Re: L4-HURD , POSIX, UNIX

[Marcus Brinkmann]

At Tue, 1 Nov 2005 17:10:58 +0100,
Filip Brcic <address@hidden> wrote:
> 
> Дана Tuesday 01 November 2005 16:47, Jonathan S. Shapiro је написао(ла):
> > On Tue, 2005-11-01 at 13:35 +0100, Bas Wijnen wrote:
> > > On Mon, Oct 31, 2005 at 08:02:29PM -0600, William Grim wrote:
> 
> > > I like persistence as well, mostly because I feel it results in a more
> > > robust system.  However, if people already see concrete problems with it,
> > > I'd like to hear them now, not after we decided that we want persistence.
> 
> Persistence could be an option. There are issues with persistence (such as 
> reloading the corrupt code on every boot),

There are issues with persistence, but they are luckily not as bad as
you seem to believe right now.

Note that even in a persistent system you can restart services, and
whole subsystems.  There are only a few core components which must be
rock solid.  However, the same can be said about other systems.  For
example, corrupt the root partition of your Unix system and you are in
trouble.

> and therefore it should not be 
> enforced. Personaly, I would like to see persistence in GNU/Hurd.

I don't think it is feasable to have it optional at the operating
system level.  If we go for it, then we will very likely require it.
There are some things in the design and implementation level that will
want to depend on a persistent system to really make it useful.

> > Thank you. This is absolutely correct and very much worth emphasizing.
> > The two systems that I feel I know very well are UNIX and the
> > KeyKOS/EROS line.
> >
> > The thing that I bring fairly uniquely is an in-depth understanding of
> > security and robustness issues and how to make strong security usable.
> > Whether this is important to Hurd must still be decided.
> 
> Strong security and robustness can always find some use. My opinion is that 
> it 
> must not be enforced for the regular (future) Desktop user of Hurd, but it is 
> very important for server computers and embedded devices. If you have 
> experience and knowledge, you should be welcomed to use it for building a 
> secure and reliable Hurd.

There are reasons to disagree with that.  For example, I happened to
upgrade the firewall on a Windows XP system recently.  The upgrade was
completely automatic.  What was _not_ automatic was that the internet
connection should have been temporarily terminated _during_ the
update.  What happened is that the firewall shutdown for the upgrade,
while the internet connection was open.

It took a couple of seconds for the first worms and viruses to
penetrate (iow: find) the machine.

The lesson here is that security for desktop systems is more important
than ever.

I agree that this is will also become very important on (networked)
embedded devices.

Thanks,
Marcus