That was the Saturn you’re thinking about. The Saturn had two sprite based processors and did 3D by skewing those tiles. This also presented other problems like with transparency (morphing squares into triangles causes problems with alpha blending). It was how Sega arcade boards also worked at that time and so Sega engineers were well versed in writing 3D engines like that but the rest of the development community had settled on the now standard approach of triangles. Couple that with the lack of an SDK and a dual processor system in era before developers were used to writing for such hardware and you had a very problematic console.

The Dreamcast, however, ran a PowerVR2 chip which was much more familiar for anyone with prior dev experience.

Perversely the PlayStation 2 was more complex due to custom hardware like the “emotion engine”. But Sony already had enough momentum from developers and consumers for any such difficulties to become game changing.

As many have pointed out it was very easy to develop for. The developers of Dead or Alive 2(Tecmo), one of best looking 3D fighters on the Dreamcast state stated that developing on the Dreamcast was like writing a sentence with a pen whereas on the PlayStation 2 it was like writing a sentence with a brush.

People are still developing games for it using open source libraries. There was recently a 3D racing game released.

https://www.kickstarter.com/projects/pixelheart/arcade-racin...

Was it? That’s not something I’d heard before (unlike, say, the PS2/3, Saturn, or N64, where complaints like that are common). What made the DC hard to wrangle?

Here is an overview,

https://www.reddit.com/r/retrogaming/comments/dn275r/sega_sa...

Remember that Windows CE was just an alternative to the actual SDK, and it was something like DX 5.

This is about the Sega Saturn, not the Dreamcast.

Yeah I mixed it up, sorry about that.