No son is fungible to his Mother! Or daughter to her Mother!


Of course not, but I think it's clear that he means that troops as troops are interchangeable.

And even this is false. SOF soldiers require a lot of training, have a lot of experience, and are probably smarter (on average) and more deadly (on average) than their colleagues in the wider body of military. There is a point to which this is true - when numbers count. Rumsfeld's goal is to have small teams with massive firepower backing them up. So, say 3 or 4 SOF guys can hide and lase targets on which bombs are dropped from afar - giving them the lethality of a much, MUCH larger force. A real example of this exact method is on of the operations in Afghanistan. However, this small force was surrounded by a few hundred local Afghanis on horseback - since 4 guys with rifles aren't much of a match for a company or battalion size force once they figure out where you are.

I recall I was talking to a former member of our board of trustees once - a retired Admiral who was also VP of a major s/w company. I was attempting to convey to him why some software projects fail and others succeed - in some cases, precisely because programmers are viewed as fungible when they are not. He cut me short and said something like, "Nonsense! You get a task, you estimate the man-hours, move so many programmers to the project, budget for overruns and move on!"

I was flabbergasted really. I held my tongue, but I wanted to explain to him that that's exactly why his firm's multimillion (at least 10s of millions invested and probably more) dollar program bit the dust while a similar project from another company - using older technology - had become the de facto standard. It's particularly important in the beginning of a project - to ensure a workable design. A small team of 5 to 10 particularly brilliant programmers is vastly better than a larger team (50 to 100) of mediocre programmers. In other fields, the top quartile of performers are about twice as good (in quantity AND quality) to the bottom quartile. In software, the top quartile of programmers is about 10 to 20 times as good (in quantity AND quality) to the bottom quartile. (Reference: Jakob Nielson's "Usability Engineering.")

And this is only talking about individuals as individuals - not the effect that these people have as members of a group. Some people are poisonous to groups and others just naturally make other people work better in their presence. (I think it was in Demarco and Lister where I saw an example of this, but I'm vague on it. There was the case of the woman who didn't seem to actually contribute anything concrete to any project she was on, but somehow the people in the project just seemed to get along better and met more of their goals than the project without her.) This is an extreme case, but I've witnessed lots of cases where the effects of individuals on group cohesion and task completion have been enormous - positively and negatively.

k