The year was 1996. I was 14 years old. I had to take a RAINMAN seminar before I could intern at the "Digital City" division of media curating for "Digital City Denver: Arts Soup" sub-category. Want proof? I was on the cover of Denver Jobs, a printed monthly. The caption might as well have been, "Look, a vaguely Asian kid with blonde highlights knows how to use a computer!" The photographer stood on top of my rickety wooden desk. Fun ensued, maybe you could find it on microfilm in the basement of the Denver Public Library.
RAINMAN stood for "Remote Access Information Manager" but what it really meant was that there was a slow-ass server somewhere that took actual pseudo SMTP messages from a bizarre coding/gui interface (closest thing I can think of is if Dreamweaver were made by the 1992 MS Word team) that "submitted" your "code" via FUCKING EMAIL to be compiled on said shitty-slow pentium i486 in Texas or some shit.
Ten minutes later you'd get a list of compile errors in your AOL Account Inbox because your "screen name" was passed through the shitty in-AOL IDE to the server via FUCKING EMAIL! No, really, the only way to code for AOL was to login, click on the "page" you wanted to edit, have some manager add you to an access list and fuck up like 5 times because THEY'RE EMAILING A SERVER TO GRANT PRIVILEGES. So, you fix some formatting, apply a regex to clean the form field, hit save and then just stare into space for about ten minutes while you wait for the server to compile the change, and maybe email you with an error. It's a trade secret because patenting that shit would just make the assigned attorney die of laughter.
This is what wikipedia says about AOL Rainman:
Rainman code is kept strictly confidential by AOL; access is granted at their discretion.
Yeah, it's a secret because it's shit! If you wrote a system that compiled server side code by email, you should be in prison, not writing enterprise software. Instead, Marc is heralded as one of the foundational builders of the internet. This just goes to show that timing is more important than implementation.