Exposing youth to programming by finding the password

Introduction

Back in 2013, Marc Scott wrote a great blog post: Kids can’t use computers…and this is why it should worry you. If you haven’t read it, I highly recommend it! In my opinion, Scott’s take was true in 2013, and even more true now, over a decade later.

A favorite excerpt:

There are variants of the phrase, all espousing today’s children’s technical ability. My favourite is from parents: ‘Oh, Johnny will be a natural for A-Level Computing. He’s always on his computer at home.’ The parents seem to have some vague concept that spending hours each evening on Facebook and YouTube will impart, by some sort of cybernetic osmosis, a knowledge of PHP, HTML, JavaScript and Haskell.

Normally when someone spouts this rubbish I just nod and smile. This time I simply couldn’t let it pass. ‘Not really, most kids can’t use computers.’ (and neither can you—I didn’t add.)

She looked surprised by my rejection of what is generally considered a truism. After all, aren’t all teenagers digital natives? They have laptops and tablets and games consoles and smart phones, surely they must be the most technologically knowledgeable demographic on the planet.

An arguably-too-brief summary of the post is: despite the widespread use of technology, many people, including children and adults, lack true technical literacy. Scott shares personal anecdotes to illustrate how even basic tasks on computers often baffle users.

Why does this matter? Scott concludes:

I want the people who will help shape our society in the future to understand the technology that will help shape our society in the future. If this is going to happen, then we need to reverse the trend that is seeing digital illiteracy exponentially increase. We need to act together, as parents, as teachers, as policy makers. Let’s build a generation of hackers. Who’s with me?

I’m with him.

Software continues to eat the world, and no matter what profession you ultimately work in, the probability that you are affected by and/or dependent on computer systems is high. Note that I’m not saying that everyone should be programmers or study computer science. I’m saying that computer literacy is highly likely to benefit a wide variety of career choices.

In fact, I was finally motivated to make a real implementation of a little toy challenge he proposes in the blog post. In his words (emphasis mine):

Stop fixing things for your kids. You spend hours of your time potty-training them when they’re in their infancy, because being able to use the toilet is pretty much an essential skill in modern society. You need to do the same with technology. Buy them a computer by all means, but if it goes wrong, get them to fix it. Buy them a smartphone, give them £10 of app store credit a year and let them learn why in-app-purchases are a bad idea. When we teach kids to ride a bike, at some point we have to take the training wheels off. Here’s an idea. When they hit eleven, give them a plaintext file with ten-thousand WPA2 keys and tell them that the real one is in there somewhere. See how quickly they discover Python or Bash then.

Find The Password

You can see the project on GitHub here:

https://github.com/lukehsiao/find-the-password

This project is a simple-to-run web app which generates a list of 60k passwords, and provides a simple URL scheme for checking passwords to know if they are right or wrong.

Typically, the challenge is presented with some rewards for the earliest solvers (and it shows a leaderboard for that), but the real goal is to tap into some natural curiosity and educate.

Note that there are still several ways you can take the challenge for different knowledge levels of youth.

  1. Simple, brute-force serial approach. I don’t have a level less than this, because I think the struggle and learning to get to this point is the crux of the challenge. For those with less experience, pointers on what to search, or even setting up a working Python environment with an example file that makes 1 web request might help. Note that there are other ways people might solve this (e.g., utilizing the web request functions of Excel), so really any solution works, but this is an opportunity to teach about programming.
  2. More optimal, concurrent approach. E.g., if you have someone somewhat familiar with programming, have them implement a checker that brute forces all the passwords as fast as they can!
  3. Design the challenge. If the concurrent brute force is too easy, it’s time to talk system design. For these young people, pose the question “How would you implement this challenge?” This presents an opportunity to talk about tradeoffs in the design space. For example, the tradeoffs of the in-memory approach discussed below.

Design

This challenge is essentially inviting a bunch of people to DDoS your server. In addition, this is NOT intended to be a long-running application.

With these two things in mind, the design of this application takes the following approach:

  • Simplify user creation. No authentication, etc. Youth can add themselves to reduce admin overhead, and we don’t worry about extra persistence or auth or anything because the server will be killed after a short time anyway.
  • All in-memory. We want high throughput. A previous iteration of this server was implemented with a sqlite database, but would cap at around 15k requests per second. By solely using in-memory shared state, this server does closer to 60k rps. Eliminating backing services also simplifies things.
  • But not too much memory. Since we want to use memory only, we don’t want to use too much. For example, when generating a 60k list of passwords (~2MB), we could store that for each user after we first generate it. Instead, to save memory, we just regenerate it from a random seed that we save each time the password file is requests. This optimization makes sense because getting the password file should be rare compared to actually checking the passwords. We also do save the correct password directly so that that check is wicked fast.

So, what this looks like is essentially just 3 endpoints:

  1. A homepage with instructions, a leaderboard, and a way for a youth to join the challenge.
  2. A user-specific page that allows them to download a password file.
  3. A password-checking API.

Based on benchmarks (see the repository), this server does about 60k requests/sec on my home desktop (64 GB of DDR5 RAM and a Ryzen 7 7800X3D) when both server and load test are running locally.

Stories So Far

I’ve given this to a handful of youth, ranging from ages 10 to 18.

Here are some stories worth highlighting.

Never underestimate the grit of a teenager

One of the first times I gave this challenge, I offered a monetary reward for the first person to find the password. At the time, I was generating a list of 10k password as the blog post suggested.

One particularly determined young man had a high tolerance for manual repetition. He had cleverly figured out how to generate the list of URLs from the passwords by pasting the newline-separated passwords.txt into a spreadsheet, and then leveraging spreadsheet functions to construct the URLs for each.

Then, he discovered there are websites (example) in which you can paste 1 URL per line, and it will open 1 tab per URL (the only limit being the actual resource limits of the computer you’re running it on). In his case, he found that was about 500. This young man then proceeded to open batches of 500 URLs at a time, gluing his eyes to the top left of the screen to look at the false or true printed there, positioned his mouse accordingly, and just closed tabs fast, waiting to see if a true flashed. In his manual, brute force approach, he solved the puzzle in just a few hours!

Now, the list is 60k passwords.

This is an opportunity to teach a growth mindset

The flip side to the thrill of seeing youth dive in with enthusiasm, is seeing a youth react very quickly with defeat. This is not unusual if giving the challenge to a general group of youth. Some will have a reaction of trying several passwords by hand, realizing this is an untenable approach, and then quickly deciding “they don’t know” and quitting. It feels even more heart-wrenching when you hear defeated statements erroneously tied to identity, like “I’m not smart enough”, or “I’m just bad at computers”.

I hope if you give the challenge, you see these as opportunities. Here is a moment in time that you might uniquely suggest an opposing, uplifting opinion to someone about what defines them, and what they can do. My goal in these scenarios is to do my best to make this experience a memorable counter-example for that individual—yes, you can do it!

Conclusion

Selfishly, this was an opportunity for me to learn about full-stack web dev with Rust using Leptos. However, even as a toy or a thought experiment, it’s proven to be an interesting and insightful little challenge to give to the rising generation. It feels more real in terms of what they build—similar to Advent of Code, but just a single, much simpler problem. That is, they are presented a real problem, and need to write real software (or other solution) that is making real network requests to a real server. They aren’t just playing in some carefully constructed sandbox.

If you know of more things like this, or if you try this and have a story to share, I’d love to hear them!

If you’re a parent, I also highly recommend Programming For Kids by Borislav Nikolov, who continues to add excellent bite-sized lessons about computers as he teaches his daughter. At time of writing, there are over 404 (!!!) lessons. This is a gold mine.

Posts from blogs I follow

Cold reading an ADHD affliction

I'm sure there are truly pathological cases of ADHD out there, and maybe taking amphetamines really is a magic pill for some folks. But there clearly is also an entire cottage industry cropping up around convincing perfectly normal people that they suffer …

via David Heinemeier Hansson November 20, 2024

You can use almost anything as a key file for your encrypted storage device

Imagine that you have an unencrypted drive containing your private data and one day it starts throwing a bunch of errors. You have backups of the data so you’ve got that part covered, but would you feel comfortable sending the drive in to be warrantied? Yo…

via ./techtipsy November 20, 2024

Testing with Go and PostgreSQL: ephemeral DBs

Let’s say you created a Go program that stores data in PostgreSQL — you installed PostgreSQL, wrote the Go code, and everything works; great! But after writing a test for your code, you wonder: how do you best provide PostgreSQL to your automated tests? Do…

via Michael Stapelbergs Website November 19, 2024

Generated by openring-rs from my blogroll.