How to learn to code if you're not sure where to begin

As someone learning web-development to build the product I want, while continuing my unrelated non-tech day-job, my key learning has been:

1. Very ambitious plans are detrimental to progress. The problem with those is you'll inevitably fall short one of these days, and maybe a few days at a stretch. When that happens, you should be OK with cutting yourself some slack - because that's the only way you'll get right back to it.

2. Don't make a 3-month learning plan. One, it won't mean anything. Have a vague sense of where you want to be in 3 months, but have a 1 week and a 1 month plan.

3. To the extent possible, stay loyal to your 1 week plan. There are a lot of distractions on the way (let me learn this alternative stack to do what I want done, let me set aside my current online tutorial course for this other one that looks like is getting more love, etc.) but staying the course - at least on the 1 week plan- will push you forward.

4. Conversely, don't be too wedded to your 1 month plan. Things will change in a week from now - maybe you'll really find a better way of doing something, maybe you'll realize that last 1 week only got you 30% of the way you wanted to come in that week - so you will necessarily have to recalibrate. That said, you should always have the 1 month plan, but it's function is provide context to your weekly plan more than anything else.

5. Stack: if you already know something, just get started with it. Among the major backend stacks (JS, Ruby, PHP, C#, Java) no stack is really leaps and bounds better than the other: each one has sufficiently big communities to get your questions answered. On the front-end, you do need JS, but again depending on your use-case you don't need to go all-out React/Angular/Vue. For a surprisingly large number use-cases, plain old JQuery will be just fine. If your backend stack is C#, there are productivity benefits to using Blazor (C#) on the front end as well, so you can reuse your backend code on the front end, and don't need JS for the most part.

6. Don't wait to first learn everything from courses and tutorials (authentication, APIs, hosting, 3 kinds of databases, user roles, the list never ends) before starting out to build your project/tool/app. That will keep you from starting for the longest time, but worse, is also not the best way to learn. By all means get to a state where you can write basic CRUD programs, but then get started actually building your product while learning the topics that you need, when you need them.

I found the article pretty empty.

Here is the plan I would give someone if they want to learn to code but don't know where to start.

1. Learn Javascript. It runs everywhere on the web and you can publish to the whole world for free.
2. Pick something you want to build. Cut the idea in half. If you want to build a golf app, cut that in half to be a page that displays golf scores.
3. Cut that 2nd idea in half. If you want a page that displays golf scores, cut it down to display 1 golf score.
4. Figure out what you need to do. aka Display a golf score.
5. Google how to do it 'how to display number on web page' etc.
6. Keep googling bits and pieces until you figure out how it works.
7. Ask for help if you get stuck longer than a day.
8. Keep doing that, building little things you want.

As a software engineer, I agree with all these points. I have a degree but have found that over time it's not something employers ever really look for. They are more interested in the latest projects I've been working on. Practical knowledge and application seem to be more important than theoretical knowledge and a piece of paper, in my experience, anyway.

This is really helpful. I've always hired developers, but learning to code is always at the back of my mind. You make some really good points regarding knowing what languages you need to know and how a degree isn't always the best option. You've confirmed what I've thought for a while. Thanks