kip/bot/blog

In the midst of many many posts and talks about generalism, it was interesting to see this post about feature creep reminding us of the UNIX credo “to do one thing and do it well.”

I seem to spend a lot of time reveling in craft, worrying about the loss of expertise and specialty, and wondering if we’re not going too far in our talk about generalism. I came across the feature creep post above, by way of this sparse post by Jon Howard:

I found both these posts bafflingly intriguing. The one above was so fast and minimal as to be koan-like, and who writes about feature creep anymore? That seemed so old school. But some pattern recognition module in the brain kept me focused on these two posts, struggling for meaning. And then!: feature creep is a way of describing how generalists and agencies accrete skills! And maybe it works . . .

Feature creep, traditionally understood, is problematic to the maker and the user of the product. The maker, by being spread thin, misses deadlines, goes over budget, dilutes the initial impact of the final product, and makes that final product hard to maintain, service, or evolve. To the user, feature creep results in confusion: the subtle differences between choices waste users’ time parsing meanings; the number of choices can be paralyzing and unsatisfying; finding things is difficult; and the cruft underneath somehow rises to become the experience of the product.

At its core, feature creep is worrisome, because it spreads out talent across a wide number of things, at the expense of doing any one thing well or in an elegantly integrated fashion.  The accretion of skills when one becomes a generalist runs that same risk:  “I’ll learn to do this and then this and then this.” We mistake conversancy in a skill with competency as we move from skill to skill, discipline to discipline and then what do we have?
When I look at resumes or talk to designers in the generalist vein, I find myself struggling with the long list or big concept. Long lists of skills and competencies have my pattern recognition module seeking what’s missing so I can understand what unique value that person could bring to the work. The big concept (”it’s all design” “it’s all advertising in the end”) leaves me suspicious.
Perhaps the tools we have for managing feature creep (does it add inherent value to the product/person, does it add noticeable increased value to the perception of the product/person (will it help me sell myself) could help with the balance.

A nice contrast to Clifford Stoll’s aversion to going deep, is a Freakonomics blog piece about, god help me, sports. The piece is about “expert performance” (no Wikipedia entry!) and the related concept of “deliberate practice” (still no Wikipedia entry!). From the blog post:

When Anders Ericsson and his colleagues in the “expert performance� movement — we’ve written about them before, and we’ll write about them again — try to explain what it is that makes someone very good at what he or she does, they focus on “deliberate practice.� This means that, your level of natural talent notwithstanding, excellence is accomplished mainly through the tenets of deliberate practice, which are roughly:

1. Focus on technique as opposed to outcome.
2. Set specific goals.
3. Get good, prompt feedback, and use it.

The piece goes on to cite an article about an early trainer who saw the young A-Rod practice, and practice hard, at his hitting and fielding, focusing on weaknesses and mixing things up.

This looks like a bottom-of-the-T approach (technique, incremental goals, incremental feedback, iteration) but it has resulted in a complete player (top-of-the-T?). It’s hard to generalize too much about anything applied to A-Rod, given the universal recognition of his inherent talent. (Reggie Jackson, in the second article cited comments on A-Rod’s seeming enjoyment of the training routines, said, “A lot of things are fun when you’re great.”) But Stoll’s rapid move to boredom with anything done more than two times, in contrast to this look at the development of greatness seemed too close together to not document.

Clifford Stoll’s TED talk may not work for everyone. There’s a hippy daftness that may sometimes feel forced and a self-dismissive “what I do so is so boring” that may feel condescending, but about six minutes in he is charming, oddly moving, human, and clever. There’s also some cool stuff in there, like klein bottles, a grade school experiment to measure the speed of sound, a tribute to Moog, and a pervasive Richard Feynman tinkerer-thinker mode. He constantly grounds himself in tinkering that leads to bigger ideas.

He also has a line that sits in nice contrast to my current reading of The Craftsman and pre-occupation with expertise:

The first time you do something it’s science.

The second time it’s engineering.

The third time, you’re a technician.

He was saying this in reference to his boredom with hacking and computer security. (He first came to prominence with a fun, witty, popular computer science of his detection and catching of East German hackers in Cuckoo’s Nest, a book I still remember with a smile 15 years after reading it.)

On the other hand, he seems to have been making Klein bottles for many years and is still getting something out of it.

=-=-=-=-=-=-

I’m a little disappointed in myself for not knowing Klein bottles. Wikipedia has several pop culture references listed that make me think I should have known it: Futurama has Klein beer sold in Klein bottles, and Magic has an Elkin bottle card.

Maybe it’s not my fault It’s just something damn hippies seem to do: