How Cows Solve Arguments

[I have told this story several times, in different media. Unfortunately, my retelling on X has disappeared into the sands of time. So, I am retelling it here, as I reference it from other pages on this site. This is not intended to be a stand-alone essay… just a fixed place I can point to to tell this story.]

In 1990, the heads of “Exploratory” work (POCs for next-gen product) at Bell Labs decided that we should work on the “Seven ‘Selfs’”: self-healing, self-inventory, etc. That is, the question was how future telecom networks could work autonomously. I was the architect in the Local Access Exploratory group, and I was teamed up with my peers from the Switching and OSS (Operational Support Systems) groups to try to solve this complex question.[1]

The three architects started to design a control architecture that was flexible and extensible enough to solve all of these issues. And we soon got into heated arguments about things which often came down to that we had different definitions of the terms we were using in the discussion. Each of the three groups - Switching, OSS, and Local Access - had different jargon with different ideas about things as fundamental as “a link”. We’d talk past each other, get annoyed, and finally figure out that the whole argument was moot because we were talking about different things with the same words.

Then, one day, the OSS architect[2], who’d grown up on a farm, announced that we were no longer going to talk about links. Instead, we were going to talk about cows. Cows? Yes, cows because in the realm of telephony, there is no concept of a cow and therefore no preconceived notion (much less, three preconceived notions) of how they work.

So, we set about discussing cows. We invented cow-terms (like “bison” or “Holstein”) for cows of specific types and properties. We used cow-terms (like “butchering” or “grazing”) as verbs. And so on, until we succeeded in creating the needed architecture (albeit in cows).

Then we wrote a dictionary where we translated each cow-term into a telephony term and defined it based on our discussion. And finally - most importantly - we enforced that dictionary with our groups until the groups could effectively communicate based on common terms with the same preconceived notions.

Ultimately, the project was a success. Key elements of it were prototyped and demonstrated on real network equipment of the day. There were two patents[3] that presaged telecom concepts decades ahead of their reinvention. It was a triumph!

Because of cows.

[1] To give you an idea of how complex it was, consider that it used to take me two hours to convey the questions, constraints and issues we were working on… to a technical audience. And then an hour to explain how our architecture solved them all.

[2] His name was Douglas J. (Doug) Murphy. I would link to him if I could, but I can’t find him.

[3] See “Method and apparatus for establishing connections in a communications access network” US 5,386,41
and “Communications access network routing” US 5,381,405