NetBeans started as Xelfi student project at faculty of mathematics and physics of Charles University, Prague. Originally we wanted to create Delphi for XWindow system, but later we decided to switch to Java and create the first IDE for Java written in Java.
In 1995 me and eighty other students were supposed to select a final software project to finish our studies. There was a huge gathering where various professors presented their projects hoping to attract attention of students. There was a project to control real robot - the previous group managed to move it and even speed it up, until it hit the wall. The goal for next project was to slow it down or even stop it to avoid the crash. Nice, but a bit too hardware oriented and I was more a software guy. Then there was a project to build a database system for a hospital - nice and valuable goal to help poor patients, but I was not huge fan of SQL & co. And so on, so on. When the presentation was over, I still didn't know what project to join.
However I wasn't alone. Several other guys were disappointed with the presented choices. We started to chat and agreed what we would like to do: we'd like to build an IDE - something like Delphi, but to run on Linux - e.g. for XWindow system. Luckily there was a visiting professor from U.S. - AdamDingle - we contacted him and (as he didn't know what it is like to lead seven students during software project) he agreed to do it. We officially announced the project and started weekly meetings. A usual meeting lasted an hour and we chatted about the great design, plans and then went into a pub. It was hard work, but we managed to do it for almost a year. Without producing a single line of code!
During that time I managed to participate in another project where we were using the new, cool language called Java. It didn't feel that bad, so at one of Xelfi before-the-pub meetings I suggested to use Java instead of C. The response was overwhelming: no! Thus we continued in our regular meetings for next three months, still without writing a single line of code. Then I proposed to use Java once again: this time the suggestion got accepted. So we all started to learn and code in Java. The progress was good: Java really delivered on the promise to run everywhere (I was OS/2 user, there were few Linux users and of course bunch students using Windows) and provided single GUI toolkit on all these platforms (something we struggled to unify when working on the C version of Xelfi). In a few months the IDE was working and ready to help us pass the final exam.
We passed the exam with excellent marks. Some people say that it was because at that time nobody from the reviewers could execute Java program. So they couldn't try Xelfi themselves, but had to trust our demo and we obviously knew what features to avoid from it. The only material they could study alone was documentation, and we had an enormous amount of that thanks to another Java's invention: Javadoc! We impressed everyone by generating hundreds of pages describing (well rather just listing) all the methods we had in the system. No surprise everybody was impressed.
However it was not a fake - we actually liked what we did, so some of us tried to push Xelfi further. We started to offer Xelfi as a shareware asking for $20 per copy. We have not made much money this way, but (and that is more important) we attracted attention: we were contacted by Roman Staněk who was seeking for investment opportunities and in autumn 1997 the company which later turned into NetBeans was established.