[mhh__]

I'm not overly familiar with rust's documentation in particular but in my limited experience of teaching programming to a few unfortunate souls, the issue is usually figuring out what framing device to use.

The actual level of English in good documentation is usually fairly "low" - i.e. keep it simple, assume your reader wants to learn (Too many people suffer from I'll let you google that for me). That and a lot of non-native speakers probably speak better English than me (I know more words but I'm lazy)

I know how to program, I've been bitten by memory bugs - rust's design makes sense to me (i.e. the borrow checker is relatively intuitive because I've written and worked on compiler/s). That is difficult to convey in 10 minutes to someone who only knows how to plot a graph in python (for example).

My approach is usually to state all my assumptions as I go, while naming as much as I can as I go (e.g. googling for operators can be very difficult) when explaining things. In writing, I find that writing documentation away from the code can be quite fruitful otherwise the overall structure and meter of the code can be hard to divine.

[bscphil]

> I know how to program, I've been bitten by memory bugs - rust's design makes sense to me (i.e. the borrow checker is relatively intuitive because I've written and worked on compiler/s). That is difficult to convey in 10 minutes to someone who only knows how to plot a graph in python (for example).

I think you're getting at an important issue here. I've been programming (as a hobbyist, not a professional) for over a decade, have contributed to a bunch of open source projects, written software of my own. Mostly in languages like Python, but I've dabbled with C, Rust, etc. The stuff about lifetimes in this guide is still borderline incomprehensible to me. The reason for that is that the basic issue here isn't the difficulty of the language used to describe the concept, but that the concept itself is complicated. Maybe it needs to be complicated, but I don't think there's any way to distract from the fact that Rust is a very hard language to learn.

[da39a3ee]

> someone who only knows how to plot a graph in python (for example).

Hey, I've been writing Python daily for 12 years and I struggle with that!

First you have to get matplotlib installed. That often works nowadays, but you have to be sure you've installed numpy first, and on a bad day you'll find yourself looking at some sort of C compiler error and copying stuff from stackoverflow to make libgfortran be found or whatever.

Then you try to remember the bizarre way people import matplotlib. Your mind briefly flits to memories of blog posts explaining that matplotlib has two parallel APIs, and you start feeling annoyed about that. Then you start feeling annoyed that one of the APIs is based on matlab; why should a key part of the python ecosystem be apeing matlab? It's not like anyone has ever thought matlab had APIs, or anything really, worth emulating. Anyway, it's something like `from matplotlib.pyplot import plt`, or some nonsense like that. So you spend a few minutes failing to remember, and then googling for it with a rising sense of frustration.

Then you get this error:

_RuntimeError: Python is not installed as a framework. The Mac OS X backend will not be able to function correctly if Python is not installed as a framework. See the Python documentation for more information on installing Python as a framework on Mac OS X. Please either reinstall Python as a framework, or try one of the other backends._

And now you're really quite annoyed. What the fuck is a framework? It's some MacOS thing but it's always been clear that it's a non-UNIXy concept that you don't want to and shouldn't need to know and you kind of suspect it shouldn't even exist. But holding that opinion would require finding out what it is.

So you google for how to overcome that. And then... there's some weird thing where you have to put a line in the middle of your inport statements before importing matplotlib... and finally you wish you hadn't been so stubborn and had just used an IPython notebook, but you just kind of hate them despite how beautiful and well-engineered they are. They can't be version-controlled, you hate typing code into your browser, you always end up with an incomprehensible mess of mutable state, etc.

I used to use R and, back then, I enjoyed making plots and did so efficiently and reasonably well.

[mhh__]

Ignorance is bliss when you can just copy and paste example code and swap out the plotted function (is what I was alluding to)