Don Box points out the Death of Hypercard.  I loved HyperCard.  I lost my 'programming virginity' to it when I used it to build a custom web browser and hyperlink editor as part of my Psychology Honours Thesis.  It will always have a special place in my heart as my 'first time'. 

Sure, I'd done some great work in primary school with the Logo turtle, and spent a long time learning Basic programming on the Commodore 64 (annoying my brothers who wanted to watch the TV) but I'd never 'released' anything for other people to use.  Here are some things I learnt (more about the thesis at the end).

What I learnt from my 'first time'
Here are some learning experiences that stuck in my mind:

• Estimating development time is difficult.  I was incredibly optimistic and ended up being late and having to cancel the first day of the experiment.
• Coding through the night is a waste of time.  In desperation I had stayed up programming through the night to get it finished.  After a sleep the next day I realised that everything I did post midnight was useless and had to delete it.
• Pair programming is a great way to learn.  After pulling the all-nighter my brain was fried but the program wasn't finished.  Luckily a friendly network admin in the psych department paired with me and helped me solve a complex problem.
• Milestones and frequent, iterative releases are a Good Thing.  I tried to do it all in one go and make it 'perfect' rather than focussing on getting it working end-to-end and gradually adding features.
• Usability testing is important.  Although I showed the program to friends, I didn't actually get anyone to use the software before I started the experiment.  Having to watch 200 people struggle with my poor interface was an important lesson in the importance of doing some discount usability before spending time writing code.  Ironically enough, one of the main authors I referenced in the thesis was Jacob Nielsen (in his academic phase, before the crazy days of the dot-com).
• Books are a great way to learn.  I love books as a way of learning something (I even taught myself to juggle from a book).  I used Danny Goodman's The Complete Hypercard Handbook (here's his view on the death of Hypercard) to teach myself how to implement my solution.  I got everything I needed from there (even implementing my own scrollbar with user-drawn arrow heads and mouse tracking).  Later I got his JavaScript Bible, which was the most expensive book I'd ever bough at AUS$70 but translated into a AUS$2000 contract to build a multiple choice system for the uni.  I couldn't believe the financial leverage!  Recently I've dampened my love of books slightly in favour of learning by 'just writing the code'.

What was my thesis about?
My thesis was looking at whether hyperlinks could be useful in helping people learn (the thesis is now available to the public at my uni library).  I built a set of pages describing key topics in first year psychology that the first year students could use to revise for their exam.  In my study there were three groups of students - those who browsed a set of hypertext pages (we'd say web pages today) which had links authored by the Lecturer, and two other groups that had to link the topics themselves.  One group had to choose from a list of associations to describe the connections between two topics and the other group had to type a description of a relationship themselves.  My hypothesis was that the groups that created the hyperlinks would do better at remembering the material (and get a better exam grade) than those who just passively read the material.  In the end there were no significant relationships (ever the way in honours topics), but it was a great excuse for me to learn how to program.