This site is now browsable with netcat/plain TCP

2020-01-20T00:00:00.000Z

This is getting out of hand. First, I wrote an unrelated post, and Hacker News got a little upset with me about my blog requiring JavaScript. I ultimately agreed with them. To redeem myself, I updated my site to not just serve static HTML, but to be entirely browsable with nothing but cURL. Full background and details in this post.

However, I kept going down the rabbit hole, wondering how simple serving blog content can realistically be. Here's how far I've gotten. If you have netcat installed, you can browse my site by running this command:

nc apitman.com 2052 <<< /txt/feed

There are further instructions at the top of that "page" describing how to navigate.

It uses an extremely simple protcol. You can use any TCP client. Just open a TCP connection to apitman.com on port 2052, write a path in plaintext, and it will return the contents if found. I call it the newb protocol.

Note that newb connections are not encrypted. If you want private browsing, you'll need to use HTTPS, either with cURL or a browser.

If you don't have netcat, it's simple to write a newb client in your favorite language. Here's one in node:

#!/usr/bin/env node

const net = require('net');

const client = new net.Socket();
client.connect(2052, 'apitman.com', function() {
	client.write('/txt/feed');
});

client.setEncoding('utf8');

client.on('data', (data) => {
  console.log(data);
});

And Python:

#!/usr/bin/env python3

import socket

with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
    s.connect(('apitman.com', 2052))
    s.sendall(b'/txt/feed')

    chunks = []
    while True:
        chunk = s.recv(1024)
        if not chunk: break
        chunks.append(chunk.decode())
    print(''.join(chunks))

Those can both easily be adapted to take the path from the command line. And the host address too if anyone else decides to implement the newb protocol ;) Speaking of which, the server code is here.