Apollo Guidance Computer History Project

First conference

July 27, 2001 

Human-Machine Interface

RAMON ALONSO: Now I'd like to bring up a story about the interface, this business about that. We've heard it a million times, but ... 

At the time of the organization of the computer design was that Eldon was in charge of the design of the physical computer and there was another group, Jim Nevin's group, that was the human interfaces. And he had to make the rules as to how, if there were displays, how big they should be. If there were letters engraved in some kind of surface, how big they should be because they had to look through them through the visors. And if there buttons, how big they should be because they had to do it with gloves, and all kinds of things. But also, he was convinced that probably the thing to do is to have a CRT -- and remember, this is a long time ago, and that this somehow was going to be the case.

Now, in the meantime, the computer design is proceeding along, and as I said before, the computer design consisted of, here is a cubic foot. By the way, this tall skinny one is for the display, fill it. Over here there is this other space, fill it, and that's it.

So as I remember it, maybe it's not right, but we had the political imperative of having to show that this computer worked for all of the various firemen that came visiting on that. And there was a lot of well, we haven't decided yet what to do about the displays and this, that and the other.

And one time coming in to work they said you know, at least for the time being, why don't we do something simple. Three vectors, all of these engineers like three vectors. They split three rows of numbers, and those are the vectors. Aligned rocket, display time, fire rocket, recovery, and this thing. We proposed it to you and you said "Well, go ahead." You said it's not going to do any harm. (Laughter)

ELDON HALL: I think you need to step back one step before that. If you remember when we had the meeting before, what is it called, for the support contractors, the information for them to bid on down in Houston. At that point we were proposing analog kind of junk; analog meters, and all of this kind of stuff. Because I even got into a little debate with Dave Hoag here. He probably has forgotten it. I talked about time. Dave said you had to have time like that. And I asked him, well, how do you write these down on a piece of paper? You write it down like this. He doesn't remember it now. I don't even know what you're saying now. You write time down like in a digital format.

DAVE HOAG: Yes.

ELDON HALL: So you were pushing analog types of displays, and buttons and stuff, and so were all of the astronauts, and so was everybody else. Now go onto your story.

RAMON ALONSO: Now, the thing is, how do you fill these cavities with displays, and what shall it be? Well, Al Hopkins and I debated this for a while, and we had heard about these things called latching relays, which required no power to stay on. Sounded good. And we also heard about these things called liquid crystal displays, which also sounded good. (See Hugh Blair-Smith's annotations) The original seven digit, seven segment display for calculators came out of this. That's the one thing you can point to that came out of the Apollo project.

So over about a year and a half, some displays were put together. They were kind of homegrown programs that could demonstrate that this computer was doing something, and display time was an obvious one. I forget, there were two or three others.

So we put this and we could show that the computer really worked with this thing, and visiting firemen would come. After a while, we began to get a very interesting pattern of questions.

Is this verb and noun stuff going to stay for the flights? Well no, there is another group that's doing that. But time goes on, and it's not so clear that this isn't going to stay for the flight. And the questions got to be more and more is this verb or noun going to be there? So we began to ask why. Fascinating reasons. It's not serious enough. It's not serious enough. It's not military enough. It's not scientific enough.

So we began to say okay, we can put something else, but we have got these strictures on how big the letters should be, and all of that. What would you put there where it says verb and noun? Verb you might squeeze action in there, but object of the action is kind of long. You can't fit that in there. So people, of course, could be frustrated, and we being young -- especially Herb Thaler and I -- 

So we got to be really, really good at this game, and two things are worth noting. The last guy who came up is, I'm going to tell an out of school story -- it was Dick Battin. And he looked at this thing and he said "You can't have that." We said, "Why not?" We baited him, you know? He said you can't. What would you propose, and all of that. And his final, final, final line of defense was because the astronauts wouldn't understand it. (Laughter)

Now, the end of the story is: the astronaut Dave Scott spoke at the Computer Museum and talked about the Apollo guidance computer. And in there, one of the final things he said was and I don't know who thought of the verb and noun, but that was the greatest thing because we really could tell what was going on, and it was simple.

JOHN TYLKO: He also made the comment a year ago at that dinner event that he was surprised that more modern computers didn't use that type of architecture. He said that.

RAMON ALONSO: Well, it was again, right place, right time. I mean, this served the function, was in there first, wasn't any good reason to get rid of it. There really wasn't anything better to change it with, so that was one of the reasons it went there, and it was economical.

__: And everything else would have been bigger. It wouldn't have fit in that little space.

RAMON ALONSO: Exactly. It wouldn't have -- The buttons have to be way big and all of that.

FRED MARTIN: I was amused at the story about a year ago -- It's not a funny story, but when they lost the Mars mission on -- I can't remember what the name of that mission was -- but they eventually figured out that somebody had screwed up inches and centimeters, or meters and feet and so on.

I was amused by it only because we had so many conversion constants in that Apollo computers, we had so many ways of presenting things. And the astronauts, if they wanted it in feet per second on this one, but they wanted it in kilometers on this one, we gave them whatever they wanted, more or less. And the number of conversions in the computer were a lot. It was all programmed with fractional arithmetic and it all had to be scaled, and so on. I can't remember, it may have been, but I can't remember that we ever experienced a single error, computational error, because of this myriad of conversions between metric and English units in the system, and there were a lot. Everybody in the lab who was doing algorithms and calculating things and so on, more or less all used metric systems, in our calculations, in the way we derived things -- maybe not all -- but that was my experience. I remember dealing with a force in the equation, and all of that was in newtons and so on. But then you had to come out to display, and you had to come out to display what people wanted. If they dealt in miles an hour, that's what they wanted, or nautical miles an hour, or feet, or feet per second, or this or that, that's what you came out. Some of the angular things were in radians, some of them were in degrees.

DAVID MINDELL: Different astronauts preferred different units?

FRED MARTIN: Well, they couldn't have that. I'm sure they did, but in the end they got together. All I'm saying is that we, on the programming side of it, we sort of did not necessarily have the discipline to say that all of this is going to be one way. It wasn't one way. There was a lot going on in conversions.

JOE GAVIN: I have to say something in defense of the astronauts at this point. The pilots went through two revolutions in the way they thought about flying machines. The idea of having commonly accepted flight characteristics that were suitable didn't really gel until after World War II. Some of the airplanes that were flown in World War II would be absolutely unacceptable in 1955. So there was an effort to standardize on what characterized suitable flying characteristics. That was the first thing.

The second thing came along when we all went to irreversible servo-controls; where there was no direct link between the pilots and the surface that was doing the controlling. That led to artificial field systems. And that engendered a whole new set of considerations as to what is satisfactory.

I was involved in one of those early airplanes, and it was amazing, the difference between professional and test pilots as to what constituted a satisfactory field system. And in fact, in one airplane, we cut special cams for each of four pilots and got their -- This is what I like. Then we tried to sell the others on those. So there is a wide difference between people that has to do with your natural reaction time, your background, your experience. It's amazing to me that we do as well as we do, when you consider how variable people are.

DAVE HOAG: In that context, Mercury astronauts visited the lab very early on. I sat with them and said look, you're starting a whole new realm. Suppose we start right the very beginning metric. You're learning in an environment where the sizes are different anyway, so if you stayed with metric -- And they essentially agreed with me.

So then I went and talked to my engineers about this. No, we can't do that. We can't do that. And we had to move on. (Laughter) I couldn't force that issue. That's the other side of the coin. The astronauts, at least the Mercury 7 were willing, but the design team was not.

DAN LICKLY: But for every flight phase or thing, there is a display, three numbers up there. In fact, Fred may have examples over there. You would show what you had proposed as the main three things you wanted to display; velocity, altitude rate, something that you thought were the three significant. And that went through a lot of people in Houston, but it ended up most of all in the astronaut office. What do you want to see now there, here, and here? They only had three choices but they could change the units.

But once they signed off for it, that was it for the flights. You couldn't personalize it from one flight to another. It was during this phase here is what you saw in the displays. And there were a lot of arguments about what they'd want. Sometimes they wanted what seemed like strange numbers or strange units.

DAVID MINDELL: How about the equivalent on the control side? I mean you mentioned the question about limits and software checking errors.

MARGARET HAMILTON: Yes. I remember one very memorable time when we'd come up with a solution for how to warn the astronaut that bad things were happening.

So I went up to the hardware people. I know Albert Hopkins was in that meeting. Maybe you (pointing to Dan) were there too, I don't remember who all was at that meeting. But I told him that our solution was to have this display come up when there was an emergency, and basically knock everything out and give them a warning. 

He said "That's impossible" and I said, "Why?" And he said "Because the computer is not on all of the time. The display is not on all of the time." I said, "Why can't we just leave it on?" And all of a sudden it was very quiet and he said "Let me think about it." He came back and he said "We're going to leave the computer on."

So now we were able to come up with a high priority display to warn the astronaut, but it was a major decision that was made just sort of between us, the hardware and the software group at Draper.

DAVID MINDELL: To what degree, this is something that I know is still debated in aircraft flight controls, to what degree would the computer allow the pilot to do something that was dangerous?

FRED MARTIN: I think to a very great degree. I mean, the story I heard when I first was working on Apollo which fits in line with some of the things that were mentioned here was that it was a silly thing to say, but the astronaut could kill himself with a pencil if he wanted to. 

Because the burden on the software of putting in checks on -- There were so many things you'd have to put checks on, and this was a limited computer with only so much space, that it was almost a policy decision to say that we're going to eliminate -- 
I mean, unless the thing was so obvious or so easy to do, we weren't going to get into it, because you'd be into it forever. So it was left pretty wide open, and he could get himself in a lot of trouble.

CLINE FRASIER: The choice for the astronauts came down, not initially, but when the machine filled up, to functionality or things like warnings and error checking and things like that. So in general, the point of view was we'll take care of that. We want the functionality.

Hugh Blair-Smith adds:

Ray Alonso speaks of liquid crystal displays but I’m sure he meant to say electroluminescent displays. LCDs weren’t ready for prime time for years afterwards. The importance of latching relays was that EL displays required a considerable voltage. The segments in LCD displays (had they been available) could have been held with latches built from NOR gates. I don’t remember whether the latching relays were later replaced with SCRs. One of the engineers working on the EL displays was a totally cool character named Larry Baxter, who came to us originally as an on-site field service engineer attached to the Honeywell 800 (later 1800) mainframe. We could see that a brilliant intellect was being wasted on a non-brilliant job, and recruited him. Aside from being a sharp EE, he was a passable poet and a virtuoso trumpeter; he played the Purcell Voluntary at my wedding.

Astronauts' role in design

site last updated 12-08-2002 by Alexander Brown