The final Fortran compiler project was finished and shipped and I was promoted and asked to join the Systems Planning department. There were about 6 of us, one other was Bob Freiburghouse - the designer/leader of the compiler that we had just built.
Our job was to propose new projects - Bob and I were working on plans for a new operating system to replace the mess that was running on what was now named the Series 200. These machines were attaining great marketing success. They were being sold to customers of the IBM 1401/1410, a machine aimed for small businesses. There were two programs called Liberator on the H200: one translated IBM Autocoder (assembly language) programs into H200 Easycoder and the other translated executable programs. The translation of executables was important because many customers had lost their source code. The 1401/1410 and the Series 200 were similar in that they were variable word-length character oriented machines but all the details were different. Emulation which is what is used these days was out of the question since the machine wasn't fast enough.
At the time IBM was promoting a path to the 360 for 1401 users, but the early versions of the 360 were somewhat shaky and the H200 was about 2-3 times as fast as the 1401; thus with Liberator, Honeywell offered a good upgrade path. Now, while planning new systems, the biggest bane is marketing success with the current product line since no-one wants to rock the boat and risk sales and thus there is tremendous skepticism in anything new. We had some good ideas but it was an uphill, frustrating battle that wasn't being won.
Another thing I was working on was software development tools. All our system software was written in assembly language and card input was the way to write programs. I went to several conferences and read lots of stuff and started promoting writing software in a higher level language and using some form of a timesharing system as the environment for writing software. At the time Burroughs was the only computer manufacturer that was doing its OS in a higher level language: Algol was their choice for the B5000 series. Again, although there was management interest, there was extreme skepticism: too hard to learn, too slow, what if the compiler...?.
Honeywell was a member of the MIT Industrial Liaison program and sometime in late 1966. Bob and I went to MIT for a series of presentations by MIT, GE, and Bell Labs on their new project: Multics. This was really exciting and our appetites were whetted. It was an advanced system for multiple users with some level of security, written in a higher level language, and there were some pretty creative people at MIT and Bell Labs working on the project.
As it turns out several people we knew had left Honeywell to go work at GE. Some quick phone calls, interviews, a bit of delay because for the first (in a long series) attempt by GE to cancel Multics was in play. MIT/Darpa talked them out of it and soon job offers arrived. With great pleasure, Bob and I each went to our common boss on the same day to hand in our resignations and we were off to new jobs.