Kodak's patents on CMOS image sensors (which were invented in the 1970's) ran out just about the same time as all the complementary technologies became commercially viable. Things like signal processing, data storage, and energy storage.
Think about how you would perform the DCT to do JPEG compression in the early 1990s. You'd need an expensive, power-hungry ASIC. Think about storing 10-100 MB of data. That would be a 2.5" spinning hard disk.
10 MSps ADCs (fast enough to shoot a frame per second) were not exactly exotic, but still very expensive. 16MB of DRAM was a luxury on Unix workstations. How much of that do you want to spend on a framebuffer?
Think about how you'd power all this hardware. Heavy NiCd batteries with memory, yuck!
The first DSLR that really took off for photojournalism (i.e., good enough quality to use in a printed newspaper) was the Nikon D1, in 1999. It's not that nobody had the right idea (Kodak had been selling digital cameras since the late 1980's), but it took that long for technology to catch up with the idea of a digital camera. It turns out (gasp!) that nobody wanted a digital camera that weighs 40 lbs, stores fewer photos than a roll of film, and has lower image quality.
It's fun to blame Kodak management for being short-sighted, but I really don't see a winning scenario. Short of a "Bet the whole company on CMOS in 1980 on the assumption that multiple enabling technologies from 3rd parties will improve by 10-100x in 10 years and that consumers will actually like the new product while we develop the right competencies and starve investment in growing lines of business." That doesn't sound like a way to get your year-end bonus.
