An intelligent object-oriented database architecture to represent knowledge - 1990
With a great deal of naive optimism, i embarked on a mission to design a database that would be able to understand a very simple set of "knowledge-like" relationships. I wanted a system that would let me organize my notes, file my various collections, and represent biological data in a way that understood the relationships of environmental factors. Several months of my life were completely absorbed in the process. I consulted the work of many academics (e.g.. the book "Intelligent Databases: Object-Oriented, Deductive Hypermedia Technologies") who seemed to be just as far from working code as i was.
Years later I learned of similar projects - Xanadu, which has many similarities to my thing. Like Xanadu, thing would revolutionize how computers store information, never had a workable front-end, and years went by with nothing to show for it, ending in frustrating defeat. Also similar is Cyc, an attempt to do symbolic AI, encoding basic knowledge, on a massive scale, which is still persevering with hopes of being a useful product... someday.