# Wednesday, 06 December 2006

Man, I've been making software for a long time. I first came across Fred Brooks' essay No Silver Bullet: Essence and Accidents of Software Engineering when the mail (you remember, the physical mail) brought me the magazine (you remember physical magazines) and it was the cover story. Apparently that was the April 1987 issue which is bringing us perilously close to the twenty year mark. (And if I needed to, I could get my hands on that copy within minutes... in fact I'm sure I've held it within the last year or two.) And now people are talking about it again. Larry O'Brien has a long post with links to some other posts of his own and by Wesner Moise. He also captures what I consider to be the essential quote from the essay:

I believe the hard part of building software to be the specification, design, and testing of this conceptual construct, not the labor of representing it and testing the fidelity of the representation. We still make syntax errors, to be sure; but they are fuzz compared with the conceptual errors in most systems.

The thesis is that software development gets easier or faster or more accurate only by degrees: you cannot adopt structured programming or object oriented programming or aspect oriented programming or functional programming or agile techniques or anything and expect to be ten times faster or a hundred times faster, no matter what people tell you:

There is no single development, in either technology or in management technique, that by itself promises even one order-of-magnitude improvement in productivity, in reliability, in simplicity.

It was true then and it is true now. And it will still be true when my kids are as old and grey as I am becoming. Worth reading and rereading, and not waiting twenty years between rereads. Oh and by the way, this is the same Fred Brooks who invented the heavens-I-wish-it-wasn't-true rule: Adding more people to a late project makes it later. That's from 1975 and you can't escape it either.

OK, there is one way you can achieve an order of magnitude improvement in productivity: hire the right people. The good ones are ten times as fast as the OK ones, and infinitely faster than the none-of-their-code-ever-ships ones. But that's not a technology or a management technique, so it doesn't count for our purposes.


Wednesday, 06 December 2006 20:10:13 (Eastern Standard Time, UTC-05:00)  #    Comments [0]