But the real challenge comes when tackling the core of the legacy, especially the data, that has many dependencies and is hard to untangle.

And picking too many quick wins can increase overall complexity that makes untangling the core even harder.

2/4

Hypothesis: Because "sociotechnical architecture" is necessary, but no way near sufficient? That sociotechnical is about the social being at least equally important as the technical. Conway is just a signal, an eye-opener to the real problem that is joint optimisation of the two.

Hypothesis: Because "sociotechnical architecture" is necessary, but no way near sufficient? That sociotechnical is about the social being at least equally important as the technical. Conway is just a signal, an eye-opener to the real problem that is joint optimisation of the two.

“Anarchy doesn’t mean you can’t organise, but it should be done voluntary, functional, temporal and small”

Really do think it is risky to use that model to implement our IT systems, an the other way around, that we limit our framing to what we can do in code. What we implement is a reductionistic version of it, a closed system. The model we create for understanding should encapsulate the complexity.

It's also that people seem to read weird things into it, like all models are equally wrong, or all models are equally useful, or all models are useful on the same parameters, or that models are uniformly wrong or useful irrespective of context. I could go on.

Am yes, one model require Contextualism as world view, while the implementation can use Formism or Mechanism.

Really do think it is risky to use that model to implement our IT systems, an the other way around, that we limit our framing to what we can do in code. What we implement is a reductionistic version of it, a closed system. The model we create for understanding should encapsulate the complexity.

Said something along these lines in my talk at @dddeu.bsky.social and @virtualddd.com. I think we should treat modeling as we ought to treat systems thinking, namely as an epistemological tool. A way of framing things, making some sense of complexity.

Mostly my fault that we didn't though, as I was stuck at work most of the time. Next time!