Apollo Guidance Computer Activities

AGC: Conference 4 - Failures

Apollo Guidance Computer History Project

Fourth conference

September 6, 2002

Failures

FRED MARTIN Of course, not all of our projects were rousing successes. For example, Dan and I negotiated an absolutely sweetheart deal with IBM to produce some sort of compiler for the PL/1 box that was exceedingly difficult to do, only we didn't realize how difficult it was. And then we signed a Time and materials contract with them. So every hour you spend they pay for the hour. It's not a fixed price, but at some outrageous rates.

DAN LICKLY: We'd been tired of paying huge rates for IBM equipment, so we said we'll flip it around.

FRED MARTIN We set to work on this compiler and some of the people who were brought into this compiler were very theoretical types and decided they were going to do this compiler in some optimum fashion and so on. Well, it dragged on and dragged on and dragged on, meanwhile we were collecting--the meter was running like crazy. We made a lot of money on that contract.

DAN LICKLY: We had a whole Russian contingent working on that.

FRED MARTIN But in the end, in the end, they pulled the plug, because they weren't getting the product.

ALEXANDER BROWN: That's interesting. Correct me, if I'm wrong. So from 1972 you had a big NASA contract, developing the language, which became Hal, was this named after Hal Laning?

FRED MARTIN Yes.

ALEXANDER BROWN: And then from about 19--you mentioned about 1975 doing some work on what became ADA.

DAN LICKLY: Slightly later. We went on the ADA competition, around 1978 ADA had a very elaborate competition. They tried to have a bake-off with a number of companies. We were on the two finals. It was us and a French company and two other companies, Stanford Research I think. And everybody on that thing was pretty smart, so they eventually down selected to two and then they down selected again to this one. And that became the winner of this competition, but by then everybody's language--they had forced people together, so that the ADA language now was not very different at all in the contenders. It had certain principles that they wanted.

And then we set about building some ADA compilers, which turned out to be in some cases disastrous. Fixed-price contracts and things with the Air Force that didn't work out and--so it was a fair amount of money that was lost in the years of trying to build ADA compilers.

I don't know--in the final analysis--after Tucker got involved in it, if you add up all the ADA money, whether we made money or not. I have no idea, but some of those were very difficult and they were very trying times. People tell me that '95 was just a terrific thing, really coherent, excellent language, very useful, but I don't think anyone is ever going to use it except demos.

ALEXANDER BROWN: By that part of the '70's, in 1978, how big a part of Intermetrics' business was the language design, compiler design?

DAN LICKLY: We were a good sized group.

JOHN MILLER I would say it was probably maybe 25 percent of the business of the company.

DAN LICKLY: We had some huge groups, Pennsylvania, California.

JOHN MILLER We had people working on--we had about 13 other offices. We were doing a lot of work in things like software verification. We did a lot of work in the signal processing of acoustics and in anti-submarine warfare. The compiler operations of the company stood out in the industry very very well.

FRED MARTIN But you couldn't ride that horse forever. And it became very automated, the technology sort of passed it by into other areas. And so people were not going to build these compilers anymore, they were either going to use something from Microsoft or use something else ... or UNIX.

DAN LICKLY: Maybe the next go around with the government will be JAVA based.

FRED MARTIN But we do still have a group that's in Houston. I don't know how many people there are, maybe eight or something like that, that are keepers of the flame. I mean they keep that stuff for NASA and every shuttle flight there's something that happens that they want these people on there. So that still goes on, which is kind of remarkable, but it is, it still goes on.

ALEXANDER BROWN: So you were picking up government contracts--were you largely doing contracts for the government or were you sort of sub contracting to defense firms?

JOHN MILLER I'd say the dominant amount of contracting was directly with the government, not always, but it tended to dominate that.

ALEXANDER BROWN: And this continued through the '70's and the '80's?

JOHN MILLER: Throughout this. We built a compiler in the late '70's for the Galileo project. You think about that, the thing is still cruising around out there and every once in a while sends back a message. And you say, "Wow! You built that 30 years ago."

Now and then - reflections on careers


site last updated 02-01-2003 by Alexander Brown