Re: Debian Installer
On Mon, Sep 16, 2002 at 06:11:06PM +1200, Philip Charles wrote:
> > If we want to spawn GNU version 1.0, that's fine; since those
> > people don't have to deal with other developers changing things beneath
> > them. But in Debian, a lot of infrastructure has to be in the right
> > place at the right time before a release can ever occur. (Just witness
> > the recent libpng, libstdc++ and Perl goodness.)
> They were nasty. However, if the pool system were to be more extensively
> used then there could be more frequent Hurd releases. What might be
> unstable for Linux could be quite stable for the Hurd and v.v. It would
> be a matter of creating proper Packages files for the Hurd.
The idea is that we're got a lot of dependencies of Debian's
system itself. Right now, policy is going through some big motions; and
debconf is going through big changes. So either we release a version of
woody far later than woody actually released, or try to release with a
partially broken Debian system.
> > Besides, on a personal note, I am quite enamored by the fact
> > that I can sit down to any Debian hamm/slink/potato/woody machine and
> > expect it to work exactly like any other hamm/slink/potato/woody
> > machine. Having independant releases would break that universality that
> > we get with the "Universal Operating System."
> I would not expect Sarge/Hurd to be the same as Sarge/Linux, otherwise why
> bother with the Hurd at all? However, I would expect quite a lot to be
> common to both.
The reason to bother with the Hurd is that a) the kernel is not
broken, and b) the Hurd has significant technical advantages.
However, that does not mean that we do not use the same
packages, nor the same programs. Although specific pieces of the
operating system would be different, I would not expect large pieces of
it to differ from Debian itself.
Remember, although GNU isn't Unix, and GNU certainly isn't
GNU/Linux; it definitely looks like both.