Part II: Leaps Forward in Computerization
None of us would forget the strange sounds that signaled a new era in truss design: the pounding of the keypunch, the whirring of the card reader, and the ratcheting of the line printer that emanated from the 10x10 room in front of our drafting bullpen. Yet, when you walked into that room to submit a coding sheet, not a sound came from the desk-sized device that triggered those actions, the IBM 1130; only a series of flashing lights. When I arrived on the scene in 1970, that machine was efficiently cranking out truss designs, one every couple of minutes, changing the truss world.
But it took two prior years of hard work by the engineers who sat in the back of the bullpen to get the 1130 going, all while coping with the onslaught of increasingly complicated truss requests coming through our office. Those stalwarts faced even more pressing deadlines than I faced when I ran the department a dozen years later. Tempers were short during the run-up to 2.4 Million housing starts, magnified by the unreliability our only delivery option, the U.S. Mail. When we finally got designs to our customers, they still had to hand calculate cutting lists. Thus the urgency to gain more productivity from the computer – which continues today.
The result of this work was that truss member design and joint plating began to be done by the computer. Our refocused draftsmen initiated the process by completing 7 lines of data on a computer coding form for each truss. Those lines of data were keypunched onto 7 computer cards, which were stacked on top of a batch of similar truss “jobs” being consumed by the whirring computer card reader. The output of each truss design was printed on a single page of fan-folded paper, which was eventually formatted to comprise the text of a sealable truss drawing.
Our design drawings were thus transformed from ungainly blueprints to Xerox-able pages consisting of master truss drawings and computer output, with minimal hand annotation. Drafting dexterity was largely replaced by cut and paste – except for non-standard and non-triangular structures. The former became the object of my first job, programming non-standard joint types. The latter was assisted by an ingenious IBM program called STRESS. This versatile tool gave us all the capabilities of the Purdue Plane Stress Analyzer, a full four years before PPSA became available. Ironically, we were reluctant to use STRESS because it required us to go backwards and do all the hand drafting work we had become loathe to perform. Even if we plodded through the tedium of entering each joint coordinate and member property into STRESS, we had no geometry check of the structure. Fortunately, IBM came to the rescue in 1971 by enabling compatibility of one of the first available graphics output devices, the CalComp pen plotter.
In retrospect, no single industry innovation has improved productivity as much as our first computer work, and we barely scratched the surface in those early years. But most of the credit belongs to IBM’s masterwork, the System/360. Though we could afford only a slimmed down version, we gained most of its game-changing features. What I had programmed on the 360 worked fine on the 1130 four years later. Compatibility of hardware and software was the genius built into IBM machines that has been successfully emulated by Microsoft software a generation later.
The enduring lesson provided by IBM’s early success is that “size matters.” In the computer business “size” refers to the number of users, not necessarily the size of the business. Droves of new users stampede to innovative technologies. IBM gained an unprecedented 8000 orders (new users) for 360s in the first year of the 360. Ditto for Microsoft with Windows, or Google with their browser, or MiTek with Sapphire!
Next Month:
CMs Gain Access to Computer Power