Linux was a kernel in more than one sense of the word. It was an OS kernel written by a beginner, which became the kernel of a low-tech mass-movement away from high-tech free software possibilities. I want to replace Linux with something just as useful, but much more maintainable.

I currently have very rough requirements and design based on a "right tool for the right job" approach which should attract multidisciplinary involvement: Erlang is perhaps the "right tool" for networking, at least. Perhaps a logic language like Prolog or an object-relational database like Postgres is the best way to manage software version and build dependencies in both a local and a distributed way. Maybe Ada is the best language for device drivers. Perhaps one of these message-passing languages (Ada or Erlang) can be a Rosetta Stone for all the others. Maybe Common Lisp can provide each processor's main loop. C++ policy-based design is a powerful static code-generating design-level paradigm, but can D or other metaprogramming-enabled languages do it as well as C++?

OS theorists and hardware designers will be needed to keep the programmers under control.

I feel this project needs to happen. I've done a lot of research, and every university seems to have their own kernel, and they're all pretty boring. The only new and exciting stuff seems to have happened in 1995 or earlier. Microsoft is coming up with some new ideas, and some of them are very good, but they won't be free.

Ultsys is just a working title so that it'll be the same as ultsys.blogspot.com where I'm documenting the background of my thinking. Blogspot is not a good place for technical discussion so this might be a better forum, or not.