Learning Software Architecture

I think of it this way. Software architecture can not be taught. (This is not particularly uncommon, and so not a claim to being particularly special.) But you can sensitize people to the problems they will encounter and solutions they should consider. You can't "teach" someone into being a software architect of 10 "software architecture goodness units", but you can accelerate the rate at which they accumulate SAGUs by doing this.

In fact, per my parethetical, honestly I would say most things work like this. You can't just pluck people up and drop them five or ten years down the road of experience for nearly anything. What you can do is give them a car and show them the road rather than letting them hack through a jungle with a machete and no map.

One of my major criticisms of how we teach is that for the most part everyone in the system operates in the belief that we can just stand in front of people and say things for a few hundred hours and they will be magically equipped with completed and refined skills after that process. Too much focus on trying to drop people into some mythical state of "educated" rather than accelerating them through their own journey. A journey that can be highly guided and structured for sure, but where you understand that for all you can clear the road you can't actually take steps for the student.

"It is also subtraction: removing unnecessary abstraction, ceremony, cleverness, and control."

This reminds me of one of the most distinct menaces to software architecture I encounter on a regular basis, the person who follows Hexagonal Design or Clean Architecture or MVC or some other Answer To All Architecture, rigorously following it to the letter and insisting that any system that doesn't follow the Answer To All Architecture is prima facie a bad system that needs immediate redesign to bring it into conformance with the ATAA.

The "answer" to something as complicated as software architecture doesn't fit in a book. In fact it doesn't even fit in a human head; for all the experience I have in network servers I wouldn't expect to be able to sit down and architect a high-performance game engine that could compete in the marketplace. I have ideas about how to do it, sure, but I know that I don't know all the details that people who operate in that space do, and plenty of those details will matter. No one has the breadth of experience to truly tell us all how to architect software, under all circumstances and conditions. Anyone de facto operating as an architect needs to bear that in mind, and while consulting the prior experience is part of that "be in a car on a road and not hacking through a jungle", the supposed set of blessed answers must always be mediated through the task in front of you and not vice versa.

Thanks for sharing! Really agree with #2 even though ultimately - we can only minimise rather than totally eliminate.

> I would like to spend my time more on gaining a mental model of the projects I work on, but I get very demotivated if I start disliking things like the programming language, certain arch. Choices or anything that gets too complex that doesn't seem like its worth my time

Let it go...

Not one single project is free of warts. Not a one. And honestly, if the programming language is that big a deal, you're better off jumping ship. We should all be polyglots, but that's up to you.

This guy knows exactly what he's talking about. However, it feels more like a "if you know, you know" rather than relaying the experience/data he's accumulated.

Let me just say: A lot of people think architecture is how do you build a very complex system with tons of moving components (databases, queues, scaling, reduandancy, failover, dozens of services). I think expert achitecture is being able to answer how do I correctly solve the problem using the fewest of those.

Genius-level engineering is inventing the zipper, usable anywhere, lightweight, cheap, sturdy, simple to make.

(I believe he's alluding to this by saying companies follow conways laws and essentially socially construct a lot of work to match up with the number of teams, hence creating a completely unnecessary nightmare of complexity just for the social incentives)

You can make the feature with one colleague, or you can call in five more people to weigh in and do their parts of the work.

If you involve five more, the result could be better in theory, but it will certainly take MUCH more time because of communication overhead.