Re: A comment about changing kernels

[Jonathan S. Shapiro]

On Thu, 2005-10-27 at 20:36 +0200, ness wrote:
> > If the goal of Hurd includes providing a more secure and robust
> > foundation, then I think that either L4.sec or Coyotos will be the right
> > kernel choice. Obviously I have a preference, but the Hurd decision
> > needs to be driven by the Hurd requirements and not by my preferences.
> > 
> That is right, of course. But maybye it would be nice fo you to explain 
> why you think coyotos is better than l4.sec.

I don't know if it is. Heck, L4 was a pretty good design, and L4.sec is
better. I would say they have it somewhere between 90% and 95% right.
Coyotos, of course, is perfect. But only on Mondays, Tuesdays, and
Wednesdays. On other days of the week we are actually working on it, and
it is somewhat less than perfect. :-)

> Marcus already mentioned that l4.sec is no alternative, as the time of 
> release is too far away.

This could very well turn out to be an issue for Coyotos as well.

> I can't 
> even say whether it is persistent. And your documentation page isn't 
> that helpful, as the kernel spec is mostly incomplete.

Coyotos will be persistent. We started on a kernel spec, and decided
that we needed to do a quick prototype implementation of certain things
before trying to capture the spec. Unfortunately, I'm the prototype
author, and as you may imagine progress is slower than I might prefer.
Eric Northup has been working out the PATT stuff, and has made
significant progress there. Two other students are working on things
that will eventually lead to verification. Our job, after all, is
research.

>  (btw, I tried to 
> subscribe to the coyotos lists but I got no mails - has there been 
> traffic this afternoon?)

It's definitely a low volume list at this point.


shap