Re: Changing from L4 to something else...

[Ludovic Courtès]

Hi,

"Yoshinori K. Okuji" <address@hidden> writes:

> The recent debate in this list makes me very afraid of a possibility of 
> extreme cost which never finish in a reasonable amount of time or human 
> resource. No matter whatever design mistakes the Hurd has, the reason why it 
> is nearly unusable even after 15 years is that most people have merely 
> dreamed and never implemented their dreams.

I think Marcus, Neal, and others would be in a good position to say that
they did try hard to implement their dreams.

You seem to forget that one of the reasons the Hurd has been
next-to-dead for a number of years is that even its developers and
original designers seem convinced that it has a number of "unfixable"
bugs.  More precisely, "unfixable" here means: "not fixable on the
current Hurd-on-Mach implementation".

Even as a simple "lurker", I now know those flaws too well to feel
comfortable advocating the use of the Hurd on Mach.  Moreover, some of
these deficiencies (e.g. the passive translator vulnerability) are
clearly not as widely well understood than those of "regular" Unix
systems.  A well-trained Unix administrator may be able to cope with
security issues in GNU/Linux, but he may certainly be unable to cope
with those that GNU/Hurd _adds_.

Really, I don't think the rationale for switching to L4 was satisfying
people's "security paranoia": it was about fixing some of those
"unfixable" issues.

Thanks,
Ludovic.