My Aha, Unix! moment was the Unix man pages, especially that they had a section for BUGS.  The very reality of it attracted me.  As Gloria Steinem said, "Something doesn't have to be perfect to be wonderful!"  I notice that on Linux the older man pages still have BUG sections, but the newer ones don't.  Telling.   Even more telling is that 'man python' gives you a lot of information, but at the end where the Bugs section used to be is a section labled "LICENSING"...

I did have the opportunity in the early years to demonstrate Unix to several dozen people, mostly users of the (IBM) mainframe computers and the GE/Honeywell Time Sharing System.  The sequence that initiated gasps, confusion, and ultimately joy was:

%  echo hello joe > hijoe

% cat hijoe

hello joe

At the time, permanent file storage was a relatively new concept for mainframes, and the implementations were very influenced by space constraints and punched card images.  The IBM was worst, because for them a disc file was made to look like a tape drive -- "records" that had multiple card images on them."  In order to create a file, you had to submit a job (punched cards again) using a Job Control Language whose authors are hopefully all burning in hell at this very moment.  And the job failed if the file was already there, ...   The time sharing system was not much better -- still had the notion of card images in mind, but also an initial size, a maximum size, and a lot of settings for who could do what with the file.  In the time sharing system, a special subsystem took control and asked you roughly a dozen questions, one at a time.   It was quite common to botch one or more of the answers, in which case you got to answer all the questions again.  No wonder when the file was finally created, the system replied "Successful!".

Typing the above created shock and awe followed by questions like "what's the blocking factor" and "what device is it allocated on".  Followed, mostly, by a dazed joy as they finally got it...

Steve