Wednesday, November 19, 2008

Is there anything worse than a critic? :)

Michiko Kakutani's review in the NY Times of Malc's new book had one part that really makes me wince. Books like Gladwell's (and T-Fried's) can legitimately be criticized for pushing the stories or research too far, but I don't have too much of a problem with it. Part of being a critical reader would be tracking down the original research if something really bothers you. A lot of the social psych stuff that Gladwell cites is widely known (i.e., not stuff that is sitting off in the academic corner waiting to be discovered) and already debated. Similarly, Friedman has more formal training in middle east politics (B.A. in Mediterranean studies from Brandeis and Masters in Modern Middle East studies from Oxford), and on-the-ground experience, than I do so I am willing to give him the benefit of the doubt, and check out things that don't seem right to me. I have more problems with Carl Sagan than with Malc or T-Fried :)

On the other hand, I've never been comfortable with the "10,000 hour" heuristic that Malc talks about, even if it is at least partially credited to Nobel-winner Herb Simon. [Herb Simon anecdote: When I was on sabbatical at the Software Engineering Institute at CMU I wanted to attend a seminar by Anita Borg (Mary Shaw was there also). There was a group in the room and the time for the seminar was close, so those of us in the hallway were getting antsy. Finally, one of us opened the door and asked if the current occupants were "about done". Eventually they left. Yes, it was Herb Simon and his graduate students. I felt a little bad about being part of that angry hallway mob. Actually we weren't angry, and we weren't a mob, but it is the only time I've seen a Nobel laureate kicked out of a room]

Back to Kakutani's review. This paragraph about the airline accident is frighteningly ignorant of the huge amounts of human factors research that has been done in this area.
Mr. Gladwell similarly raises the notion that cultural traditions may play a role in plane crashes, that the 1990 crash of Avianca Flight 52 over Long Island might have had something to do with the pilots’ being Colombian. He quotes Suren Ratwatte, a veteran pilot involved in “human factors” research, saying that “no American pilot would put up with” being held up by Air Traffic Control several times on its way to New York for more than an hour if he or she were running short of fuel. And drawing on the work of the psychologist Robert Helmreich, Mr. Gladwell argues that the pilots came from a culture with “a deep and abiding respect for authority” — which suggests that the first officer was reluctant to speak up when the exhausted captain failed to do so, and that both men failed to talk forcefully to the air traffic controllers, who were tough New Yorkers, unaccustomed to the pilots’ polite language.
The relatives of the 583 people who died on two 747's at Tenerife and those that died in a frozen Potomac river because of poor communication between first officer and captain might disagree with Kakutani. This phenomenon is so common that NASA Ames has studied the heck out of it.

Let's end this on a good note: A front-seat view of a landing at St. Maarten. Click on "more info" for probably the most informative explanation I've seen on YouTube :)

Tuesday, November 18, 2008

Malc at Pop!Tech

Another Malc video, this time he is talking at Pop!Tech 2008 about his book about genius, Outliers (there's also a Pop!Tech talk of his from 2004 discussing the Blink material).

In Outliers he talks about the "10,000 hour rule" for becoming an expert in just about anything, made famous in a 1993 article "The role of deliberate practice in the acquisition of expert performance" by Ericsson, Krampe, Tesch-Römer (Fresno State people can click here).

Thursday, November 13, 2008

Tokeneer

Tokeneer sounds like a fan of the Lord of the Rings books and movies (OK, that would be Tolkieneer) but it is the latest of a public example of a "fully" formal software development effort. You can read a few paragraphs about it in the most recent Dr. Dobb's here and peruse the official website. Software engineering students will be interested in some of the statistics, particularly about productivity
The Tokeneer ID Station system’s key statistics are:

lines of code: 9939
total effort (days): 260
productivity (lines of code per day, overall): 38
productivity (lines of code per day, coding phase): 203
defects discovered since delivery: 1
In some ways I find this a little depressing since I see some of thee same people who were doing FM when I was doing my PhD in this area in the late 1980s (Tony Hoare, for example) and the number of lines of code that are being verified (my PhD advisor formally specificed and proved a secure UNIX kernel in the late 1970s, now doing voting machines).

That said, this NASA formal methods workshop at Ames sounds encouraging, and I recognize some of the names from people I bumped into when working with the formal specification of the space shuttle jet select group at JPL.

Monday, November 03, 2008

Pinker

I forgot to mention that cognitive scientist Steven Pinker (you remember him) was on BookTV this past weekend for, yes, three hours. He's an interesting guy, and there is at least one caller from Fresno.

Magazines

Two things from the latest MIT Technology Review: and an article from the Honolulu Star-Bulletin about using an 800mm lens to take pictures from the stands at a UH football game, and from Scientific American, a sort of ridiculous article about using "malevolent logic to define the dark side of the human psyche". hmm.