Is it just me or is this an unviable learning strategy? My approach for learning anything has always been to follow small tutorials, build an intuition, and only consult docs when I need some specifics.
It's only on the second pass that I am trying to go through each section carefully and make sure I really understand before moving on, including seeking outside help or resources if I feel confused or stuck.
This feels fun for me, and the casual first pass makes it easy to figure out if a book or language truly appeals to me.
I also feel, strange as it sounds, like for me it save time compared to learning in small increments through tutorials. It lets me more quickly absorb the basics for things that are already more or less familiar, and then I can focus on exercises and examples only for the tricky stuff.
When I first started studying computer science, in high school, the biggest productivity gap was between the students who tried to work only with what they were directly taught in class by the teacher and the students who decided to go explore the language/stdlib API docs on their own. There was a lot of 'wow, how did you do that!?' from the former group and a lot of 'it's built in, check out this part of the manual' in response from the latter. But somehow no amount of exchanges like that could convince the former group to take some time to RTFM in a comprehensive or unguided way, so it stayed that way.
The docs about the concepts are better, those are worth reading.
For example, a lot of C++ programmers miss the full spectrum of std::algorithm. A lot of Python+NumPy programmers miss out on some useful indexing tricks.
I'm still suffering from not taking a rigorous approach to unix shell.
For me, the ideal thing is a book, and great docs can feel like this. Something well thought out, comprehensive, deep, and importantly error checked/reviewed. The downside is books are increasingly rare and get outdated really quickly.
Some of it is a matter of it being hard to even have the language to know how to ask the questions you have, so a bootstrap or quick start to let you just get in there and start looking around helps a ton and makes the full documentation a lot more accessible.
While I basically tend to learn like this - (solve my current problem and move on, since I have limited time to spend on things) - I find that this type of learning can miss some things. Like when the documentation says this is important etc and you never know about it/find out about it the hard way when your stuff doesn't work properly.
- read high level architecture docs, - try to do some small coding - get confused and google until I get my goal done - go back to reading high level architecture docs
You're (or at least, I will) never absorb and become productive in all that material in a first pass. But at the same time reading the intro for the mentality of how the big picture works helps a lot to scope into what to focus on.
Nonetheless for just getting to your goal and not becoming really good in something this is not necessary or can be done later