Intro: I stumbled onto the notes for this post on a back-up disk. I’ve re-worked and expanded them today. The 15 year old grandson is now 19. He just built himself a new rig for gaming. A lesson in delayed gratification, it took months for all of the parts to be delivered in this era of pandemic.
Back in the 1980’s, I was involved in a group charged to implement three-dimensional seismic imaging technology for a service company. My qualification for membership in the team was the prior two years working in another seismic 3D start-up company, where innovation and determination became failure and frustration which manifested as constant corporate re-organisations.
Striving to defeat failure with lay-offs was was like cutting off an arm to solve obesity. The lighter we got, the less dextrous we became. The uncertainties became a common bond of anger among the employees, it in turn became a daily challenge. Our building was shot-up twice, once with several rounds from a rifle from a long distance and the second time, a drive-by shooting that was standard for LA at the time. One was indiscriminate, the other a targeted miss. We found the bullet in a hole in a window frame the next day and fired the security company. A security guard who was asleep while on duty never knew that a window frame took the bullet meant for him.
The endless staff reductions saw me promoted to data processing centre manager long before my time. My first task was trim to the department workforce from 75 to 50. Not even a month later, the decree was was to reduce 50 to 15. The miracle of it was that I survived six months.
That 3D start-up company was the brainchild of a past-president of the Society of Exploration Geophysicists. A brillliant man, his excellent idea was to manufacture and run an innovative seismic recording system that could record a magnitude more seismic data than anyone else. My direct boss had the ear of the founder and was mediocre but arrogant (or ambitious). He kept the MBA certificate on his wall office lest we forgot his qualifications.
The MBA did not prepare him for the computing and geo-scientific challenges implicit in having a magnitude more data than anyone. In those formative years, in my fourth job in four years after graduation, I managed an enormous computing centre without any management training under a boss who didn’t really understand the business. No matter how much effort we put into solving the geo-scientific problems, the boss and the founder never understood the computer run-time challenges that sound trivial forty years later.
Think about Moore’s Law for a split second. If I express the empirical relationship in increasing transistor density as a doubling of computer capacity every two years, then, going back in time, the computing speed in 1981 was less than one ten-millionth of that in hand today. Capacity and speed are like pixel-count and resolution, inappropriate substitutions but they work for plain speech communications. Our computers in 1981 were blindingly fast for the day but mind numbingly slow in hindsight.
To illustrate the challenge, I’ll mention one particular computer job that ran uninterrupted for seven months, just one processing step of many that was part of an overall project with 18 months turnaround. The biggest issue in geophysical computing is always how to get the data into the machine. We rewrote the code to merge several steps into one so that you read each bit of data just once and acted on it repeatedly before summing it into an evolving final image which we saved every 36 hours. We minimised the i-o and maximised the use of devices called array processors. Testing and parameterisation for such jobs was not for the faint hearted and pre-dated the rigorous and rational QA/QC and testing plans in use today. You don’t want to spend seven months to learn that you chose the wrong parameters. This was innovation on the bleeding edge for a technique called pre-stack time migration, over a decade before its mainstream use.
Suffice to say that when bankruptcy protection in the form of US Chapter 11 arrived, my insolvent employer blamed their hardware and software supplier for not meeting the contractual performance criteria. There was an element of truth to this assertion, in my opinion, but some might say that my employer’s greedy or naive over-expansion could have been a contributing factor.
And so we left California and headed back to Texas. In my fifth job, actually my third time at the same company, where I was involved in a group charged to implement seismic 3D, we were an industry renowned full-spectrum seismic service provider. We manufactured land and marine equipment, we ran data processing centres around the globe and even developed seismic interpretation systems. We had headstrong managers who believed in what they budgeted for. We were there to make it work. So we built a fleet of innovative seismic vessels with cutting edge source and receiver arrays. We developed and sold huge channel-count onshore recording systems. And I was working in a team whose role was to work out how to process all the incoming data, how to reduce the vast quantities to interpretable and dependable images in every shorter time cycles. Remember this was over 30 years ago and you might think your phone or tablet could probably handle it now. That might be true for many algorithms, especially if you linked hundreds of tablets or phones to work in parallel, but as I wrote earlier, the biggest issue in seismic computing is always how to get the data into the machine.
A few years into this fantastic technological journey came an industry nadir of 1986 and within a year, we found ourselves merged with another industry renowned full-spectrum seismic service provider who had embarked on a competing parallel development and expansion. And failed. But like us, they couldn’t see it let alone admit it.
Here’s the point. There were several common ingredients in these experiences that taught me personally valuable lessons.
One life lesson came from battles that took place over control of the ‘vision’. There was a tendency among most of my employers to hoard or squander political capital when in fact, collaborative and tactical practices were what was really needed. Competitive forces blinded most managers. Failures were inevitable but blame trumped insight. Two generations on, I’m pleased to say that my 15 year old grandson understands this at an earlier age than I did. A poster has appeared over his desk that commands him to ‘Build your own dreams or someone else will hire you to build theirs’.