My Encounter with Fernando Flores

I had the great fortune to meet Fernando Flores when he was working with Doug Engelbart at Tymshare in the 70’s. Flores in a session that lasted perhaps an hour or two described to me his ideas for a class of software to make enterprises work more effectively. He had been pushing these ideas, I later learned, since he had been cabinet minister for Chile’s president Salvador Allende. Flores was highly computer literate and I ran development of fundamental technologies at Tymsahre at that time but I did not know him even by name. Someone had suggested to him that it might be good for him to tell me his software ideas. It was for me a unique experience. I had only a casual knowledge of ‘institutional software’ either from the ‘theory of institutions’ perspective or the ‘software engineering’ perspective. Indeed I had then little practice in dealing with issues dear to Flores’ heart.

I understood Flores’ ideas in the context of vague institutional principles, for which I had little knowledge or respect. I am not a ‘company man’ even though I have high regard of companies as an institution. I evidently remembered a number of the goals that Flores promulgated, perhaps simplified versions considering the conceptual chasm between us and time limitations.

It took me several years to come up with a logical criticism of notions that had left me uneasy at the time. The notion, as I recall it now, was for the computer to keep track of obligations and agreements to perform certain tasks by individuals and parts of the enterprise in support of other parts of the enterprise all to the end of supporting the ends of the enterprise.

My enlightenment came when I first encountered the engineering term “load shedding” which engineers would specifically add as functions of mechanical systems to keep local failure from growing to global failure. A fuse or breaker in an electrical circuit is a load shedding mechanism. Engineering system designs without such considerations may be fatally flawed; they may completely destroy themselves upon encountering the first design flaw. I knew about fuses and why they worked, but I had no concept of the larger category of load shedding. I now had a non-pejorative name for something that had seemed taboo. This allowed me to think more clearly. With the new category, and its name, the old Flores concepts came back to me along with words to describe what had bothered me.

If Flores had notions of modifying goals or admitting that previously agreed goals were infeasible, I do not recall hearing them. Perhaps he had not thought thru implementation details. I now suspect that Flores had plans for failure contingencies but as a second class portion of his system—just as electric generator salesmen don’t boast of how the generator may shut down to protect itself.

Lore in capitalistic countries about conditions in Stalinist Russia suggested executions upon failure of plans. Similar things not as destructive happen in U.S. corporations. I recall a computer manufacturer had a long scheduled delivery date for a complex project. A couple of weeks before that date the company announced a two year delay in the delivery date. I recall hearing thru friends that the only people in the company that were surprised were top management. I suspect that upwards information channels in the company did not carry bad news well.