Re: where do NEW packages go?
On Sun, 19 May 2002, you wrote:
> On Sun, May 19, 2002 at 01:41:56PM +0200, Eduard Bloch wrote:
> > Why do you use the term "port" here? "Debian" defines a base structure
> > for an complete operating system.
> Because we aren't porting GNU/Linux to a new processor. Debian only
> defines the base structure for the GNU/Linux operating system. Not for
> *BSD and GNU/Hurd. That's why *BSD and GNU/Hurd ports are having
> That's true. But the infrastructure was designed for GNU/Linux
> only. Now there are GNU/Hurd and *BSD too. Because of that, we need to
> rethink the infrastructure. I want GNU/Hurd to be part of Debian, but
> to make that possible there are changes needed for Debian.
> Why is it that difficult to see that Debian GNU/Hurd is a distribution
> of the GNU operating system? And Debian BSD a distribution of the BSD
> operating system?
There lies the core of the 'misunderstanding'.
The difference is that many people involved in Debian, will say that
Debian, while being a distribution following a set of well-established
procedures and well-defined policies (that make it as great as it is
IMHO), *is an OS* in its own right.
That is is a stronger word than 'distribution', and personally I think
the word is justified, because Debian delivers a more coherent system
than most distributions of a base OSes + tools; one of the reasons
probably being that all of its packages follow similar procedures for
development and release.
On the other hand, you seem to say that Debian is merely a distribution
of an existing (real or more abstract) base system, be it *BSD,
GNU/Linux or GNU/Hurd, plus a set of application packages. And,
consequently, that well-established practice in a base system
(filesystem conventions) should always take precedence over Debian's
I hope the difference is clear, and perhaps that already helps to
smoothe things a bit, but I'd like to add my own two cents to this as
If, for a minute, you'd try to adopt the 'Debian as an OS' point of
view, you'd see that in order to get things like changes to filesystem
conventions or patches to important packages accepted, you'd have to
play by Debian's rules. Even if those changes have already been fully
accepted in the abstract base OS (in this case GNU, or more precisely,
Also, if you consider that people have chosen to look only at the ABI
aspect of the differences between the various base OSes (to determine
whether or not that's a good thing is left as an excercise for the
reader), which allowed them to model different base OSes the same way as
machine architecture ports (hence using i386-hurd, as a slightly
different ABI than the 'normal' i386 one for debian), you'll also see
that because a release is done for all architectures at once, Debian
GNU/Hurd cannot follow its own schedule independent of Debian GNU/Linux,
because they only differ in ABIs/machine architectures ('ports' is an
awfully confusing word in this context).
If you want to hold a plea for modeling different base OSes independent
from different architectures, be my guest -- I think there could be some
good arguments that could be considered, I don't know.
But one thing for you though: please allow people to ask you to change
your point of view, even if just for a minute, in order to try and
understand things as they are. If you remain convinced of your own way
of modeling the world, you'll never learn a thing that doesn't fit the
model, nor will your model ever advance.
Essentially, that's the same thing you asked that from the Debian group
(albeit not too politely). So, it's only reasonable that you do so as
E-Advies / Emile van Bergen | email@example.com
tel. +31 (0)70 3906153 | http://www.e-advies.info
To UNSUBSCRIBE, email to firstname.lastname@example.org
with a subject of "unsubscribe". Trouble? Contact email@example.com