Multics
01 Dec 2012

Site History: ASEA

Location

ASEA, Västerås, Sweden

First Installed

1979?

Configuration

unknown

Honeywell Folks

Site Analysts: Jim Mackenzie, Steve Hopkins.

Application development: Ed Ranzenbach, Charlie Spitzer, Mike Meyers. Plus two more Europeans.

Sales team: Gunnar Andersson

Branch manager: unknown. Account manager: unknown.

Honeywell hardware guy: Janni ?.

ASEA Folks and Consultants

Allan Anundi, Tim Ehrler, Roland Orre, Åke Larsson, Ingemar Enkler, Olin Sibert, Soren Soderholm, Haken Bergman.

Anecdotes

[Charlie Spitzer] I don't have a lot of memories about ASEA except for what I worked on. The machine was sold to ASEA for a year or two before I got there.

I was stationed in Västerås on a short term contract, arriving for 3 months in late summer of 1980. They were impressed with me and asked that I stay on, but I then left to sell my car, pack my belongings, and returned in early spring of 81. In april of 82 they canceled the project and I left may of 82. I don't know when it was installed, but I believe it was early 80 or late 79, and I don't know when it was deinstalled. I think it was there until at least 83.

The onsite system engineer was Jim Mackenzie. From ASEA, he went to work at a Multics site in France in 1983, and then went to Stratus London for a number of years. I went over there with Ed Ranzenbach and Mike Meyers. Jim was already there a year before I got there, and he was responsible for being the sysadmin and keeping the system up. Ed came from (I believe) RADC and we did most of the work on the Multics side of the project, and Mike was a non-Multics program manager from HIS McLean. Another name was Steve Hopkins, an HIS SE on a 6 month loan from one of the Paris or Grenoble sites.

ASEA bought a Multics because they had contracts with various governments to track paperwork (manuals, design docs, build sheets, etc (basically every piece of paper)) that went into building a nuclear power plant. These documents were produced on a minicomputer and then uploaded to Multics, which had to store, retrieve, and possibly send them to test machines, which did some sort of robotic testing on production lines, as some of the documents were actually robot programs to do automated testing of parts being built. ASEA was very big in the robotic manufacturing area (both using robots, and building them), I think top in the world. hence the need for security.

I worked on trying to get the comm working between Multics and the set of mini computers that fed and received these documents/programs and test results from the robots. I don't remember what they were, but it was some kind of European type computer, and not anything available in the US. This was really before FTP was broadly available. At one point they wanted use Hyperchannel to connect them, but that was just being developed also I believe, so never came to fruition. We didn't connect to anything outside of Sweden except for talking to kernel/comm experts on System M in Phoenix. I also worked on the backend database storage/lookup/retrieval for the documents.

A couple years after I left, the project broke down at ASEA, and eventually Multics got deinstalled. I lost contact with the HIS salesmen at that time, so I'm not sure that the info will be recoverable.

I remember the document storage project: the project was to store this giant set of documents in a tree structure, they wanted to use the dir hierarchy to name them, and there was a limit (10? 12?) on the tree depth that wasn't enough. They tried to get me to change that, and it took me quite a while to convince them that they really needed a document storage database mgr to do the translation to/from ID to file pathname. That's the part I worked on, along with trying to jimmy the login supervisor to allow the Modcomp's to log in and look like a user who just typed fast.

The project manager, Haken Bergman, didn't like that Ed and I listened to music at our desks on these new fangled devices, the just-released tape Sony Walkmans.

[Olin Sibert] ASEA was my first big consulting project. I was to be responsible for making the ring zero Hyperchannel software and integrating it into MCS. The whole idea was crazy: using a pair of $100K Hyperchannel adapters to connect Multics to an obsolete 16-bit Modcomp minicomputer was a project that should never have gotten started. As I recall, they figured that out for themselves and switched to Motorola 68K systems instead of Modcomps.

It's funny--a nuclear power plant records system actually makes sense for a Multics installation, but I don't remember that part at all. What I do remember about the Modcomp/Hyperchannel exercise was very different. I recall that the idea was for Multics to host an updated version of ASEA's proprietary cross-development environment and OS for the Modcomps. They'd been using Modcomps forever, for robot control, and had done a whole bare-metal OS and environment for programming them. But everything about it was obsolete, and they figured out that modern microprocessors, and maybe Unix, would be a vastly more cost-effective solution.

I was there for three months in 1982: February, June, and November. June was the prize. As I recall, the November trip was contractually obligated, but by then they'd decided to give up on Modcomps and none of us had any work to do except sit around the office during the prescribed hours (and not drink coffee at our desks lest we violate work rules).

I remember they had a giant paper tape reader (attached to a GCOS system, maybe?) with reels 3 feet in diameter.

Final Shutdown

1983?

Information from Charlie Spitzer, Jim Mackenzie and Olin Sibert.