Many, many years ago I owned and ran asciifun.com. Mostly it was created as a way to paste figlet output into code comments and other random stupidity happening on irc networks. Back in 1999, this blog was running self written code in php. The blog was loosely based on php nuke, but I decided it had too many security holes and started writing a system which used directories and files as the content repository. As an experiment to learn how to code php to interact with the underlying FreeBSD system, I wrote the asciifun codebase in a weekend. Unfortunately, I haven't been able to find the code in any of my archives. The astute reader will notice the code versions at the bottom of the page:
: index.php3,v 1.7 2000/02/19 22:49:53 deimos Exp $
That's version 1.7 of index.php3 committed to RCS on 19 February 2000 at 22:49:53.
Overtime, my personal figlet to web interface gained popularity. I had no idea. I didn't bother looking at log files or tracking anything about the site. I remember logging in to the server hosting the site at some point in 2002 and noticing a load average of like 88. This was running in a simple single cpu desktop with like 16MB of ram. Yes, this was cutting edge in 2002. Might have even been a 486 or Cyrix 686 I had laying around. A load of 88 was insane.
The core challenge was for every form entry on the site, I spawned a figlet process in the system and gave it a maximum runtime of like 30 seconds to generate the text output. Yes, the coding was that good. I ended up re-writing the code basically keeping a figlet process running at all time, sort of turning it into a daemon. Eventually, I completely rewrote figlet into php to provide the same output as the command line program. Mostly, figlet is open source so I just translated it into php. This ran much better under heavy usage.
Eventually, I was approached by a few advertising networks and asked if I wanted to run their ads on the site. I turned them down because this was a hobby expriment, not a business. (Besides, wasn't 2002 the first dot com bust? Who had money for stupid sites with no business plan?)
Later in 2002, I was approached by someone who wanted to buy the domain. They offered real money and just wanted the domain, not the code and such. They claimed to have some idea for business about coding education.
Time passed and I eventually took the offer, as I could use the money for my growing race car business. Of course, the new owners filled the site with ads. Clearly they knew something I didn't.
I think what surprised me were emails like this:
thanking you for a wicked site, truly ascii heaven. VERY fast responses and slick minimalist interface. 2 thumbs up....
Oh No! Something's not working at asciifun.com - I can't put any banner comments in my code now... Help! Seriously though - one of the truly great remaining 'old skool' site on the net. I hope you can fix it soon (I really do use it for creating banner comments in my code.)
People really missed the site and wanted to know if they could have the code so they can run their own. I just gave people the source code and let them go about it. This was before GitHub, SourceForge, and the like. Plus, the code was tiny and dead simple (all in one php file).
I eventually got the domain back when the banner ad experiment was over and redirected to my other site, rootme.org. That's a whole other story in itself.
If you want a fun site in the same vein, look at qaz.wtf for now. The future is unicodefun!