Thread by @swardley

X : Why do you need a system of theft with pioneers, settlers and town planners (PST)?
Me : Gosh, complex. Ok, let us start with PST overlaid (A) onto a map and I'll just add a line for the evolution of computing infrastructure from genesis to utility.

Me : Now, lets add some illustrative examples (A). I'm not going to go through putting these in the exact places, just purely for illustration but it'll do. Even imperfect maps can be useful to communicate concepts.

X : Are you saying the pioneering of computing infrastructure was long ago?
Me : In the concept? Yes. Of course there are many sub components and even below electricity provision (which is commodity) there is pioneering going on in generation ... Dyson sphere anyone?

Me : Now, the evolution axis (x) comes from the evolution curve. So, let us add the points onto that.
X : Is that a diffusion curve?
Me : No, you can't measure evolution over time. In fact, each node actually does goes through a diffusion curve ...

.. which looks like this. Remember this is illustrative. Each point goes through a diffusion curve and so there are many diffusion curves (A) in the evolution of an activity, each going to 100% adoption of its applicable market (B).

Me : To make my life easy, let us take two of those diffusion curves, add the various social groups (A) such as early adopters and laggards as per Rogers.

Me : Now, now let us add Moore's chasm (A) in roughly the right place. What we have is two chasms and two diffusion curves in the evolution of the act of computing infrastructure. In fact, we only have two because I ignored the other examples.

In practice ...

Me : In the evolution of a single act you will have many diffusion curves and many chasms.
X : Are they all the same?
Me : Good question. In my experience, no. To explain ...

Me : When we cross stages e.g. custom built to product or product to more commodity / utility rather than product to product then we seem to experience significant barriers. I refer to these as points of inertia.
X : Why?
Me : Hmmm, complex but without invoking co-evolution ...

Me : .... the problem is we build capital in the previous way of doing things. It's this capital which make it difficult to jump stages (A) which is why you need to steal it away from those people (B) ...

Me : Hence settlers have to steal from pioneers and town planners have to steal from settler.
X : Steal is a harsh word.
Me : I'm not very good with words but it reflects that it has to be taken, a pull mechanism rather than a push mechanism.

Me : Now, to complicate matters, the amount of capital we build up seems to vary with evolution i.e. inertia seems stronger for a product to commodity / utility shift than a custom to product shift. There are 16 different forms of inertia by the way and some of these vary.

X : 16?
Me : Well, that I've been aware of for quite sometime. There are probably more. The killers at the product to commodity transition tends to be "changes to government, management and practices" and "political capital".

X : Is PST recursive?
Me : No, but something is.
X : How do you mean?
Me : Ok, first it's worth remembering I stopped research on organisation in 2007 -

https://twitter.com/swardley/status/1087511545091899392

- because, well you can only research on an organisation you're experimenting on ...

Me : Now, if I look at the characteristics of PST they are quite distinct but as @b6n and others argued with me a decade or so ago, there is something recursive in there.
X : And?
Me : And what? I can take a guess, but no more ...

Me : Let us go back to our evolution curve and take one example, the IBM 650. One of the first "product" like things ...

Me : Let us take the diffusion curve and rather than looking at the time for the entire activity and the applicable market, let us expand this to ...

... just the adoption of that one example and the time it takes to diffuse i.e. just like a normal Rogers curve.

X : And?
Me : Well, this is a diffusion curve for a product. I'm going to stretch here widely. Let's add @KentBeck 3X's to it because I quite like that idea ... so we've got explore, expand, extract.

Me : Let us take this back to a map. We've got 3X's around one diffusion curve and its chasm. Let us leap wildly and say you get a 3X around every diffusion curve ...

... that would mean Pioneers have an "explore" phase as well as Settlers. But they are very different i.e. Pioneering is exploring the truly unknown (A). By the time you're on the path towards a product (B) you have an MVP and users you can talk to (from the custom built world).

Me : If we add this all together, you'd have an evolving act with many diffusion curves and chasms, specific points of inertia, an attitude shift from pioneers to town planner plus some sort of recursive 3X of explore, expand and extract.
X : Is this right?
Me : It's a guess.

X : Could you test it?
Me : I don't run companies these days. But, I would look for artefacts which might occur from this. There is an "obvious" one which means it's almost certainly wrong and therefore makes me think this guess is wide of the mark.
X : What's that?

Me : Well, in the extract phase (and there are three), you would expect in the Pioneer stage for big consultancies to arise, in the Settler stage for rent extraction to occur and in the Town Planner stage a tendency for monopoly utilities ...

... alas, that seems sort of right which means it's almost certainly wrong.

My general rule of thumb is the obvious solution always turns out to be obviously wrong at some point in the near future. You'd need a lot of experiments to test whether recursive is useful.

X : I assume that means you don't know?
Me : It's a guess, that's all. You'll have to experiment to find out.
X : On what?
Me : Your organisation.
X : I'm not happy about that. Can't you research it?
Me : Sure, give me control of your organisation and I'll find out.