Hacker News Reader

Comment by nerdponx | Hacker News Reader

Maybe this is a silly question, but if people are building these from scratch as a hobby project, why don't we have GNU Hurd working yet? Is the problem nowadays that the goal has shifted from "make a kernel" to "make a kernel that's bug-for-bug compatibile with Linux including driver support"?

View on HN

I'm not an expert, but my understanding is that HURD made some uniquely poor choices that make it difficult to work with in ways that other microcernels don't necessarily suffer from.

Be helpful if you could elaborate on those choices

HURD (mach) -> Hurd-NG -> back to HURD (mach):

http://walfield.org/papers/200707-walfield-critique-of-the-G...

One step forward two steps back...and no actual progress, that kills a project who is in competition with something like linux (who is obviously not a perfect design but just pushes forward)

Not a hobby project. Used for teaching at a university. See http://www.helenos.org/wiki/FAQ

I think it came about at a time when Stallman was trying to make a power grab for Linux and a lot of people saw it as empire building rather than as a serious project they could get behind. Meanwhile Linux was hitting The Network Effect at a time when people were itching for a challenger to Windows, and The Hurd was never going to be that.

Because it isn't cool to work on Hurd, most likely.

It would be cool if they had a direction...aka let's build on mach...oh that's old let's make Hurd NG....wait even less people work on that one? ...ok let's get back to mach...and what do we do now?

Having a kernel without goals is like a sandbox without having fun, no one cares.

I missed that that, yeah.

Why do you think that GNU Hurd isn't working? The Debian Hurd port is completely self-hosting and makes regular releases.

For one they still haven't finished their 64 bit port.

Reply on news.ycombinator.com