Re: Foundation (was Re: The Kernel (fwd))

[George Bonser <grep@oriole.sbay.org>]

On Sun, 25 Jan 1998, Rick Jones wrote:

> This is just unreal.  I must be in the twighlight zone.
> 
> We "had", past-tense, people on the lists that wanted to discuss these
> issues and possibly take on the process of getting it done, but you shot
> their suggestions down at every turn because "you" didn't want to do it
> that way, saying there are larger issues...blah, blah, blah..., I asked
> you 3 times to tell us what you are doing and you blew me off.

Scuse me? Who wanted to work on boot disks? Let's back up a second.  
There is much groundwork to be done before anything is coded. Before
anyone starts working on a package, I would hope that they become
familliar with the debian package making process and start a dialog with
the debian maintainer if that software exists within Debian. This might
save you a lot of time considering that packages will be submitted as
diffs to the pristine source.

I am not trying to kill the creativity, I am trying to save people time by
cluing them in on the groundwork they need.  Having a great idea about how
to improve xyz-1.2.3 and spending a month working on it might not amount
to much if you can not properly create the package and if a new release
from upstream is issued just before you are finished.  Get a dialog with
the Debian person, most of these people also have dialogs with the creator
of the program. If a completely new way of doing something is just around
the corner, chances are good that the Debian maintainer might know about
it in some cases.  Other packages are very static.  Some are orphaned.  It
is possible for the SEUL maintainer to ALSO be the Debian maintainer of a
package if it is orhpaned.

> 
> Now I find out that you are just putting together some install disks
> that Debian is going to put out anyway, if all you want to do is get an
> install, instead of hashing out a valid structure and time table, and
> getting people working on things.  

I wanted to learn how the install process worked and I wanted to get the
ball rolling with SOMETHING.  I came here and saw a bunch of people
standing on a hilltop looking at a rock that they were going to roll
arguing about where exactly they wanted it to stop.  Bruce Perens made a
comment that all he had seen is a lot of talk. Once it was decided that
Debian would be chosen as the basis, that provided a general direction for
the rock to roll in so I decided to push it and get it started.  We can
guide it on the way down but we need to get it started.

> You worked for 3 days on getting a combination of libs together to
> create a set of disks that somebody else is already working on, pushing
> people away from the list that wanted to get involved?  This was your
> "larger issue"?  Working on something that we won't even be using, since
> we will be using the Debian's initially?  Which, you said yourself.

Right, someone else is working on it.  Someone in Debian. An important
concept of SEUL is to make the system easy to install.  An important part
of the install process is the boot disks.  Understanding them is key to
changing them.  Having something that does not work at first is great
because in the course of fixing it, you learn a lot about how it works.
 
> This was the "larger issue" you had to devote all your time to instead
> of hashing out the boot process and what should be in the kernel, so we
> could apply it when the real disks are ready?

We are not ready to code anything yet.  I have said several times to go
ahead and discuss how you think it should look.  In the meantime, I will
go about creating the baseline. Since the baseline is not going to have
any modifications, participating in the discussions would absorb time best
spent doing that.  Other people are in charge of the parts of the
distribution.  Discussions of the kernel changes in the context of 0.1
mean nothing because there won't be any, that is all I was saying. The
changes would be incorporated starting with 0.2

> You could have spent all this time getting a list for the "Linux core"
> together.

And I will as soon as I see what 2.0 is going to finally look like.  I had
a core-0.0.1 directory up at the same place as that boot disk for three
days.  Did you look at it, comment on it, suggest additions or deletions?
Nobody else did either.  I removed it today and, as soon as I get done
here am going to write just that.  A suggested list for Core.  

> Writing config dialogs, documenting the different structures
> needed to make SEUL work the way we want it to, you could have been
> talking to kernel developers about the various patches and what to
> expect in the kernel during the development of SEUL so we might have
> gotten somebody working on a program to take advantage of new changes,
> or whatever.  There are a lot of things you could have been getting
> *DONE* for SEUL.  But this was your idea of a priority?

Hey, I tried going down that road when I threw out the idea of
incorporating patches. That is not my juristiction. I will incorporate
whatever is decided to be incorporated by the leadership.  I will also
express my concerns.  I DO NOT have veto power over what goes into the
distribution.  My job is to make sure it all comes together and works. If
discussion turns to something that is potentially troublesome, I will
express my opinion or alert people so that they at least KNOW about the
potential problem.  If it is decided to go ahead, I will set aside my
personal feelings on the matter and do what is best for the team.  It does
the leadership NO GOOD though to not have potential problems pointed out
and mitigated in some way so that they so not cause trouble for the end
user.  EU is the heart of SEUL. If we put them in a risky situation, we
must give them a safe path.  I just want to make sure we do that.

> 
> Doesn't that fall under the install leaders area anyway?  Why are you
> working on it?

Because as far as I know, nobody else was.  


George Bonser 
If NT is the answer, you didn't understand the question. (NOTE: Stolen sig)
http://www.debian.org
Debian/GNU Linux ... the maintainable operating system.