You’re certainly right about one thing: there are plenty of strong feelings about methodologies.

I’m not sure that programming will ever have the kind of failure rates that you see in construction. After all, construction failures are minimized by building the same thing over and over again, and honing the process over time. In software, you (ideally) never create the same code twice. The second time, it is abstracted out, and you never have to write it again. As a result, we get much greater reusability gains than construction, but we also don’t get any of the reliability that comes along with repetition.

In the end, getting good coders in an environment where they can communicate and think clearly is FAR more important than any methodology. For all the hype, no methodology has ever accounted for more than a 5-35% gain in productivity, and that eventual gain is only *after* a temporary 10-50% *decrease* in productivity while everyone’s getting used to the new process. That’s still usually worth it, long term. However, “order of magnitude” increases in productivity (that is, by a factor of 10, or 1000%) are simply fantasy.

When agile fails to deliver magnificent gains, it’s common to blame the people, not the process. That is actually reasonable. For *any* change in productivity, positive or negative, the people doing the work are the single most significant factor. Successful management in software is 90% picking a good team, and 10% removing obstacles.

I will say, though, posts about Agile Scrum generate the most traffic and the most interesting comments. I should write about it more often!

I don’t think any of the leaders in the Agile community have ever suggested Agile is a magic bullet.

Sorry if the reference to the famous Brooks article was a bit obscure. (It would have helped if I’d used the correct title: it’s a silver bullet, not necessarily magical.)

I’ve been told by an Agile consultant that Agile methodologies “regularly result in an order-of-magnitude improvement in performance.” It’s not hard to find others saying this. The definition of “silver/magic bullet”, in my mind, is an order-of-magnitude increase in productivity. Agile is neither a magic nor silver bullet. The “order of magnitude” claim is 100% weapons-grade balonium, designed to sell books and put asses in seats at lectures and conferences.

I also find that your relations of degrees to the positions is exagerrated, at least in my experience, many years in the Java-based dev environment, degrees mean pretty much squat.

Way to pick up a non sequitur and turn it into a straw-man. I’ve said before on this blog that degrees don’t matter as much as merit when it comes to programming. I presume that you’re referring to this line: Trust me, if they wanted to be project managers, they wouldn’t have degrees in computer science.

I’m not saying that all programmers need to have degrees in computer science. For example, I don’t technically have a degree in anything, and it’s never been a problem. Quality output is the best credential in software, and the only one that matters. (I’ve met some real bone-heads with advanced CS degrees and no understanding of the art and science of software development.)

However, I’m sure that you’ll agree that you’ll find a higher proportion of CS grads in engineering than you will in project management. The point is, most programmers didn’t become programmers because they like meetings and timeline estimation. In fact, most programmers despise these activities as boring busy-work. They’d rather be coding. They are more useful when they’re coding. So let them get to work.