John D. Mitchell's Blog

March 2005 Archives

Communicating Conundrum

Kim Burchett posted a great story illustrating the Misunderstandings that happen in the software business. Hilarious, in a sad, scary, and all too true way.

Metcalfe's Law broken

Way back when, Bob Metcalfe, one of the creators of ethernet, coined a formula that says that the value of a network grows on the order of n². During the internet boom, that so-called "law" was used in bazillions of pitches to (attempt to) justify all sorts of aggressive growth strategies.

Andrew Odlyzko and Benjamin Tilly have recently written a paper, A refutation of Metcalfe’s Law and a better estimate for the value of networks and network interconnections which dissects Metcalfe's law (and Sarnoff's law). They also present arguments for a better rule of thumb: the value of a network grows on the order of n log (n). This new guideline seems to fit our intuition as well as actual practice much better.

Steele Fortress

Guy Steele is leading a group to build a new programming language called Fortress. In homage to the old SATs, Fortress is to Fortran as Java is to C++. :-) That is, Fortress is about doing high-performance number crunching. Alas, Fortress is not yet available for us to play with. The complete article is: The Soul of a New Programming Language.

The most telling quote:

"I'm now not convinced that a single programming language can serve everyone's needs, because the needs are so diverse." —Guy Steele

Googlezon

EPIC 2014 is an 8 minute, Flash movie from the year 2014 about a very 1984-ish outcome to the juggernaut that is Google.

Hilarious, in a scary way. :-)

Belief of Control

Well, I am a sucker for discussions about risk and software development. There are some interesting tidbits in Tiwana and Keil's: the one-minute risk assessment tool article for the ACM's Queue magazine. Alas, there are some fundamental problems with the article. So, definitely read it, but with a few grains of salt :-).

First off, while they noted the (potential of) self-selection bias, that's buried in a sidebar.

In conjunction, they completely gloss over the fact that they are only dealing with the stated assessments of the participants (i.e., the upper-level managers) of these projects. That is, they miss the entire point that the stated reasons rarely coincide with the real reasons.

In addition, they only talked to high-level managers — where's the talk with the people actually doing the work?

Building on those mistakes, of course the outcome of the research supports the notion that the managers can exert a lot more control over the success or failure of projects. Now, you're probably nodding your head and agreeing — fine, that fits our underlying intuition but the article and those managers are making rationalizations — not revealing reality. The first problem with the rationalization is the Belief of Control. I.e., the belief that managers (and other people in positions of authority) have that they can control everything. [Though, I love the ending of the article with the paraphrase of the Alcoholics Anonymous prayer. :-)] On another hand, in the vast majority of projects, of course the managers can control the success or failure of the project — because the managers are the cause of most of the problems. That is, the managers consciously and unconsciously set things up to fail through e.g., ignorance, neuroses, politics, greed, and just plain incompetence. Even worse, as the project unfolds over time, the tendency of managers is to apply those same broken approaches to "fixing" the problem and thereby usually just make things worse.

"We have met the enemy and he is us." —Walt Kelly, Pogo

This work is licensed under a Creative Commons License.