Thanks for your interest, Dan. What you are talking about here are existing software frameworks I am not trying to reinvent: from very low level (programming language constructs) comparable to nuts, bolts, and hoses, to mid-level (API calls) comparable to simple assemblies like water pumps or shocks, to high-level (frameworks like Spring, caches, and databases) comparable to the complete engine. Anything beyond that is a car/truck, and cluelessly bolting two trucks (or even engines) together to create a bigger truck or engine doesn’t work. There are no blanked statements e.g. “componentization” in any engineering whether mechanical or software. The right size as you put it becomes art.

I am currently back to the corporate world (gotta make a living), and am witnessing everything I described: worse than ever. Gartner/Forrester based decisions to buy IBM or Oracle, naive hope to rewrite billions of lines of convoluted COBOL code in allegedly “almost DYI” “business rules”, a new name for another myth: “customizable”: “low code/no code”, and last, but not least RPA, that gives “robots” a bad name (just like it happened with AI): zip-typing and duct-taping amateur outdated applications on an industrial basis. Lots of clueless MBA hopes based on Gartner “magic quadrant” assurances. In the end there are no guessing or “almost” in engineering. It is precise. Leave generalization about components, performance, and tools/languages to those two “executive” consulting firms, along with equally clueless Forbes.