A lightweight HTML parser that actually works

The input is parsed into a tree which is then processed recursively to produce a single list of boxes.

And then what?!

Here's a whole browser:

Start with a url given on the command line.
Parse the url.
Use http to deliver the url into a cache file. The name of the file is returned to the main loop.
The file is parsed into a tree.
The tree is displayed.
References are listed.
The user enters a number or ^D to quit.
The reference is combined with the current url to make a new one.
Continue with 1.

Looks simple, eh? Let's do it!

Done: line mode browser in under 1000 lines.

#!/bin/sh echo "Content-type: text/html" echo echo ""