IRC, freenode, #hurd, 2011-05-22
<silver_hook> Since apparently Hurd's aim is a very stable and transparent
system ...why aren't there any companies backing it up?
<antrik> silver_hook: it's not in a state yet where it would be
commercially interesting
<antrik> silver_hook: and after some epic failures in the 90s, few
companies dare to invest in microkernel development...
<silver_hook> Isn't MacOS X running on top of Mach?
<antrik> yes, but it's not a true microkernel system
<antrik> for one, it's single-server, which is boring
<antrik> also it uses co-location, i.e. runs all the system code in the
kernel address space -- they are separated only formally
<antrik> even NT is more of a microkernel system I think
<silver_hook> Oh, OK, I'm not that knowledgeable about kernels to know
that.
<antrik> well, now you know :-)
<silver_hook> Yup, thanks :)
<antrik> most people don't know this, so don't worry
<silver_hook> I was just wondering that it might be potentially an ideal
server system, right?
<antrik> well, *potentially* it might be an ideal general-purpose system,
which includes server use... though personally I think the advantages of
the architecture are more visible in desktop use, as servers tend to be
rather streamlined, with little need for individualisation :-)
<antrik> however, it still remains to be proven that true (multi-server)
microkernel operating systems actually work for general-purpose
applications...
<silver_hook> antrik: I mean regarding hosting or virtual servers.
<antrik> so far, they are only successful in the much simpler embedded
space
<antrik> well, yes, the Hurd architecture in theory allows very much
flexibility regarding virtual environments... I once blogged about
that. not sure whether server applications really require that
flexibility though. I think most people are pretty happy with the various
virtualisation/container solutions available in Linux. again, the
flexibility is more relevant in the desktop space IMHO
<antrik> dosn't mean it wouldn't be useful for servers too... just not as
much of a selling point I fear :-)
IRC, freenode, #hurd, 2011-07-09
<antrik> gnu_srs1: regarding your question why people aren't interested in
workin on Hurd: Eric Raymond explains it pretty well in his famous
"Cathedral and Bazaar" paper
<antrik> people are more likely to work on something that *almost* works
for them, and where they only have to fill in a few missing bits
<antrik> the Hurd doesn't almost work for anyone
<antrik> actually, you should probably reread the whole paper. it's
essentially an analysis why the Hurd failed compared to Linux