Meanwhile, the former creator of Delphi had left for MS and created C# that at the time was already beating Delphi at its own game.
By the mid 2000s Delphi was already dead in the water: poor unicode support, no 64-bit compiler, and a language that hadn't innovated much since the 90s.
It's a shame, really. Delphi in the 90s was on its way to rule the world. Only then suddenly it wasn't.
It's worth noting that Kylix's failure to take off had a lot more to do with Borland's variously stupid, obnoxious and immoral business decisions than it had to do with the upheaval in the Linux desktop world. It was a strong enough concept for a product, and the base technology was implemented well enough, that all they had to do was not screw it up... and guess what they did?
All these years later I remember saying "fuck it" when Borland decided to wrap up some important bug fixes into the (full price) upgrade for Kylix, rather than just giving them to existing users. The smug uselessness of the one or two Borland employees who liaised with the community online has kind of stuck in my memory as well. The technology deserved to survive but the company sure didn't.
RIP. You were great once.
I also remember a coworker had been approached by MS to work on the DotNet marketing campaign and he declined, as it was seen as kind of a failure with a very uncertain future. Hindsight is 20/20.
edit: personally I have found memories of Delphi3.
It has all the downsides of service based DB without all the advantages of file based DB.
And the documentation is way below average, backwards compatibility is a joke.
Not to mention the 3rd party software had to design their own(horribly bad) way to handle geometry in firebird... and they keep transactions open forever - and bloody firebird prior to 2.5 had no way to drop the dead connections.