Research Paper - The Basis of Our Work

[Daryl Kulak]

Okay, so much for the easy stuff.

Attached is the first of Hong's research papers, published in the online journal BioTheory.

I would love for you to read the entire paper (26 pages) and give us your thoughts. Although the work is meant for an academic crowd, I believe that Hong has kept it pretty readable.

If you want to understand the very basis of the work we are doing with "Agile + Rigor" you will need to understand these concepts.

My suggestion is that you print out the paper and read it on the plane or in a chair after the kids have gone to bed. It will take some quiet time for you to digest it.

There are two more papers that follow this one, but for now I'm just going to post this one.

Thanks.

[John]

I have a profound respect for Daryl and Hong’s concepts and I always agree, when I can understand them. My education is in engineering and public administration. How I rose to the top of the IT food chain at ODOT is a bit of a mystery. It was kind of a Forrest Gump thing. Having Daryl and Hong as technical mentors certainly helped. I mention this because I find Hong’s research paper almost incomprehensible. I am a reasonably smart guy. By the third grade the nuns at my Catholic school were so disturbed by my questions and grasp of concepts that they suspected I was possessed. They responded by moving me into the 5th grade where I spent all of my time trying not to get the crap beat out of me by my older, bigger and resentful classmates. Daryl mentions that the paper is meant for academics. No offense to academics, but what good will that do? Most executives are, at best, disdainful of academics. And if I can’t get through this paper, none of the decision-making executives I know would even try. Would a translation of this research paper for executives be helpful?

[Mike Cornell]

The main problem is that there are many other research papers which we're not familiar with that make it a bit tough to read.

The premise is simple enough (assuming I understood it). I got the essence to be: Replacing all human activity with machines is troublesome due to the complexity of the edge cases of that activity. The edge cases are impossible to predict for humans and very difficult for humans to deal with (Black Swan events), or often ignored by many businesses (Long Tail). If the human can't model it, the machine will be unable to react to it properly.

I'm sure that is only scratching the surface of the paper, which then dives into pieces which help to prove the model. That's about where my eyes glazed over....not due to the content, but due to me :)

[Daryl Kulak]

Mike,

You are definitely getting a flavor for the theory.

However, let me add a couple of thoughts to what you are seeing as the premise of the paper.

I'm glad you referenced the Black Swan and Long Tail examples. A Black Swan is a high-impact, low probability event and the Long Tail represents the significance of how a lot of low-impact items can almost match the high-impact items.

Bringing this back to software development teams, a machine model team misses many opportunities because of the limits of machine-thinking. It is not really fair to state these as Black Swan edge cases, because there are far too many of them. The machine team suffers primarily because of missing these opportunities to improve, being locked into whatever the direction of the top manager is. The opportunities are not statistically insignficant, like the Black Swan, they are everywhere.

Looking at the members of the team, is it right to say the majority of the people belong in the Long Tail? That doesn't seem correct. Again, many of the opportunities and potential increases in productivity happen within the hits at the top of the curve, not just in the Long Tail.

As I said, I think you have the beginnings of what we are trying to say in the research paper. I just don't want to minimize the impact of the opportunities by saying they are Black Swans (statistically insignificant) or Long Tail (low-impact items).

One more thing. Hong and I feel that using theories like Long Tail or Black Swan, which are meant for products/goods/events, is not appropriate when discussing people issues.

We feel that we basically need a whole new set of theories that tightly link science (like statistics, mathematics, etc.) to human value. It is one thing to treat a book as belonging in the "long tail." But to treat a person as that? We don't think that's acceptable or profitable or productive.

Thanks for this comment Mike. You generated a great discussion between Hong and me this morning and hopefully this reply encapsulates that discussion in a good way.