Thread by @HeyChelseaTroy

I once wrote a blog series about how I level up as a programmer. Now I'm also a teacher in a CS program.

My advice in that series and my practices as a teacher both differ from what the "top" programmers say they did to learn.

Let's talk about why. 1/

chelseatroy.com/category/level…

The first oft-mentioned plot hole in "asking top people how they did it" is survivorship bias.

I.e., for every person who succeeded by doing X, there are 999 people who failed while ALSO doing X. The secret sauce wasn't X. This is true. There's another thing at play though.

2/

Here it is: the things that people at level n of advancement do to get to level n+1 might be different—and in fact, even the exact opposite—of the things that people at level n-k need to do.

I think about this a fair amount at athletic competitions. When I'm competing...

3/

THE MOST IMPORTANT nutrition consideration for me, outside of hydration, is glucose availability. Reaching my redline intensity requires access to glucose.

So what do I eat at functional fitness competitions? Gummy bears. Candied kiwi. Some athletes chug sugar drinks.

4/

This is, like, LITERALLY the exact opposite of what a personal trainer would tell MOST people who are trying to raise their fitness level to do.

And yet, if you interview me like "how did you get more competitive at fitness," the answer that will pop to mind for me is...

5/

1. I focused on endurance work on advanced skills in the ranges of motion that are hardest for me
2. I learned how to fuel myself on competition day

What I leave OUT is "Well, first I spent 5 years building a cardio and strength base and ignoring advanced skills entirely"

6/

But if someone who is new to fitness wanted to do what I do, they need to start with that. NOT with what I do now.

It is the same, I think, with building programming skills.

If a homework problem is an "interesting challenge" for me, it is a BAD FIT for my students.

7/

My students need opportunities to learn basics. They need opportunities to read and modify other people's code. They need this MORE than they need opportunities to WRITE greenfield code, even, in my view.

My class is the first time a lot of students get practice doing that.

8/

(btw if you're interested in my approach to teaching, I write about that too)

9/

chelseatroy.com/category/teach…

Similarly, when I was first getting the reps in to build my strength at reading and analyzing code, I did a ton of commit tracing.

This technique is all reading and looking things up. That writing? It's copied-down code that I didn't write.

10/

chelseatroy.com/2018/07/21/lev…

This is one reason that I recommend most folks DON'T look to the stars in their field as mentors.

I recommend they look to people a couple levels ahead of them, who are, among other things, more likely to remember what THEY should do.

11/

chelseatroy.com/2019/03/21/lev…

I think it's worth considering expert advice in the context of the prerequisites that the advice might require to...well, to work.

Where those prerequisite skills came from is often misrepresented in, or omitted from, the conversation.

This is kinda why...

12/

...I make a point to write as I learn. Because 10 years from now, when I'm good at this, my memory will not accurately represent how I got there.

But my writing FROM THAT TIME will represent that journey much more accurately and usefully for folks who come after me.

13/13