The box-and-arrows paradigm for systems, built in the 50s and enjoying popularity briefly in the 80s, is overrated, and has been outmoded by the likes of complexity theory. This is due to the fact that box-and-arrows systems like those made by Club of Rome to predict civilizational collapse carry strong assumptions as to the nature and structure of underlying variables and as such become very brittle as the size of the system scales. The norm is not the closed-loop circuit models that initially inspired systems thinking, but open-loop energetic models where any structural element is more like a rarified pattern than an ontological atom.

The result is a discipline that has transformed into managing uncertain outcomes in large heterogeneous models, i.e. complexity theory, rather than reducing everything to balls-and-sticks. Meadows was famous for devising "12 basic places to intervene in a system", nowadays the focus is on hedging bets adequately such that interventions don't catastrophically fuck up.

That said, some of the basic tooling is still flexible enough for basic business problems and some of the old gems are able to explain important concepts found in other fields without getting bogged down in the math.

https://www.amazon.com/Early-Retirement-Extreme-Philosophica... is my favourite, it's not about retirement, it's about using systems thinking to devise a robust lifestyle.

https://www.amazon.com/Introduction-General-Systems-Thinking... will make a good complement to Meadows and should give you a calculus to rigorously think of systems with.

https://www.amazon.com/Introduction-Cybernetics-W-Ross-Ashby... for its explanation on entropy, I mean requisite diversity, which will you give you an approximate mental quantity of how "powerful" any given system is.

https://www.amazon.com/Sciences-Artificial-Herbert-Simon/dp/... and https://www.amazon.com/Design-Rules-Vol-Power-Modularity/dp/... I haven't read either of these, but Herb Simon is extremely influential and has great thoughts on the notion of system hierarchies (nearly-decomposable systems is a great concept for design). The second book is about the properties of modular systems, which will help grok the reasoning behind a lot of refactoring techniques.

Good luck.

I would not say that we are not doing science.

I would rather say we are doing science of the artificial---which looks a lot like design.

See the Sciences of the Artificial by Herb Simon (a true legend):

http://www.amazon.com/The-Sciences-Artificial-3rd-Edition/dp...

Both inspiring:

Doug Engelbart's 1962 "Augmenting Human Intellect: A Conceptual Framework"

http://www.invisiblerevolution.net/engelbart/62_paper_top.ht...

Herbert Simon's "The Sciences of the Artificial"

http://www.amazon.com/Sciences-Artificial-Herbert-Simon/dp/0...