Ask HN: How would you teach teens to code in 2019?

Teens/Childs quickly loose interest in something. Find out what they find "hip" or cool, don't expect it to be what you find cool. I would rather support them to program their own Instagram app then to propose them to some hardware programming environment (thinking of educational boards) if they don't like it.

Also, learning programming was applied autodidactics for most of us, isn't it? My feeling is that this is the only really working way of learning programming (have worked 10y in university, teaching programming: Doesn't work well).

It really depends on your kids and on your relationship to them. Maybe you should present them a vast amount of programmable devices and interfaces to explore together interesting projects.

If they have interest in it, by all means. Programming is a very slow learning curve, which can be frustrating for most people. I would personally start with very highly specialised / abstracted programming environments. If they like video games, using things like construct or, harder, Unity might be a good inbetween. Because of scenes editor and such, you can create something nice just with a series of drag and drop, and just write a few simple scripts to make it a bit more alive. It's also good to start on existing tutorial and try to customise them.

If they aren't interested in video games, it might be harder to sell programming as a hobby. I guess there is something a bit more magical to "hey, I managed to make this cube move when you press the joystick, and then when you press A it jumps" vs "hey, I have a sign-in form for my app".

Well it is difficult because there are so many distracting things and programming environments are not very engaging for kids. My two daughters had great fun assembling Lego Mindstorms and Robotis Bioloid kids (with the smallish metal screws). We then downloaded the examples programs to which I did a few changes in front of them, and tried to explain stuff. But clearly they showed no interest in the programming activity. So I did not insist. I may try some time later with an Arduino: I think it is by far the less sucking programming environment for young people, and it basically works (and is text based).

Do they show an interest? Same as playing an instrument you can get them one, teach them a few tricks, send them to classes and then, well, they either want to pursue it or not.

Agreed that they should show an interest before you embark on this. Finding joy in general education is more important than job skills for that age.

That said, Ruby and JS are not great languages for teaching fundamentals. They (especially JS) have a lot of weird design decisions that don't translate to other languages and don't map neatly onto problem spaces.

If you insist on JS, start with TypeScript at least. But maybe D, Kotlin, Go, or Dart would be better to teach general concepts.

Python would be a good add-on because they could actually use it to write quick utility scripts and, eventually, data analysis.

Instead of linear programming exercises, it may be easier to find problems they find interesting and help them Google their way through solving them. Maybe they have some tedious homework they could learn to automate.

The problem you'll find is that kids don't do much work, so programming (i.e. making a computer do work for you) isn't necessarily that exciting yet.

Personally, I would think twice and instead help them find what they are good at (or have interest in). Future will include content creation of sorts.

Like the rest, if they are interested. I think, again like others, that something with better fundamentals than JS would be good as then JS won't be hard to take up. I learnt coding in the early 80s by programming games. But when it came to making money I used those skills to build business applications. And I liked that as well (I like the feeling of creating something that works 'out of nothing').

yes, definitely. My father teaching me to code as a kid was a very happy time!

Today, I would start by playing a bit with the browser developer tools, and then move straight to shadertoy to produce cool opengl demos.

code for code's sake is usually boring. using code to do something you already want to do can be fun.

so start by finding a project. even better if they find the project on their own...

You should differentiate between the different aspects programming involves. I sure do see logical thinking as the most important aspect of development. Also Algebra is important, but well above the understanding of a younger child. If you look for example into Game Design you get a lot bigger area of interest that is combinable with programming, such as UI, Art, 3d, Sound, Story, Mechanics and Level-Design, etc.

Getting them directly into programming directly will likely be the wrong attempt, because (in my opinion) programming requires a lot of frustration tolerance and if done the wrong way will spoil it for them. Also combining an existing interest with development, such as using their painted characters in a game might create interest in content creation.

In our Game Design BA, we also give a summer school course for kids (12-18 years), that teaches them to develop a game in GameMaker during 2 weeks. Other visual development systems such as GameMaker might be good as well. Afaik lego also has a visual programming system for it's mindstorms system.

If you you insist on classical programming languages i'd suggest python with some ui based libraries (for example pygame, not because of gaming, but because of graphics, sound and interaction aspects) or processing.

I also consider gaming as an important aspect of learning. But the game has to change often, othervise it's just enjoyment to get perfect in a system that you already mostly understand. I usually reach this point after 2 to 8h in a game, because i want to understand the logic of the game. If you enjoy the story, you have to go trough the whole game usually. To reflect on the things you played and experienced is good to learn as well, but usually spoils the fun. There are a lot of games that teach logic thinking and often contain logical structures systems or train logic thinking. (minesweeper, sudoku, kerbal space programm, bridge builder games, lots of construction games (minecraft)...)

If I had to teach a class right now, I'd teach them on Raspberry Pi's and with some command line / terminal capability that includes SSH. I would probably bring in mathematics, taking care to separate the two pedagogically. I think the rest of the details would depend on the age of the kids in the class.

I did a video series with this in mind: gameswithgo.org

Teaching programming fundamentals via game projects. Things mostly from scratch, no engines.

I learned to code on GW Basic in the 1990s. My first programs were simple games like: try to guess which random number the computer picked, you'd input a guess, and the computer would respond with higher, lower, or correct (IF-THEN-ELSE). You'd get to keep guessing until you win (LOOP).

Those might be mundane coding examples for teenagers, but it taught me the basics of logic.

Functions is one thing I wish I had learned about a lot earlier than I did. Also how to commenting out lines of code, rather than deleting them. And how to use the debugger. These are no-brainier concepts now, but I was 10-years-old and self learning from books with some guidance from my dad and brother.

I would definitely say stay in the web development area of programming. It's really expanding at an incredible rate. HTML, JavaScript (All the frameworks), CSS, PHP, etc.

When I was quite young, I started 'programming' by copying code out of magazines. I would run home from school, open up a magazine, write code until dinner, run downstairs after and write code until I had to go to bed. If I repeated that often enough, I would eventually (if I was lucky) be able to compile my code, fix the 14,000 missing pieces of punctuation and then have a (very primitive) game to play and share with my friends.

I don't recall learning much, but I was at an age where I was more concerned with getting Nario to jump over a bucket than with actually learning. I likely did learn something, maybe even just a passion for sitting in front of a computer and making it do cool things. I'm sure I also learned that if you really want something, just put in the hours and jfgid. For me though, the operative part was games. I wanted to play games and at some level, I wanted to learn how other people built amazing games for my Tandy.

The point of all that (other than 7:30am nostalgia) is that, for me at least, I had to be motivated. I had to want the end result enough to push through the grueling bullshit of it all. To this day, I believe that the only way to learn to code is to fall in love with a project to such an extent that you can't stop yourself. Whether that's a good lesson for teenagers will have to wait...my kid is only three, so I have a decade of self doubt to interfere. :)

This is more "resource" than "curriculum" but have you seen Glitch.com? It's a combination hosting service/dev environment/playground that has lots of example projects that can be forked and modified, many of which are things that they might find interesting (messing around with Spotify playlists, etc.).

I would direct them to a game style coding session as gaming can be a stepping stone for some.

For me gaming led to wanting to mod the games, which led me learning how mods were created and learning to 'break' the game.

I am not affiliated with them however I did stumble across a DApp programming 'game'

https://cryptozombies.io/

If they like the above at that point I would direct them to other programming languages.

First one is Beaker Browser. I know this isn’t a popular technology - but that’s not important to a teenager. I have been teaching some teenagers how to build little JavaScript apps in Beaker (no npm or git yet) and it has completely caught on for most of them. These are kids familiar with Scratch but not text-based languages.

The nice thing about Beaker is:

* They can write dynamic apps without needing to write server code. (We are working on note-taking apps and home pages that save to the file system.)

* They can share their stuff and link to each other without needing to publish anywhere. This also makes it a kind of private network, so it feels like a safe place to make mistakes.

* Has a built-in text editor and file manager. Eventually they will learn more - but let’s start with just the HTML, CSS and JS - that’s already a lot to take on.

* Concept of versioning is built-in, so hopefully this will make it easier to teach source control.

Other good tools that I have enjoyed using to teach elementary-age kids: Twine is amazing. And Scratch (as well as Scratch Jr on touch devices) are good preparation.

To determine interest level, I like to have them watch over my shoulder while I build something small for five or ten minutes. (Don’t go too long.) I start by just showing how HTML tags work. This will excite many of them - they may even start to conjure up ideas right away - if it doesn’t excite them, then perhaps the interest isn’t there. (Or they might not understand something like HTML yet and need something simpler.)

If your kids get into coding and if they're doing Javascript (which I think is a good place to start for lots of reasons, but python might be better) I can personally recommend Free Code Camp for learning basics and intermediate topics in javascript web development. The interface is thoroughly gamified and it's very effective for people who are, but ONLY after they're enthusiastically hacking away at web dev. They'll be much more successful at learning the basics once they realize that they want and need the basics.

How to teach programming to your loved ones: https://www.youtube.com/watch?v=BleOgPhsdfc

when I was learning programming as a kid it was all about video games. Quake was a new game and it had an embedded language called QuakeC that controlled the game logic. That was hugely fun for making Quake mods. That set me up to do programming in Perl and C.

I would probably play to their interests with projects that are more concrete than abstract. If they are into video games, lots of games have mod languages and they can start by giving themselves more health, more power, different weapons etc. If they are into social media maybe javascriptlets to edit the DOM on a website to change peoples tweets and take a screenshot, twitter or insta bot that automatically responds to certain types of posts, etc.

Teenagers should be able to grasp non-toy languages and the concept of searching StackOverflow for how to do stuff :)

Glitch is amazing. It solves the development environment problem better than anything else I've ever seen - you can spin up a new environment entirely from your web browser without installing anything, and seconds later you have a dynamic website that other people can visit.

It's also a really great learning community, thanks to their "remix" feature (like forking on GitHub).

I'm super grateful that my first exposure to programming (high school in Eastern Europe, 10+ years ago) was in C++. I don't think we even did a lot of C++ specific stuff so it was probably closer to C but using cout instead of printf. Covered things from if-statements and loops to sorting and tree-searching algorithms. It doesn't sound flashy and it somehow goes against the trend to try to make learning exciting, but it lays a really nice foundation in your learning.

I would not focus just on programming but on projects. I would get a few Raspberry PIs and some electronic components and let them build and program devices. Build a simple weather lab for temp, humidity, etc. Build a silly traffic light that is triggered by a motion sensor. Have a lamp that can be controlled from a web page.

Programming isn't just about web apps and mobile games, so teach them how to build simple projects. Their eyes will open up to all the things computers control all around them. That way they get to decide what they like, hardware, software, both, neither.

I started with BASIC programs written on my TI-86 to automate math problems; they may find that a useful way to get started, and learning a little BASIC shouldn't be much more brain-damaging than learning Javascript.

I kind of have a problem with training people to "code" rather than teaching computer science; feels sorta like comparing a "wine and painting night" to a course in oil painting, in that you might shortcut to something acceptable-looking but with little understanding of how or why you did it. But they're kids and will probably quickly tune out during Dad Explains Linked Lists :)

Ideally I'd suggest starting off with some smaller guided exercises to get them familiar with the language, then work with them to spec out and implement a more interesting project. In my CS courses in college we implemented the minimax algorithm for tic-tac-toe and checkers, starting with a tic-tac-toe engine which just printed out its desired move and evolving along until we were displaying a checkerboard as it played along. Could be a good project for you since it's implementing something they're familiar with (board games), doesn't use anything weird like networking, and can introduce concepts such as recursion and complexity ("if we have it calculate the next move, how many possible moves are there? What if we calculate the next two moves? The next three?")