Tension Between Information and Action

I have some interesting scrawled notes from a train conversation with Richard earlier today. He was talking about software engineering management (don’t go away, I’m going to relate it to other things in a moment), and how there is a scale upon which you can slide too far one way and things go wrong. This is about varying how much managers tell their staff about what is happening:

Insufficient information given to workers < ----------> Too much information given to workers

The information here is detail about goals and motives, about why the software is to be like the spec/design. If you tell them too much about this, rather than just giving unjustified orders, then they spend ages worrying about it and don’t do their job. If you give too little, they will either not do what you say, or worse do exactly what you say, obeying your precise orders letter for letter. Either way the software won’t achieve the goals because the programmers can’t incorporate them in every creative decision they make at every level. Richard says this is what goes wrong with most software projects.

Here is another way of looking at the scale:

Hierarchy < ----------> Network

One way of analysing political power is to think of it as coming in three forms. These are hierarchies (think inside a multinational corporation or scientology), networks (think an old boys club of company directors, or a workers co-operative), markets (think buying vegetables from a choice of farmers). These all mesh in a great recursive tangle with each other, which I won’t get distracted into describing now.

Now, just ordering your workers about without explanation is an extreme form of hierarchy. Letting your workers know everything that you know, and letting them make all decisions is an extreme form of a network. The slider is the same, and it applies to any organisation.

It also applies inside you, yourself, and how you decide things:

Acting on gut instinct < ----------> Thinking hard, getting too much info

Some people are prone to over-analysis and never take any action. They are stuck too far to the right of the scale, investigating in detail the causes, theory, opinions on a political issue before actually doing anything about it. Of course, you can never fully understand everything, you are always riddled with doubt, and so you do nothing.

On the other hand, many unthinkingly sit on the left of our scale. They always do what comes naturally to them, so are over-exposed to, say, advertising and other manipulation by the society that they live in. They probably make good soldiers, or good hedonists.

There’s a sweet spot in the middle where you balance action with research. If you are working alone, you have to go further to the left or else you’d never get anything done. If you are a large corporation, you can go further right because of division of labour, and the extra benefits of scale when you do improve something by thinking harder.

Final amusing thing. Even the most junior programmer is really a manager whose staff are stuck infinitely far to the left on our scale. The computer obeys their exact orders, taking no account of the programmer’s goals, to the absurd extent of crashing and failing in a way even the most obedient soldier never would. When you manage people, you are working more over to the right. Perhaps this partly explains why programmers who become managers instinctively manage without giving information about goals, as that is how they are used to “managing” the software that they write.

Software is totally obedient < ----------> People think about what you tell them

Leave a Reply

Your email address will not be published.