Tuesday, September 20, 2005

Process thinking

The Unfinished Revolution
Alan Key: It is amazing to me that most of Doug Engelbart's big ideas about "augmenting the collective intelligence of groups working together" have still not taken hold in commercial systems. What looked like a real revolution twice for end-users, first with spreadsheets and then with Hypercard, didn't evolve into what will be commonplace 25 years from now, even though it could have. Most things done by most people today are still "automating paper, records and film" rather than "simulating the future".
Visit to Bootstrap
In 2003, the five members of the Meta-Learning Lab huddled around a conference table at The Bootstrap Institute to listen to Doug Engelbart recount the story of his life's work. Doug told us that for fifty years, he had pursued one goal: augmenting the human intellect--to boost mankind’s collective capability for coping with complex, urgent problems.

No success
I asked Doug who best exemplified the spirit of his work. Who had put his ideas into action? "No one," he replied. It was tragic. Fifty years and yet to reach the finish line. Augmentation of intelligence was the very reason Doug invented the mouse, windows, bit-mapped graphics, word processing, and the network. How could such a brilliant mind have failed to enlist some takers?

How to bootstrap
Doug's objective is simple and compelling: "A-B-C's of bootstrapping. Any organization's stock in trade is called here an A-activity; its ordinary R&D work to improve on A is called a B-activity. The bootstrapping strategy serves to improve on B and is called a C-activity. The value of C may be perceived as garnering compound interest on an organization's intellectual capital."

The work of Chris Argyris is quite similar. "Double-loop learning" involves reflecting on one's learning, asking if there's a better way to accomplish it. Intellectually appealing, Argyris's work has not been adopted widely. When the Meta-Learning Lab talks of "going up to the balcony," it's another case of improving the process and not just the instance at hand. In spite of showing enormous potential returns on investment, not one corporation has shown an interest in funding a project.

What's going on here?
Waiting for someone to arrive at the North Berkeley BART station this afternoon, I pondered why organizations had failed to pick up on something with such enormous and obvious potential. Boiled down to essence, the process of augmenting human intelligence is no more than applying continuous improvement to learning. As Doug wrote long ago, his mission includes:
An Italian tailor has a private audience with the new pope. Upon returning to his neighborhood, his friends ask what this new German pope is like. "He's a 42 Regular," says the tailor. Every profession has its blinders.

Physicians and surgeons
When I was in my twenties, I had an annoying medical condition. I visited my physician weekly. He was a bright guy, a graduate of Johns Hopkins, very articulate. But he couldn't figure out what was going on. He checked my blood. He checked, ahem, other fluids. He sent tissue samples to the lab. No luck. I seemed to have some condition unknown to medical science. After almost a year, by which time my condition had deteriorated significantly, my doctor said, "Ah ha. Maybe this is not a medical problem at all. It might be a surgical condition." I went to a surgeon. He immediately recognized what was going on and scheduled me for surgery at St. Mary's Hospital ten days later.

Going to a specialist I mistook for a generalist, I almost literally lost my ass. I can't resist inserting a bit of humor here. Question: "What do you call the person who graduated last in his class at medical school?" Answer: "Doctor."

You see what you're looking for
Remember the movie Bullitt? Steve McQueen chases the bad guys through the hills of San Francisco in his souped up Mustang. Their cars race down Fillmore Street toward the Marina, jumping into mid-air as they crest each street crossing. I have seen the chase a dozen times since the movie was released in 1968. It still takes my breath away. Not long ago, my wife Uta told me, "Watch the Volkswagen." Now I have a different perspective on the chase scene. Block after block, Steve McQueen passes the same pokey Volkswagen beetle. The movie makers knew that the eyes of the audience would be so glued to Steve's car that they would not even see the surroundings.

The Cobbler's Children
Giving other people advice that's meant for us is commonplace. That explains why most psychiatrists are nuts and why most great learning designers have learning disorders. (See "The Cobbler's Children in the chapter Learners & Their Brains.)

Engineering Focus
Doug is an engineer. An engineer is a person who maintains or controls an engine or machine. The engineer's job is to design or build machines, engines or electrical equipment, or things such as roads, railways or bridges, using scientific principles. Engineers are notorious for underestimating the impact of the human element. You'd no more rely on an engineering perspective to address deep organizational issues than you'd go to the surgeon for a headache. The logical approach of engineering is the opposite of the chaos of evolution. Engineers plan; evolution just happens. Engineering works top down to meet objectives; evolution happens bottom up and only the fit survive. Here's how an engineer might represent the process of augmenting intelligence or double-loop learning:

This is mechanical; it is not how nature works. And therein lies the problem.

Not an engineering problem
Doug and others had addressed process improvement as an engineering problem, when in fact it was a nature problem. Looking at a single, logical case (as implied by the diagram) is folly, for we're talking about human nature here, and that is anything but logical. At Accelerating Change 2005, Steve Jurvenson described a similar situation in science. (His blog has a rich description of this.)

Over dinner at O Chame
Last night it was my pleasure to have dinner with Dr. Herwig Rollett, the Head of Research Cooperation at the Know Center in Graz, Austria. He told me he'd been thinking about this area, which he calls collaborative cognition. It's described by its output rather than its process. DNA computing works this way.

Another design model
I had drawn a design model of a bucket into which several streams were pouring, and Herwig said that fit his understanding of DNA computing. You throw a lot of single DNA strands into the bucket, mix it up, and strain out those that have found a pair. This pairing up, sort of a mating ritual, is equivalent to massive parallel computing, and you don't find out what has been created until the end of the process.

Diversity required
As the business saying goes, "If two people in business always agree with one another, only one of them is necessary." When working with people instead of DNA, it's important for them to bring different mental models, whose combination may create something new and better.

to be continued

September 21
This evening I went to a presentation by Brian Behlendorf on Apache on the UC campus, after which he and I retired to a local cafe to talk about informal learning. He had brought up DNA computing so I asked him for links. He told me to check Andrew Hessel at University of Toronto, Drew endy @ MIT, and to search for Synthetic Biology.

The Apache Story

In 1995, Brian and seven others took advantage of NCSA's open license of its server software and began reworking it. Brian named the effort Apache because it was one of the last tribes to go and it's also a play on "patches," the medium by which the software is maintained. The Apache email list grew to 100 participants. (Every email from day one is still on line.) Three years later, Apache was running 60% of the servers on the internet. There was no company, just the name. They're weren't and still aren't any full time employees. It's a 100% volunteer effort. They became a non-profit foundation to limit their personal liability.

Apache's primary means of communication with its developer community is via its mail list, and with CVS ("the Volkswagen of version control tools") and its modern replacement Subversion. Bugzilla tracks bugs, and a Wiki coordinates discussion. Apache runs on concensus; the Linux project is a more directive air traffic control model.

Three rules for running an open source development effort:

1. Be humble. The Java community of Apache became toxic; it was disbanded.

2. Be transparent. Make decisions in the group. Keep the decision-making process open to all.

3. Think of the user community. Set up self-correcting mechanisms to address any potential Tragedy of the Commons situations. (Spam has ruined email. IM and VoIP are next in line.)

Look for passionate people. Fred Brooks says a top programmer outperforms a mediocre one by two orders of magnitude.

Communications skills are vital. These should be taught in school.

Everyone on the Apache team was self-taught. They read the online documentation (the "MAN" pages.) They explored. Linguistics majors were natural programmers. People learn Apache through observation and osmosis. They lurk. They watch how others respond to challenges. They incorporate other people's ideas in their work with grace.


From Doug Engelbart's Bootstrap Institute
Among the Institute's missions:
A-B-C's of bootstrapping. Any organization's stock in trade is called here an A-activity; its ordinary R&D work to improve on A is called a B-activity. The bootstrapping strategy serves to improve on B and is called a C-activity. The value of C may be perceived as garnering compound interest on an organization's intellectual capital.

Augmenting Human Intelligence

Comments: Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?