Re: Contributing to the Developement of Debian Installation Program
Hi,
> Few days ago I discussed in news with a Debian developer about
> contributing to the development of the installation program.
>
> I've used Debian since 1.2 and I've always feeled that it's a great
> operating system. IMHO the weakest point is installation: it's too
> much work even for an experienced user and it can be very hard for a
> person who doesn't know much about PC hardware. Personally I have
> never feeled that this is a big problem but many people seem to disagree
> with me. So I'd like to help.
This is also my opinion. Currently, I am writing a completely new Install
process. I think, one of the main reasons, why RedHat, Suse et al
are getting more "seats" is, that they have an install process, which
is more convenient to the novice, impatient user. Although *I* do
like dbootstrap & deselect, I think it is rather an obstacle for a
person getting involved the first time. My first time with Debian was
2.1, so as you can see, I am *very* new to Debian. Luckily, I am
*not* new to Linux and Unix in general.
Currently, I am poking around the 'boot-floppies" package, to see
what is done there. It is a bit confusing, so I am having a fairly
"fun" time with it :-) Here is what I intend to do:
1. I assume a CDROM in the target system, which can be
detected at boot-time.
2. I want to create a boot-floppy ( single disk ) loading the
kernel and a small initial ramdisk. The ramdisk contains
*only* the tools necessary to mount a iso9660 filesystem
into the ramdisk-root tree, along with the obvious stuff for
the boot-process itself. The CDROM will contain the same
for direct CDROM boots. I want to create as much space
as possible on the boot-disk for the upcoming Linux
kernels, which I assume to be larger than the current
kernels to some extend.
3. The CDROM-based ( live ) filesystem will contain the libs
and tools for the installer itself, along with stuff to start
networking and things like that.
4. The installer will ( at first ) be only text-based initially, for
two reasons. First, it will allow a working installer earlier :-)
Second, for an X-based install, I intend to wait for Xfree 4.0,
since it makes handling of the various graphics cards easier -
only one server, loading driver-modules etc...
In order to *avoid* dselect in the install process initially, some
dpkg-profiles will be created, allowing things like
'devel-system', 'X-workstation' or 'gnome-workstation etc.
Similar to the current install process.
I will do this for x86 initially, along with Alpha being the second
and SPARC (4m/c) being third ( I have these archs at hand ).
Currently, I am at the very beginning, but I am just starting working
full-time on this ( yes, there is a commercial motivation, but all
my code will be GPL - I can give details on demand ).
I have the CeBit-show coming up ( I am German ), so I will be
interrupted for about the next two weeks, but after that, I
should be able to come up with code.
Currently, I am not involved in any way with the Debian project,
but if there is interest in this, I will happily change that :-)
It is my attitude to come up with code first, before I apply for
some 'developer' status within Debian.
I didn't even intended to do this 'pre-pre-announcement', but
maybe this helps in case someone finds my approach
interesting enough to help. I would sure appreciate help.
> The developer asked me to post to this list and ask what there's to do
> and told me that at least there's testing.
Great. In case you are interested, I would appreciate comments and
testing.
Sincerely,
Thomas Weyergraf
--
-------------------------------------------------------------------------------
Thomas Weyergraf kirk@colinet.de
Reply to: