If you are reading this on delphi-dolphin.com, you might reasonably be wondering what a blog about AI, social housing, and digital transformation has to do with Delphi, or a dolphin. The answer is a 29-year-old act of naming desperation and a student project which got out of hand.
I was 24, four years into a degree, with the finish line in sight. The university called me into an urgent meeting. They presented a letter assigning all intellectual property in our final-year project, Delphi-Dolphin, to them. Sign, or risk losing your degree. I signed. So did the rest of the team, summoned individually rather than together. The project was a tutorial application for Borland Delphi. We had built it ourselves, taught ourselves the technology, and given it away free on the internet. The last point is what got us into trouble.
A beta test which worked too well
About the name first, because it is the question every reader of this site will likely ask. Delphi-Dolphin is an backronym. Delphi ONline Programming Help INteractive. I know. Even in 1997, finding a short, sensible, available domain name was close to impossible, so I worked backward from what was available, and DOLPHIN was the least bad fit. The contrivance required to make the letters work was significant. The cringe has never fully faded.
The application behind the name was something we genuinely cared about. Our assessment brief asked us to beta test the software with real people. The web was young in 1997 and largely treated with suspicion. We took a decision which felt bold: we put the application on a website and let anyone download it.
Around thirty thousand people did. The feedback loop was real and immediate in a way formal user testing rarely is. We iterated. We released a second build. People kept downloading.
Then the magazines came calling. PC Plus included Delphi-Dolphin on their August 1997 SuperCD. We were on the bottom strip, listed as "On-line Delphi tutorial". No name, no fanfare. A single line of text alongside "Over 500 free Delphi components". Bottom strip placement is more meaningful than it sounds. Cover discs in 1997 were how millions of people in the UK acquired software before broadband existed. Inclusion meant the application had been judged good enough to distribute to a mass audience. For a student project, it was unusual.
The institutional response
The university was not pleased. None of us had made a penny from the distribution. We had credited the institution in the software itself. The technology we used was not on our course. We had built it using skills were not taught on the course, for an assignment which explicitly asked us to reach real users. The result had exceeded the brief by some margin, which is precisely what an educational institution should want from its students.
What the university did instead was call us in one by one. Not as a team. Individually, which in retrospect was the point. A pre-drafted letter. A signature line. A threat large enough to overwhelm a student with everything to lose.
The grounds for the claim were weak. The execution was effective. We signed.
How UK universities handled student IP in the late 1990s was neither clear nor consistently applied, and I am wary of overstating the legal position. What I will say is this: the experience of being summoned alone, presented with a pre-drafted letter, and threatened with the most significant consequence available to the institution, left a mark. Not a wound. A lesson.
What an institution with nothing to lose taught me
Twenty-nine years later, I'm a CTO building AI-powered products in social housing, working directly with the people those products are designed to serve. The technical tools are unrecognisable. The instinct, build something which works, put it in front of real people, listen to what they tell you, is identical.
The university's response did not discourage me from building. It taught me something about institutions and power which turned out to be more useful than most of what was on the curriculum.
Three things, specifically. The first: institutions move to claim ownership of value they did not create the moment recognition arrives. They rarely move when work is invisible.
The second: the threats institutions reach for are usually the ones they hold most freely. A signature on a piece of paper is cheap to demand. A withheld degree costs the institution nothing and costs the student everything. The asymmetry is the point. If you ever find yourself negotiating against an institution, look at what they are willing to spend, not what they are willing to threaten.
The third, and the one I have come back to most often: the people who build things which work are usually the people who taught themselves to build them. The curriculum is not where the leverage is. The leverage sits in the gap between what you were taught and what you went and learned anyway. Institutions sense this gap and respond to it badly, because it exposes the limits of what they offer.
I see the same pattern in the work I am doing now. The tools for building AI-powered products are not on any curriculum either. The people getting useful results from them are the ones treating the tools as something to learn by using, breaking, and rebuilding, rather than waiting for someone to certify them. Twenty-nine years on, the disposition is identical. The technology has changed. The institutional response to capability built outside the syllabus has not.
I still own delphi-dolphin.com. You are reading this on it. The university did not get to decide what the work meant then, and they do not get to decide what it means now.