Finally I can write code faster than LLM! /s
replies(1):
BTW I hate that term with a passion.
As someone who has crafted software for the last 45 years, it’s like implying that learning the alphabet makes you a Tom Robbins.
Writing software is about understanding the problem and the information involved in such a way that you can craft algorithms and data structures to efficiently and reliably solve those problems. Putting that into code is the smallest, least important part of that skill.
That’s why software engineering is mostly language agnostic. Sure, there are paradigms that fundamentally change the way that the problem space is handled, but at least within a paradigm, languages are pretty much interchangeable. Some just have less hidden footguns.
Interface design is another thing altogether, And is either fruitless drudgery or fine art, depending on your predisposition.
There is definitely room for a subclass of brilliant interface designers that do not need deep data manipulation intuition …. But they do need to have a deep understanding of human nature, aesthetic principles, color theory, some understanding of eye mechanics and functional / perception limitations, accessibility/disability engineering, and a good dose of intuition and imagination.
In short, nothing about producing quality software is something that you gain by writing code snippets to solve simple problems. But you do have to learn the alphabet to write, and “coding” it is still a prerequisite to learning to write software. It just shouldn’t be sold to people as if it was some kind of degree lol.
Give me some kid who’s been building things with Arduino in her basement for a few years over a coding bootcamp graduate any day. I can teach her how to write good code. I’ll just pull out more of my non-existent hair trying to reach the “coder” to actually solve problems, unless I get lucky.
I've seen people who are really great at combining "recipes" off the web for anything (including hobby electronics and programming), but never really get to the bottom of things or develop clear understanding of how things work and tie together.
I imagine you'd only get more out of that kid toying with Arduino because of persistence ("few years"), and not because of the type of things they did, but I ultimately believe you'll have similar chances of developing a great software engineer out of any of them in general.
You can start anywhere, and coding boot camps are useful, just as following YouTube tutorials. But until you learn to identify, quantify, and characterise the problem and data space you aren’t really doing the job of software engineering.
My experience is that many people are deceived into thinking that language fluency is the core skill of software engineering, and coding bootcamps tend to foster that misrepresentation.
That doesn’t make them bad. It just means that often, thrashing around with no real knowledge of the tools and solving a problem with the tiny set of syntax you can get to work is much, much more educational towards the goal of becoming a software engineer than getting a good grasp of the language as it pertains to solving toy problems that require little effort to characterise.
Anyone that is willing to hack around a problem long enough that the solution emerges is doing the real job.
It doesn’t matter where they start, or how much they “know” about “coding”. The real task is to fully characterise the problem and data structures, and sometimes that emerges from the horrific beast that rises from the tangled mess of a hacking session, a malformed but complete caricature of the actual solution.
Then, all you have left to do is to code the actual solution, the structure of the algorithm and data has been laid bare for you to see by your efforts.
That, I believe, is the essence of software engineering.
I've obviously seen people who misjudge this (they can code, hire them), but ultimately, developing someone requires an amenable minds of both a mentor and a mentee on top of talent and persistence.