Hacker News new | past | comments | ask | show | jobs | submit login

"Don`t get me wrong, I am not unhappy with LINUX. It will get all the people who want to turn MINIX in BSD UNIX off my back. But in all honesty, I would suggest that people who want a MODERN "free" OS look around for a microkernel-based, portable OS, like maybe GNU or something like that."

Hehehe.




It's easy to be smug now knowing how things turned out, but lots of really brilliant people agreed with Professor Tanenbaum at the time.

The difficulties that microkernel projects ended up encountering were not easy to forecast and ended up taking virtually everyone by suprise.

It was in the spirit of progress towards better ways of architecture software that Tanenbaum and Stallman (as well as many others) chose to try a new architecture rather than just build yet another monolithic OS kernel. Being on the pointy end of technology means you often end up being the one to discover what doesn't work.


What's even more amusing about that remark is that - as of MINIX3, at least - MINIX has in fact adopted the NetBSD userland.


He is not talking about the userland, he's talking about the kernel structure, mirco vs. monolithic.


Ya...part of the design goal is that a microkernel as envisioned by AST should be able to able to have interchangeable userlands, including multiple different userlands running at the same time. So in that sense, AST was spot on.


Oh, and:

- device drivers as processes (so you can actually debug them)

- increased security by isolating various parts of the core OS from each other

- easy scaling from single machine to cluster by message passing

- treating devices as networked resources

- file systems in userspace (which we have now with FUSE)

and so on.

The benefits of microkernels go a lot further than just being able to use multiple userlands.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: