A house comprises many parts. A custom motorcycle has many parts. These things are designed and built to established practices, but in 99% of cases they aren't really "engineered" with calculations and formalised decision-making processes and the like.

Thanks for bringing up the PE distinction. This kind of argument comes up here ad nauseum. People have different views and we argue semantics.

There is engineering as a discipline (e.g., mechanical, electrical, industrial, civil, aerospace...etc.) and engineering in a more colloquial sense like "I did some hillbilly engineering and rigged up a pea sheller". Both involve solving problems.

The former typically involves the person going through a rigorous 4-5 year program where they take a bunch of advanced math and physics classes and a ton of other classes (including coding and robotics as well as more specific classes such as power electronics) that teach logical thinking and problem solving. I'm always amazed that I can more or less talk to any engineer across the country (different state, different university, different engineering discipline) and still recognize their training and that hyper analytical mindset and feel the kinship. That of course doesn't mean that there aren't similar analytical people that went in other directions, but the training is a real thing.

A professional engineer passed their fundamentals of engineering exam, got their degree, and then took their professional engineering exam. You used to have to get 4 years of work experience before applying, but I think they got rid of that. It's an awful 8-hour exam with a pretty low pass rate. Honestly though, the PE usually just means the engineer was willing to spend 3-6 months prepping for the exam and has a need for the license to stamp papers. A lot of engineers never get it if they don't need it. I believe you need it in some cases for court cases.

As far as software engineering goes. It seems to be a discipline that matches the second colloquial definition more in a lot of cases. In some cases they have actual software engineering degrees where the student takes all the standard intro engineering courses (calculus, linear algebra, statistics, differential equations, statics, dynamics, circuits, thermodynamics) before doing the core classes distinct to that major. In that case, I'd call them an engineer due to the training and think it's an appropriate job title even if they get a job slinging corporate java, which would not typically be viewed as traditional engineering. So they're an engineer...not necessarily doing engineering. Does that really make sense? My own definition breaks down if we apply this to the same engineer becoming a garbage man. They're still an engineer by training, but they're most definitely not doing engineering anymore.

Also, if Pete gets a job slinging code out of highschool...we can say he's doing engineering in the colloquial second sense, but is he an engineer in the first sense? No.

Clear as mud? As an engineer I now must flog myself for my own internal inconsistencies. In all seriousness, language is hard and sometimes imprecise. I didn't realize my rambling got so long and apologize.

That's where engineering fits in. It's in the knowledge and information space, you make things like plans, arguments, decisions, documents.

That's not to say it all needs to be written down; someone can do the "engineering" of a small project all in their own head and then implement it themselves.

The engineering process itself is a form of knowledge work. A single person can do both tasks of course.

I actually read an interesting study, probably here, about software engineers. It turns out that for as much emphasis as we put on mathematics we actually tend to be better at language, and that even our penchant for maths is less in the way of calculation and more in the way of definition and expression (which is to say, making it the language numbers speak and learning it that way).

You mean communication, the transfer of ideas. Already hard between humans, machine language and operation is completely foreign.

I'm thinking things like: do you know what's the latency and throughput of your system, and how it is affected by different conditions? Based on this, people who work on real-time systems (including video games) would be the closest thing to the engineers of the software world.

CS proves that a binary tree is faster asymptotically, and that’s still true and very much of practical relevance regardless of the CPU you use.

No, they don't. A vector has faster access than a binary tree, O(1) vs. O(log n). Maybe read some of those expensive books.

If you talk about editing the data structure, then it depends.

And plumbing is a highly skilled and essential profession. And a lot of plumbing work involves engineering to a greater or lesser extent.