A Balloon Bursts

Every now and again in the workplace we run across someone who suffers from a "Napoleon Complex."  This is a personality that grasps at any pretense of power, no matter how small, and inflates it and himself. Just like a balloon. Sometimes they burst just like a balloon. The public service is not immune to this persoanlity disorder. In fact, in some ways it encourages it.

After two years at Energy, Mines, and Resources  I had established myself as the VMS expert. I fell into the position by default; no one else wanted the position because, after all, VAXes weren't real mainframes and so were beneath the dignity of most of the 200 or so employees in the computer support centre. (A few years later about 180 of those mainframe experts were let go in a massive layoff--the mainframe day had come and gone--another story.) Besides, I had taken three five-day courses in system management in Toronto and so was also the only one in the department who was qualified for the position. I had two machines to look after when a third appeared on the horizon. The translation bureau of EMR bought a VAX to use for document tracking. The machine was delivered to the office of the self-appointed project manager. Let's call him Sammy, as he certainly would be very embarrassed if the real person were associated with this story. He should be.

Though VAXes were not mainframes they still required a controlled environment. That meant air-conditioning and basic cleanliness (isolated from the dust and debris of normal office life). Besides, this particular machine was about the size of a modern large freezer chest. An office definitely was not the right place for it. Still, Sammy wanted it to be set up in his office. One of the senior operators and I appealed to his boss who quickly agreed that the machine had to be placed in the computer centre. Sammy was not happy.

I went about setting up the system: creating accounts and printer queues, setting up the main application and the basic management routines, as well as writing up procedures for the operators to follow. One day Sammy showed up my office with a fellow who offered me a 9-track tape. Sammy was excited saying that the wonderful program on the tape would speed the system up. Right away my warning lights flashed on. I looked at the tape. It had a hand-scrawled and much-smudged label on the casing. I politely told Sammy and friend that we had established procedures to be followed and that I could not even mount the tape on a tape drive without proper authorization. Sammy was insistent that it was "his" system and so I should do as he said and install the program that was on the tape. I stood my ground. Later that day the head of the computer department came to see me and asked about this tape that my client wanted installed. I was surprised at how far up the totem pole the disagreement had reached.  I sketched out what had happened and he left, satisfied with my account.

Communications in the pre-PC days was usually by using lines leased from Bell Telephone to link a user's terminal (composed of a monitor, a keyboard, and a modem) to a large electronic switchboard in the computer centre. The switch would route the user to the correct mainframe and thus establishing a link for input (from the keyboard) and output (to the monitor). One day Sammy phoned me to tell me to raise the speed-limit on his modem. He would not listen or understand when I explained to him that what he was seeing, "9600 Baud," was it. It was a description, not a variable that could be manipulated. He started shouting that I had to cooperate. I firmly told him it had nothing to do with "cooperation;" he was asking for the impossible. Really, it was a 9600 Baud modem and there was nothing I could do about it. He may as well have been asking me to make his car float in the nearby lake.

Another feature of pre-PC days was that users shared resources on a central computer. Methods had been established to ensure that each user got his fair share of CPU access. On VAXes this was managed by a round-robin interrupt-driven preemptive system. What this meant was that access to the CPU was determined by priority. If the system lost power, it required the highest priority in order to ensure a safe shutdown as the power drained away (I was amazed to discover how much a CPU could accomplish in the micro-second of a power failure.) The highest was priority 31 in a list that began at 0. As events occurred priorities of "processes" (batch jobs, interactive sessions) were raised or lowered. For example, if an interactive session issued an I/O request (usually access to a device) it would be granted another two position points on the priority scale. With each cycle through the CPU the priority would drop by one. Interactive sessions were given a base priority of four--the position below which they could not drop. This meant a batch queue could be set up at, say, priority two, to use up CPU cycles when no interactive users required CPU cycles. This is important to get because this issue is going to come up again and again as I relate my experiences.

So, one day when the system was particularly sluggish I check the jobs running on the system and found an interactive session with a base priority of six. That meant that that process was hogging access to the CPU. It was, you guessed it, Sammy's account. I lowered its priority to four and system functioning returned to normal. A short time later Sammy's session was again at a base priority of six. I phoned him and told him that he was hurting the system's performance and asked him to please stop tinkering with the running of the system. He was upset and refused. I went to my boss, who went to Sammy's boss. I was then told to take away the access to the system (privileges) that allowed Sammy to adjust CPU priorities. Sammy was livid. Apparently he locked himself in his office and when anyone knocked he shouted, "No entry! Insufficient privileges!"

VAXes came with a communications port, usually with 16 slots. That way the computer room switch could route up to 15 sessions to the machine (one slot was reserved). There was space for only one communications port on this particular model. There were more than 15 employees in the translation bureau. The machine couldn't support more than about 12-13 users at any one time in any case because of RAM limitations (they would spend most of their computer time having their sessions written out to and read from disk.) In a meeting to discuss solutions to the problem, Sammy suggested getting another communications unit and plugging it into one of the free slots on the existing one. I assume the reader has enough experience with basic logic to realize how ludicrous this suggestion was. You have 15 apples. Cutting one of them into 15 pieces is not going to give you more apples. Fortunately no one needed my expertise to respond to that suggestion.

And so it went. These are just a few of the examples of his behavior.  Sammy became more and more frustrated as his dream of controlling the computing for the bureau he worked for melted away. His behavior got more bizarre. He attempted to copy the entire operating system files to his own account. He edited batch jobs required for the running of the department in order to "speed them up." (His changes would only slow the jobs down.) He phoned me with more absurd demands.  At last, I was asked to document Sammy's requests and behavior. He disappeared from my life--and the life of the translation bureau. His replacement was a reasonable and pleasant woman and we got along well for the duration of my stay at EMR, and later, when I was consulting to the department.

Comment: On re-reading this I realize that I didn't make clear why "Sammy" was so concerned with speeding up the computer's operations. The system was undersized from the beginning. The absolute maximum number of simultaneous users it could support was 15, but, performance fell off sharply after 12-13 users. There were 25-30 employees in the department it was intended to support. Some simple arithmetic could have predicted this. They bought a VAX 11-750, but needed, at a minimum, a VAX 11-780 (roughly twice the speed and capabilities). As Sammy had been involved in preparing the original specifications for the system he was somewhat anxious to prove that he was right and the reason that the system did not meet their requirements was because I was uncooperative.