Chapter 10

STEVE - FEB 61 back at the agency

When I got back to the agency, I first gave my report to Sam and Dr. Edwards. They were pleased with the progress that we had made, but realized that there were still a number of major issues to be resolved. I then told them, "Paul mentioned that there was a possibility of adding a small fast memory box. I told them that I was interested, and I was sure you would be also. I think that on the basis of the initial projections of a cycle time three times that of the main memory, it could speed up the streaming operations by at least a factor of two."

Needless to say, they both were quite excited about this. I then warned them that Paul had said that it would not be cheap, and that he specifically requested that Murray be at the next meeting in two weeks to address contractual issues.

Sam then told me, "Dr. Edwards has been pulled off for another very high priority project. He and I both think that you have the best understanding of the issues here, and while I want you to keep me updated on your progress, you are going to be the primary contact with IBM for at least the next six months."

I started to work through the details of the streaming process, in particular, the indexing needed for data references, and the conditionals. I also knew that we needed some sort of statistical data; since we would be looking for recurrent patterns, we needed to identify them as they went by.

I said, "OK, I'm getting pretty used to Poughkeepsie by now." Inwardly, I was elated that I would continue to be working with Sandy. I was starting to develop strong feelings about her and realized how often she was in my thoughts.

After running through the process in my mind several times, I realized that if we took our data and made several streaming runs through it, our job would be easier than if we tried to do everything in one pass. We could do one pass to search for known patterns, another to look for recurring patterns in the input, and yet another to do some manipulation of the input data and generate a partially decrypted version of the data. We could then repeat the process and after a few iterations, we would be getting close to a usable cleartext version of the data.

Steve, February 1961

Murray and I were back in Poughkeepsie two weeks after our last meeting. Paul had asked the IBM business people to sit in on the meeting, and they spent a great deal of time discussing contract terms with Murray. Finally, just before lunch, I got Murray off to the side and asked him if we could break into parallel sessions, so that we could get some technical work done. As we were about to break for lunch, Paul told us that while the group is large enough to lay out specifications for the Reaper box, to actually complete the design will take a much larger group.

Sandy and I had lunch together, and she told me, "Mom didn't forget - I have the recipe for her stew with me. We were both very happy that you were able to have dinner with us - please plan on a repeat engagement.""

I said, "Sandy, I'd like that."

The technical people reconvened for the afternoon meeting, although Paul was splitting his time between our group and the business discussions. It was pretty obvious that he was far more interested in what we were doing than the contractual issues.

Chad appeared that afternoon and presented his ideas for the enhanced tape system. We reviewed the design, which was based on Hypertape, but with the wider cartridge. The automatic cartridge changer - the jukebox - looked good, but there was no significant improvement in the data transfer rate beyond what resulted from the wider path. I told him, "That's not good enough. What we need here is a 'scheduled invention'.""

Chad and his manager started grumbling at this, but I continued, "Don't forget that this is very important to the country, and I'm sure that there will be some exceptionally good contract terms for IBM. Please see if you can come up with something better."

Stew and I had talked in one of the breaks, and I knew that he was another tinkerer. I wasn't surprised when he said, "Chad, let's get together on this tomorrow. I have some ideas about how to speed up the tape data transfer rate, both by the speed of the tape, and the character density - bits per inch - on the tape."