Chapter 43

STEVE — Monday, March 19, 1962

I called the movers as soon as I got in to work today, and they arranged to deliver Sandy's boxes and furniture on Wednesday afternoon. I spent the rest of the morning and another hour after lunch studying the two files that Dr. Edwards had given me last week. The first one, the effectiveness estimate for a weapons system looked like something could be done using the approach I had used last year on my first assignment, and I knew that I could outline the solution in a few days. The other, the onboard computer system for the missile, was going to be more of a challenge, one that I was eager to tackle.

I went into the tape library, where Samantha was back at work, having finally turned over the reaper run coordination to another librarian. Vlad had told me that it was still running quite smoothly, and that he and Lou were becoming quite comfortable with the programming. They had even fixed a bug in my code. Just as I was about to knock on Dr. Edwards door, it opened and a young man emerged, looking quite confused, carrying a couple of CDC manuals. I noticed that he stopped and talked to Sam for a few minutes as I knocked on the still open door to the office. Dr. Edwards called out to me, "Steve — come on in and shut the door." When I had done that and sat down, he continued, "How do you see us proceeding on the two projects?"

"The first, the effectiveness estimate, has a lot in common with the project I did early last year, and I'm confident that the Monte Carlo method will work here also. I need a few days to study it in detail, and by the end of that time, I will be ready to write a specification for the program. If there's someone else who could actually write and debug the program, I could work with them and make sure they were on the right track."

"What about the missile computer?"

"That's a bit more complex, and is going to take a bit more time. It looks like a very interesting and challenging job, and I'd like to work on it. The computer side of it is fairly straightforward, but there's a lot of definition and high level design work that needs to be done on the interfaces. As you know, I'm getting married in less than three weeks, and will be out for a couple of weeks for that. I'll be able to make a start on both these projects before then, but wouldn't be able to complete either of them even if I didn't touch the other."

He grumbled, "Inconvenient, but I guess I can't complain. Sam is very happy with the work you've done with Reaper, and Murray told me that the way you got the new programmers up to speed was excellent. Miss Ishido also told me that it was a pleasure to work with you."

"She was very helpful, and did a good job training her successor."

"At any rate, go ahead as you suggested. I'll have someone assigned to you for the weapons system job by the end of the week or Monday at the latest. It certainly won't be Mr. Rossi, the fellow who just left. He doesn't know anything about programming yet. Let's see, your last day is Thursday, April 5th?"

"That's right." I had a warm happy feeling when I realized that three weeks from today Sandy and I would be married, enjoying our honeymoon in Arizona.

"I hope that you'll have a progress report on the missile computer by then."

"I'll definitely have something, even if it's not complete."

He rose, indicating that we were done. "See me every few days, then, and tell me how you are coming along. I may only have a few minutes for each meeting, but I want to know how you are progressing, so that I can report to the customers."

I went back to my office. I took out both files and a pad of paper, and laid out a work plan for the two projects. It was clear that I would have to finish the programming specifications first, but I also wanted to look at the missile job long enough to see what kinds of interfacing would be involved so that I could review some of my logic and circuit design notes and textbooks. I remembered about the movers coming on Wednesday, so decided to take a quick look at the missile computer requirements so that I could spend the time I had to be at home for the movers reviewing my course materials from MIT.

SANDY — Thursday, March 22, 1962

I had spent yesterday in another all-day NPL architecture meeting that wasn't really very productive. For one thing, the program manager was out for the day, and the person running the meeting didn't have good control of the group; there was a lot of shouting and interrupting. There were some major issues about subroutine linkage conventions, program status indicators and especially levels of privilege; would there be a problem program mode and a supervisory mode, or would any program be able to execute any instruction. Since there were serious plans to facilitate multiprogramming in the architecture, there were strong reservations about this, but others had strong objections to having more than one mode.

The programmers had also raised the issue about program hangs, and there was a memo floating around that suggested that if the machine were operating correctly, no instruction should be able to hang or lockup the machine. Among other things, this implied that there would be no STOP or HALT instruction. By the time we were done, it looked like while we had previously reached some agreements on instruction formats, much of the rest of the system architecture was yet to be finalized, and if this didn't happen within the next few months, design schedules would start to be seriously impacted.

When I got home after the meeting, I poured myself a glass of wine and picked up the phone to call Steve. When he realized that I was on the line, he said, "The movers brought all the stuff today. I haven't had a chance to do anything with it other than push the boxes in the corner of the living room."

"That's fine; we can unpack them together next month. If you have a chance, open up the wardrobe boxes and hang the clothes in your closet, if there's room."

"There's a lot of room, dearest. I only have a few suits and stuff. How's everything going at work?"

I groaned. "There was another one of the all-day meetings today. I swear, sometimes it seems as if we are losing ground."

"It sounds like you are discouraged."

"Yes, and so are a lot of the other people. Too many loose ends, and there are several problems without a clear solution. The problem is that there are conflicting proposals that are imperfect, but the proponents are almost fanatical about them. There's going to have to be some head knocking and some major compromises; I'm sure I'll be done here before most of them get resolved."

"Do what you can, my love. I'm sure you'll make some significant contributions in the next two weeks."

"I hope so, Tigger. Just talking to you helps. Mom is calling me to come help with dinner; I'll talk to you again in a few days."

When I got in to work today, Paul told me that the NPL program manager had called a special meeting. He had requested the attendance of each of the system managers along with one or two of their key architecture experts. He told me, "I don't know how I'm going to manage in another two weeks when you are gone, but right now, I need you in this meeting."

When we got to the meeting, I was surprised to see that most of the usual mob that had been at yesterdays meeting were missing, but that most of the key technical people were present. The system manager called the meeting to order, and started out, "I'm not happy with the lack of progress on the architecture in the past few weeks, and we're going to set up some task forces to move things forward faster. They will be running nearly full-time until we get agreement on all the major issues. To start, I'm firmly convinced that we need a problem program mode, with limited privileges, and a supervisory mode, with unlimited privileges. I also think that any program exceptions, including but not limited to a problem program executing a privileged instruction, or any program executing an undefined instruction, or many other exceptions, should cause an interrupt rather than a hang."

He paused, and went on, "What else do you see as major issues?"

Several people mentioned the input/output system, and he said, "I'm aware of that, but there's another working group making good progress. We just need to make sure that what they do is well coordinated with the interrupt mechanism. Anything else?"

I said, "I have a big concern. You are suggesting independent work groups to work on several problems that are inter-related. That's good, but someone needs to make sure that they are coordinated so that they don't introduce inconsistent requirements on the work of the other groups."

"Good point, Sandy. As I remember, we're losing you in another two weeks. Would you be willing, for that time, to sit down with each of the groups on a daily basis, and make sure that they aren't on any collision courses? I'd like to hear from you every few days, and if there's any major conflict that you can't resolve, please get back to me immediately."