer Science: Key to a Space Program Renaissance. Man, thought I to myself, I want a piece of that! The idea was to study NASA’s computer infrastructure and make recommendations for improving it. That interested me and, having done a bit of work on an Air Force contract not so long ago (see this recent post, Metagram Software - A New Perspective on the Art of Computation) I figured I had a relevant credential or two – there was my work on knowledge representation as well. So I applied and was accepted.

The University of Maryland had a conference center high on a cliff overlooking the mouth of the Susquehanna River about an hour northeast of Baltimore. I arranged to spend my first week at the conference center and then the other nine weeks in Baltimore, near the Homewood Campus of The Johns Hopkins University, where I’d gone to school. I arranged to get some kind of visiting appointment at Hopkins so I’d be eligible for university housing.

Come the first week of June 1981 I packed my car, including my new NorthStar Horizon computer (I wrote about it here) and headed to Maryland.

The conference center was at an estate that had once belonged to an industrialist named Donaldson Brown. Large rooms, oriental carpets on the floor, decent food, wonderful setting. The stables had been converted into bedrooms and I booked a room there for the first week. Most of the fellowship holders stayed there the whole time.

This was going to be interesting, perhaps even fun.

* * * * *

Full-time study personnel consisted of 19 faculty fellows, a handful of NASA administrators, and eight support staff. In addition to that there was a whole passel of NASA people who showed up for varying periods of time, some for a day or a couple of days, some for a couple of weeks or more. They were from NASA centers across the country and represented all functions, space science, weather and atmospheric science, satellite design and fabrication, spaceflight (vehicle design, fabrication, launch, and tracking), and so forth. NASA brought in academic and industry experts from all aspects of computing who came by for a morning or an afternoon, perhaps a day, to lecture and talk. We also had visits to various to Goddard Spaceflight Center in Maryland and NASA Headquarters in Washington, D.C.

There was A LOT going on. A lot of balls to keep in the air.

To what end?

The previous year a small team of academics, including Marvin Minsky and Carl Sagan, had reviewed NASA’s computing facilities and issued a report, bound in red covers, Machine Intelligence and Robotics: Report of the NASA Study Group. They declared NASA’s computing infrastructure to be below standard. NASA, the folks who landed humans on the moon and brought them back, a computing backwater.

There is more. NASA had been formed for that moon mission; everything else was dross, or gravy, depending on your POV. Along the way NASA had also developed scientific missions – tracking earth’s weather and geography, deep space probes, telescopes – but the moon shot was the heart and soul of the organization. That was over. The Apollo program had ended in 1972. It was a fabulous success of course, but what next?

What next was the Space Shuttle program, "Space Transportation, Inc." That wasn’t so sexy. Important, but it lacked heart, Nietzschean levity.

Consequently, NASA had a morale problem. Some of their best people were leaving. What to do?

The people who’d organized this – Stan Sadin at NASA headquarters, Paul Schneck at Goddard Spaceflight Center, and Raymond Yeh from the University of Maryland – figured that perhaps if NASA were to inject some top-flight computing technology into its programs, including AI, that might help alleviate the morale problem.

That impressed me, it impressed me a lot, that NASA officials had arranged this, really, to address the organization’s morale problem. How intangible. How vital.

And that was our charge, to give these managers the documents they needed to set about improving NASA morale by upgrading NASA’s computing infrastructure. Of course, they never said that in so many words, but it’s clear that that’s what was afoot. That’s what consultants are hired to do, give the people who hire them the ammunition they need to make changes they already more or less know they have to – or, at any rate, want to – make. Sure, we were brought in as faculty fellows to do research, hands across the sea, university faculty working hand-in-glove with NASA researchers advancing human knowledge, that was the pretext. But the mission was straight-up consulting. We all knew it. That’s why we were there.

* * * * *

How to get it done? There were 19 of us faculty fellows. We didn’t know one another, though some of us had spent a summer with NASA before. And most of us, I warrant, had little or no experience as management consultants. How could we possibly produce a credible report in only ten weeks, a report about and for an organization that had a 1981 budget of 5.5 billion 1981 dollars (equivalent to 14.5 billion 2014 dollars) and I don’t know how many employees. It’s currently over 17,000 (plus a lot of contractors); I don’t know what it was then.

We had ten weeks, from early June to mid-August. In that time we had to learn about NASA, come up to speed on computing technology, organize ourselves, think and deliberate, and produce a report. Yesh!

We had a lot of help. NASA knew what they were doing.

There’s no point in even trying to give a day-by-day or week-by-week account of what happened, even month-by-month would be a stretch, and it was only two and a half months. That summer was 36 years ago and I wasn’t taking notes. All I remember is this and that, here and there.

I know that we spent the first two or three, maybe four, weeks just drinking in the information NASA showered down on us. Somewhere in there we divided into teams, but I don’t know how we did it. There must have been a meeting where NASA said something like, Hey, guys, dontcha’ think you oughta’ break into teams? And so we did. What teams? I was on the information sciences team (I think it was). Don’t remember what the others were called, there were two, maybe three others.

Somewhere in there NASA brought in a guy to lead us in team-building exercises. I’m inherently skeptical about such things. This guy was good. Think his name was Richard Horworth.

I know we spent a day at Goddard Spaceflight Center. Saw a big whirly thing – not a technical term (centrifuge?) – where an astronaut would get into a capsule at the end of a long arm and then the arm would go ‘round and ‘round and ‘round thus giving the astronaut experience of high G forces. We saw a huge clean room where technicians were working on a mock-up of the space station, though maybe it was the real thing, or a bunch of spare parts? It was cool.

Every once in awhile I’d talk away from the conference center and down to the edge of the cliff where I could see the Susquehanna two-hundred feet down. That was cool too. As were the large rooms with those pedestrian tables sitting on those large oriental carpets. Were those carpets from the days when rich people lived here? You come in in the morning, find yourself a chair at the table, and get to work. I had some nice talks with a NASA engineer named Dave Callendar, from JPL. Was he the one who had a copy of Ted Nelson’s Literary Machines: The report on, and of, Project Xanadu concerning word processing, electronic publishing, hypertext, thinkertoys, tomorrow's intellectual revolution, and certain other topics including knowledge, education and freedom, a slender volume covered in silver Mylar? I forget.

And then there’s Bob Freitas, space nut enthusiast. He’d been with NASA the previous summer; worked on a fat report, Advanced Automation for Space Missions (PDF). He had a particular interest in the section entitled “Replicating Systems Concepts: Self-Replicating Lunar Factory and Demonstration” (PDF). The idea is that you haul a bunch of equipment to the moon and assemble it into a factory – or does it assemble itself? I forget. Anyhow, it starts producing things and stuff, including the parts needed to replicate itself. When those parts are ready, the first factory assembles them into another self-replicating factory (or does it assemble itself? I forget), and off it goes.

It was a wonderful idea. Sort of a high-tech version of free beer. That first factory, the one we’ve got to haul up to the moon on the backs of high-tech oxen, that costs an arm and a leg. But all the rest are free – at least free of terrestrial cost – because they’re assembled out of stuff (freely) available on the moon. If you google “Self-Replicating Lunar Factory” you’ll see that the idea’s still kicking around.

Crazy? I guess. Vernor Vinge hadn’t yet announced The Technological Singularity (that was 1993), but this certainly had that feel about it. And as you may know, I’m highly skeptical about the technological singularity (translation: I think it’s nuts). But it was nice to know that NASA had that kind of vision about it. They hired science fiction writers as consultants too.

Such “blue sky” concepts weren’t our charge, though we could maybe include them just a little bit. We had to deal with things like the length of NASA’s procurement cycle, which was approaching the length of a generation of computer tech. That means that by the time a request had been approved, the approved technology was all but obsolete. And then there was the peculiar fact that government personnel policy had a slot for mathematicians and one for electronic engineers, but no slots for computer scientists or programmers. So how could NASA hire the people it needed?

That was our playing field, somewhere between getting a reasonable procurement policy and shipping a recursive factory system to the moon. And we got it done. We produced an actual report with actual recommendations.

The final week was a whirlwind. A guy named Fred Buoni, also on the information sciences team, took charge and organized the process. No one asked him to or appointed him. And he didn’t make a big deal of it. He just did it and we all were glad of it. The job had to be done and he stepped up. He was a retired Air Force colonel.

Me, I wrote. I wrote my own appendix for the report (An Executive Guide to the Computer Age), most of the faculty fellows did this, as did many of the NASA participants. But I also helped draft key sections of the main report, the sections that coughed up actual recommendations [1].

With the report finished, we went to NASA headquarters and gave a final presentation before NASA brass. Yes, it was a big deal. I practiced my part of the presentation before a mirror. And I wore a suit for the presentation itself.

Then it was over.

Back home to Troy.

green box.jpg
Not the Tardis, but just as real

[1] The final report has two volumes: R. A. Freitas, Jr. and P. A. Carlson, eds. Computer science, key to a space program renaissance final report of the 1981 NASA/ASEE Summer Study on the Use of Computer Science and Technology in NASA. Volume I: Institutional Needs, Status, Plans and Recommendations. NASA-CR-175294. Volume II: Appendices. NASA-CR-175295. While I’ve downloaded both volumes from the web, I can’t locate a URL for the first volume. You can download that second volume here, https://ntrs.nasa.gov/search.jsp?R=19850025389

No comments:

Post a Comment