Thanks to the invention of the first microchips, minicomputers spread to all industries from the 1970s onwards. Chuck and Ken, a father-son team of IT experts, describe the role of the so-called “dealers”, in the 3rd part of their mini-series. Thanks to Chuck’s anecdotes, we understand why the sale of minicomputers was so closely linked to specially tailored software back then.

Our first two epics were about Data were plentiful, and it was all about turning that data into actionable information. In short, it was about enabling businesses to become data-driven. The problem with this new information and “business intelligence” is that it was addictive. Large enterprises that had it, wanted more; and smaller companies wanted badly to join the game and enjoy the benefits. Fortunately, the introduction of the microchip in the 1970’s kicked off Moore’s law, the technology was very obliging, and it offered just the right opportunities when companies started to feel the need.

It continues with processes

The world of minicomputers looked a little different than it does today – with smaller monitors and uncomfortable keyboards.

The mid 1970’s to the late 1980’s saw the introduction of the minicomputer, enabled by the ubiquity of the first Microdata chips. But these minicomputers themselves were not yet commoditized in any way; they ran proprietary software running on proprietary operating systems. So-called “dealers” associated with each manufacturer sold “turn-key” solutions they developed to address a specific business challenge for a specific customer. It was not just the dealers’ job to sell the complete package (hardware and software) but more importantly, to create or tailor the software to meet the specific needs of the business. Whether you were in the garment industry, trucking industry, or hotel industry, the “Age of the Dealer” had a turnkey solution for you.

The “Commodity Era” spanned from the late 1980’s to early 2000’s.  But be careful: when we write commodity, we refer to quite different topics. The commodization of hardware i.e., the PC was a big driver, to be sure; companies no longer were locked-in to a single minicomputer vendor. But the importance of the commodization of software is greatly underappreciated. There were many flavors of important packages such as ERP, and the barriers to companies switching between them were lower than at any point before. In fact, it was these packages that enabled the IT workforce to transform from company-experts to industry-experts, all speaking a common language. And finally, there was commodization of people; freelancers could easily switch between companies, and the all-around data processing engineer split into many specialties, such as business analyst, requirements engineer, programmer, tester, etc.

The “Infrastructure as a Service” pendulum starts swinging

Not only are they all the rage today, but in fact Infrastructure-as-a-Service and its siblings (Software-as-a-Service and Platform-as-a-Service) have contributed greatly to lowering the adoption barriers and accelerating the pace of digital transformation. But before giving too much credit to AWS or Azure, it’s useful to be reminded these are very old concepts dating back to the Age of Iron. We tell this story now to set the scene for the rise of the minicomputers. Something called the Service Bureau (SB) whet many appetites for digital transformations at small companies.

When I became editor of perhaps the earliest IT magazine in the early 1970’s, every month I would ghostwrite a column for ADAPSO, the Association of Data Processing Service Organizations. Hitting their stride in the 1960’s and 1970’s, before the minicomputer revolution, Service Bureaus were a booming business. Started by IBM, who started a division named Service Bureau Corporation (SBC), their philosophy was:  if you can’t buy, then lease.  If you can’t lease, we’ll do your computing for you.”
The SBs would do as much as you needed: they could keypunch sales receipts, orders, and payables, and turn them into reports. They could print payables checks or even paychecks for you.  Anything an in-house computer would do, they could tackle. In fact, the biggest one – SBC – was so successful that the government said IBM was monopolizing an industry (computing) and forced them to divest of SBC and set it up as a new corporation. Does this remind you of any IT companies today?
The smaller service bureaus – to compete with IBM – formed a trade association: ADAPSO. Many of the SBs offered time-sharing, this was a multi-faceted set of services.  For those who had a smaller computer, the SB could hook up a data line, let them upload data, process it on the SBs bigger-faster mainframe, and download it back.  Or even use it as data entry, hooking up an entry device (card or paper tape) and uploading it that way.The Age of the Dealer caused the demise of the SBs. Minicomputers gave companies autonomy, faster results, and better overall cost. So ADAPSO went off and became a high-tech organization, not involved with end users.  And one by one the SBs closed their doors.  There are still some around, but the minicomputer avalanche did most of them in. (Chuck Ritley)

Enter the minis

We’ve just seen the Service Bureaus enabled smaller companies to craft their own digital transformations. But with their digital appetite now whet, they wanted more features, more flexibility, and lower price that what the SBs could offer. In short, minicomputers let smaller businesses kick their digital transformations up a notch.

In 1972 I was hired as the editor of the first IT magazine. My job was to find stories of how computers were used in various industries. So I traveled to many sites, writing stories on mainframes controlling assembly lines and production control. Great reading for the engineers – but not so great for our business readers.

But that fall, I got a call from a PR guy in Chicago: “Charlie! You’ve got to come here. A German company – Nixdorf – has already sold a dozen “minicomputers” to small companies.  They’re pre-programmed, ready to go. Sort of a turnkey operation.”  Now PR guys always exaggerate, but the publisher had a permanent condo where I could stay in Chicago, my boss was intrigued and baffled by the word minicomputer, so off I went.

And I was blown away!  A computer dealer – a new concept in itself – had already sold and installed 12 of these things and had orders for lots more. They took me to the showroom and there it was – the size of a desk with a box underneath and a CRT terminal on top.  That alone was strange, since mainframes rarely had terminals.  The dealer explained that was the control, and there could be 7 more scattered around the business.  “Does away with keypunching because anyone with data to enter can just sit down and enter it.” Remember, as an experienced mainframe guy seeing a minicomputer was hard for me to accept. “I’m half sold,” I told him, “but, can we visit some real end users?”

So we did – 3 of them, but I’ll tell you about the first one.  It was a medium-sized metal works company, making specialty brass and bronze parts for plumbing fixtures. The CEO was surprisingly versed in IT jargon. He told me: “A couple of years ago I was up to my knees in paperwork, and so was everyone else here. I could have used a computer, but I started thinking: how do I get a system designed, how long to get it programmed, who will supervise it?   There’s was no way I could think of.  So I went with an SB – a Service Bureau – and for a while I was a happy camper. We had the operational data I needed, but it wasn’t flexible, and it wasn’t cheap.”

So when the Nixdorf dealer offered a minicomputer solution, it sounded attractive.  He could capitalize the expenditure; the dealers’ programming staff would take generic modules and customize them to his needs and be available down the line to do more.  But best of all, it cut back on paperwork. Having 8 terminals placed in various departments let staff key in data directly rather filling out a form for the SB.  He could start with as little as 4 terminals, adding others as they were needed.  And the best part: the terminals made data available in real time – no more waiting on weekly reports.

I knew that people were important, so I interviewed his staff. They were happy. No one was a dedicated data collector, so feeding the SB with data meant more busy-work for everyone. With their new mini, they’d were free of this and could get on with their work. And everyone loved the real-time data aspect of the terminals. I got back to the condo that night, wrote the story, and faxed it off to the magazine. I felt we had just met the future. (Chuck Ritley)

Software becomes a commodity

The Age of the Dealer was not the new face of the Age of Iron, because minicomputers were not about replacing mainframes. These were two quite separate ages, because minicomputers were now put computing power into the hands of small and medium businesses (SMBs) that could never afford a mainframe.  But the transition from the Age of the Dealer to the Commodity Age was different still: at first, the SMBs now had established investments to protect.

When you reach the point where software becomes so reliable – with capabilities clearly defined – the users’ trust level goes up and parallel runs aren’t necessary. When software reaches this level of reliability and standardization it becomes a commodity.  In the 1960’s, when I worked for a large mainframe manufacturer, they had a very basic Bill of Materials processor – a forerunner of the MRP systems to come. I installed quite a few of them, and it was tough.

By the 1980’s, the minicomputer manufacturer I worked for had a full-featured MRP system: sales, orders, BOMs, purchasing, routing, full accounting – everything you needed. We installed an 8-terminal system at the HQ of the Silicon Valley Inventory Control Society. My role:  teach a few classes a week to those who wanted to know more about MRP. We hoped they’d choose our MRP. But as I waded into this, I looked and I listened, and I learned that most of the competition did exactly what we did. Unbeknownst to us we were producing a commodity. So were they.

So two separate trends melded into a direction. Software standardized into a commodity. But the trend was to abandon the mini for the cheaper PC. Moving into the 1990’s, business realized that a network of small PCs could replicate the work done on the mini – and adding an extra PC here and there was far cheaper than upgrading the mini. But there was an obstacle: taking software modules the business had grown dependent upon and moving them to another platform. This gave birth to a new specialist. Because the mini’s proprietary worlds all sprang from some common roots – Microdata in particular – enter the specialist who could seamlessly “lift and shift” the software from the mini to a PC network.  I was one of them, and as PC networking became cheaper and cheaper, we specialized in this: creating a clone of your mini-based system that performed on a networked PC exactly as the original on your proprietary minicomputer. When we handed you the keys, we went off to the next conversion, like migrant workers. (Chuck Ritley)

Miniseries on digital transformation

This article is the third of a 4-part mini-series on digital transformation. The final part will be published next week.

Acknowledgements

It is our pleasure to thank Maria Kreimer and Nikola Gaydarov for detailed discussions that contributed significantly to this manuscript.