Propossed Project: Odyssey

[Alex de Landgraaf]

Note: this may be a bit long, but i hope you will read it and give
feedback about your opinon on the subject(s) discussed. And sorry for
crossposting, just wanting everones 2cents. forgive me :)

Fellow Debian-users/developers!

Lets start by telling something about myself and the reason i am posting
this proposal.

For some time now i work (ICT) at a highschool in Amsterdam, the
Netherlands. It's not a very large department (mostly
social/juristiction education) and, well, the cashflow is drying up a
bit. My colleague (actually, he's my boss) and i talked about this
yesterday and this set of some kind of spark in me. This is the result
of that small spark.

Normally, our school pays about 100k guilders ($50k) each year for
licenses, mostly thanks to a not further mentioned Operating System.
Now, the maker of this OS has a nice plan to squeeze more cash out of
it's installed base throughout the world. Our school is going to have to
pay nearly 1000k guilders ($500k) for licences. Thats half a million
dollars people, for one year. A simple small public highschool has more
than enough way's to spend this without problems with licenses. And this
school isn't the only one...

Now, my boss is going to get most of the heat, and he and his managers
are looking for alternatives. He hit the subject: as a linux/BSD geek
(not using $$$-OS, simply refusing ;) he asked if Linux was an
alternative for this $$$-OS. The reaction was forseeable. YES! But then
i put my support/propaganda aside.

Ofcourse my thoughts went directly out to Debian, as a personal
favorite. It would take some time to create an enviroment where all
these (simple) users would be content in, it would take even more time,
testing and support for them to be glad with it, but that wouldnt be
much of a problem. The problem is that it should be simple to
install/administer/configure/upgrade (not everyone likes digging in
/etc/ or answering all these questions to install an OS), and using it
should feel like second nature! Most of these things are not yet here or
easy in Debian GNU/Linux, or in any distro in general, and that is a
pity, from a user-perspective.

Here, with Debian, we have the collective packages to make these
systems. Throughout the whole open-source community, there are countless
efforts taking place to make great software, nice to look at and easy to
use. These are succeding moderatly, and many of us use versions of them
(evolution, office clones, etc). There have also been companies
(Progeny, Corel) who have tried to make easy installers on top of
Debian. These have also moderatly succeded, but these efforts now all
but died out, as we all know...

How about updating? Apt-get and the frontends do this very nicely for
Debian, excellent as a matter of fact. A (small) problem here is the
huge amount of packages, nothing a good sorting and filtering wouldn't
be able to better, imho.
Configuration has been an Achilles' tendon/heel of all *nix in my humble
opinion. Editting configuration files is still the most reliable way to
change your system. There are some tools (YAST, Linuxconf, Webmin, etc)
that attempt to change this, and they too work (only) moderatly. For
now...

But now for the main part: the user. If you are reading this, than i
don't regard you as a real user anymore. Real users use computers as a
tool to do their humble wordprocessing, e-mail, and surfing. Some time
ago, i read something about having your aunt being able to use Linux,
and she should be able to. Not because it is free, but because it is
BETTER! With that i mean more coherent, easier navigation, the ability
to do what the user wants: to accomplish a task. As easy as possible. As
fast as possible. No strings attached.

Now, where am i going to?

We have the building blocks, the collective effort and the possibility
to make the world a better place: making a truely easy to use Free
Operating System. This is long due from our (the community's) side, but
it just doesn't seem to work out as well as everyone hoped for. Now,
with the tentacles of this Other OS grasping and squeezing the cash and
freedom from our fellow human beings, should we just let it happen?

Therefore, I propose a new project: Project Odyssey.
To take the long road from where we (the geeks) now are to create a true
alternative for the user. For long the road it will be, it may never be
accomplished. But with everything coming together as stated above, and
no viable alternative, I thought it must be better to act than to hold
still.

The goal of this propossed project:

To create a viable alternative to our main competetor, based on Debian
GNU/Linux.

Not in the way of power and possibilities, but to help the users to
accomplish their tasks that they use the pc for.

Not in the way of "just" an installer, or "just" a e-mail client, but as
a total solution. (sounds very business IT, i know...)

Not made by a company, but as a united community, knowing that although
we may not get paid for our work, the reward afterward will be worth
every second of work.

Not as an OS made for geeks to administer, and users to use, but to make
an OS simple enough for even users to be able to understand. And even
simple enough for them to install and configure to their needs.

Not as a system just used at school or work, but that is easy enough to
use at home. No, not easy enough. easier! For every user!

I know that this will take an huge effort, if we accept this project. It
will have all sorts of problems that i couldn't even dream of. But NOW
it is possible...

Now, i will halt my rant/manifesto/speach, and let you give your
opinions and feedback. Speak now, or forever hold thy tongue! :)

greetings, TIA,

Alex de Landgraaf (aka AdeL)
the Netherlands

--
To UNSUBSCRIBE, email to debian-us...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org

[Scott Dier]

> Therefore, I propose a new project: Project Odyssey.
> To take the long road from where we (the geeks) now are to create a true

I would attempt to not fork whatever work you do, but provide packages
that work within debian for these things that track unstable and could
be released with a future release of debian.

Also, I think there is a lot of need for this sort of work. I've been
talking to some admins here about starting a project for backporting
sets of packages from unstable to stable to do 3-6month update packs
after stable comes out, the response has been positive.

--
Scott Dier <die...@ringworld.org> <sd...@debian.org>
http://www.ringworld.org/ #lin...@irc.openprojects.net

Just say NO to Product Activation!

[Bernd Eckenfels]

On Wed, Oct 24, 2001 at 01:12:02AM +0200, Alex de Landgraaf wrote:
> To create a viable alternative to our main competetor, based on Debian
> GNU/Linux.

Do not forget www.boxedpenguin.com

IMHO you should not use the Debian Project as a start for your trip, go out
and do it in public.

Greetings
Bernd

[Scott Henson]

> Therefore, I propose a new project: Project Odyssey.

> The goal of this propossed project:
>

> To create a viable alternative to our main competetor, based on Debian
> GNU/Linux.

I would be willing to help on this project. I have been thinking about
something like this for a while now. Feel free to email me about the
project. I know C, C++, perl, and shell scripting.

[David N. Welton]

Scott Dier <sd...@debian.org> writes:

> > Therefore, I propose a new project: Project Odyssey. To take the
> > long road from where we (the geeks) now are to create a true

Do you want to create something that 1) can be *used* by anyone, or
something that can also be 2) installed/admined by 'anyone'? The
first is doable now.

You might have a look at

http://www.newsforge.com/article.pl?sid=01/08/10/1441239

for the first case.

> I would attempt to not fork whatever work you do, but provide
> packages that work within debian for these things that track
> unstable and could be released with a future release of debian.

Debian's goals are not to provide ease of use at all costs. And our
current packages aren't flexible enough, in my opinion, to provide
goal number 2) out of the box.

--
David N. Welton
Consulting: http://www.dedasys.com/
Free Software: http://people.debian.org/~davidw/
Apache Tcl: http://tcl.apache.org/
Personal: http://www.efn.org/~davidw/

[Wilmer van der Gaast]

Yesterday at 18:21:63 you wrote the following wise words:

> Also, I think there is a lot of need for this sort of work. I've been
> talking to some admins here about starting a project for backporting
> sets of packages from unstable to stable to do 3-6month update packs
> after stable comes out, the response has been positive.
>

Now THAT sounds interesting! I surely hope you'll put it on an apt
source on-line somewhere when finished? I do the backports myself now,
but my system is quite a mess, now.. :-( And some programs just take up
too much time to compile.

--
*=-+-______________________
|lintux-@t-lintux-d0t-cx:
| http://www.lintux.cx/ |
:http://www.algoritme.nl|
~~~~~~~~~~~~~~~~~~~~~~-+-=*

[Wilmer van der Gaast]

Today at 01:12:63 you wrote the following wise words:

> Now, my boss is going to get most of the heat, and he and his managers
> are looking for alternatives.
>

I wish the ppl at my school were that smart as well...

There are some Debian-based distro's with very nice installers and
things like that already. One of then is Progeny, we have Corel... The
other one can be found via http://www.debianplanet.org/. Problem is that
Progeny is dead now... :-/

Just install Woody as soon as it's stable. It has a nice desktop, a nice
browser (Konqueror, you can even make it support things like ActiveX, so
no blahs about not-supported-by-browser..), and you can install
StarOffice for people who want to write something.

The fact that Linux is more difficult to maintain might just be a bit of
a good thing for schools. On schools the pupils seem to like to mess up
everything they can, and now they can't do that anymore... And cfengine
is said to be a good program for the sysop to configure lots of systems
at once.

Greetings,

Wilmer v/d Gaast.

[Cesar Mendoza]

I agree with Scott. With the help of a good admin any body can use Linux
right _now_. Remember that for a user like you said the computer is a tool
so as long as the admin makes sure that the applications the users need
are there it's only a matter of time before the user feels at home with
the new system.

As an example, I put together a sid box for my wife with KDE and I explained to
her what were the equivalents in Linux to the applications the she used
on Windows and now she can do the same thing that she did before. Of
course I made sure that the machine runs without problems and update it
from time to time, but from her point of view everything is fine.

Why don't you take a group of people in your organization and create a
pilot project, so you can determine what it takes to switch everybody
to Linux. As long as your users have good admins to help them I don't
see any problem to switch them right _now_.

What you want to do it's more easy for an organization that can afford
admins, than to a home user that it's his/her own admin (That is why I
always laugh when an error message ask me to contact the administrator :) ).

Bye
Cesar Mendoza
http://www.kitiara.org
--
"A scientist once wrote that all truth passes through three stages:
first it is ridiculed, then violently opposed and eventually,
accepted as self-evident."
-- Schopenhauer

On Wed, Oct 24, 2001 at 10:16:33AM +0200, David N. Welton wrote:
> Scott Dier <sd...@debian.org> writes:
>
> > > Therefore, I propose a new project: Project Odyssey. To take the
> > > long road from where we (the geeks) now are to create a true
>
> Do you want to create something that 1) can be *used* by anyone, or
> something that can also be 2) installed/admined by 'anyone'? The
> first is doable now.
>
> You might have a look at
>
> http://www.newsforge.com/article.pl?sid=01/08/10/1441239
>
> for the first case.
>
> > I would attempt to not fork whatever work you do, but provide
> > packages that work within debian for these things that track
> > unstable and could be released with a future release of debian.
>
> Debian's goals are not to provide ease of use at all costs. And our
> current packages aren't flexible enough, in my opinion, to provide
> goal number 2) out of the box.
>

--

[Timothy H. Keitt]

Better yet, lets convince package maintainers not to unnecessarily
update all their dependencies to the latest libs in unstable so that
packages can be easily backported with 'apt-get -b source ...' My guess
is that 60-90% of the packages in unstable do not require the latest lib
versions to build, but that maintainers are defaulting their
dependencies to be >= the latest version in unstable for no reason (of
course, package name changes and package reorganization can throw a
wrench into things). If maintainers default to only depend on what is in
stable whenever possible, many many deb packages would compile just fine
on both stable and unstable.

T.

Wilmer van der Gaast wrote:

>Yesterday at 18:21:63 you wrote the following wise words:
>
>>Also, I think there is a lot of need for this sort of work. I've been
>>talking to some admins here about starting a project for backporting
>>sets of packages from unstable to stable to do 3-6month update packs
>>after stable comes out, the response has been positive.
>>
>Now THAT sounds interesting! I surely hope you'll put it on an apt
>source on-line somewhere when finished? I do the backports myself now,
>but my system is quite a mess, now.. :-( And some programs just take up
>too much time to compile.
>

--
Timothy H. Keitt
Department of Ecology and Evolution
State University of New York at Stony Brook
Stony Brook, New York 11794 USA
Phone: 631-632-1101, FAX: 631-632-7626
http://life.bio.sunysb.edu/ee/keitt/

[Mark Carroll]

On Wed, 24 Oct 2001, Timothy H. Keitt wrote:

> Better yet, lets convince package maintainers not to unnecessarily
> update all their dependencies to the latest libs in unstable so that

(snip)

Absolutely - I've been far from convinced by a lot of the requirements. I
normally like to stick with "stable", but once you want a couple of later
packages, following all the dependencies can make it more trouble than
it's worth.

-- Mark

[Colin Watson]

On Wed, Oct 24, 2001 at 11:43:50AM -0400, Timothy H. Keitt wrote:
> Better yet, lets convince package maintainers not to unnecessarily
> update all their dependencies to the latest libs in unstable so that
> packages can be easily backported with 'apt-get -b source ...' My guess
> is that 60-90% of the packages in unstable do not require the latest lib
> versions to build, but that maintainers are defaulting their
> dependencies to be >= the latest version in unstable for no reason (of
> course, package name changes and package reorganization can throw a
> wrench into things).

It's relatively rare for build-dependencies to be updated in the way you
describe, except for cases where packages are reorganized or their
behaviour when other packages are built against them changes. If
anything, build-dependencies are usually too loose.

If you're thinking of dependencies on libraries such as libc6, those are
changed because (and usually only when) they have to be, that is when
the library changes its interface in such a way that binaries compiled
against the new version won't necessarily run with the old version.

> If maintainers default to only depend on what is in stable whenever
> possible, many many deb packages would compile just fine on both
> stable and unstable.

This is a useful goal. Just don't confuse binary dependencies with
source dependencies. A package may depend on libc6 (>= 2.2.4-2), but if
you built it on stable it might well compile fine and produce a package
that depends on libc6 (>= 2.1.3-18). That isn't to say that the first
package would work with the old libc6 if you just installed it directly.

(Follow-ups set away from -project - this is a technical discussion.)

--
Colin Watson [cjwa...@flatline.org.uk]

[Alex de Landgraaf]

First, i would like to thank everyone for their opinions, some are very
interesting. I'll try to cover them all in one post, forgive me if i
don't answer everyone personally :)

Okay, lets tackle the hardest first:
- Why put a lot of effort to start a project from scratch, when that
effort could be used to make Debian better?

Note, i never said that this should be started from scratch. The Debian
we all use would be perfect to start from. But i don't believe that
Debian will be a desktop from a users point of view. I don't know how
Debian 3.0 will look like, installer, default packages, configuration
and general ease-of-use, but now Debian is a perfect system: for a
geek...
Also, Debian focuses on its rock-solid stability and sturdiness. This is
wonderful, but Debian isn't moving very fast, IMHO. There are so many
packages avalible and they all need to work well to be added to the
distribution. Using Woody would (pun intended :) work, but it still is
quite a hassle to set things up "like they should be". My proposal is to
only take those packages that would be needed in an end-user system to
create a more sleek, fine-tuned distribution where the focus lies on
effectiveness towards the end-user.
And effort used in this project wouldn't be effort not waisted on
Debian. Who says ideas could not also be fed back into Debian?

- Why use Debian? (use mandrake!):

Important point, but why isn't mandrake being used by all those users
today? I don't know, but keeping a system up-to-date is sure a lot
easier with Debian. And what will happen to mandrake if the company
cashflow dries up?
I personally havn't tried the latest Mandrake (or the new RedHat, for
that matter). Does anyone have more pro's and con's about these
distro's? Are they an alternative?

- www.boxedpenguin.com:

I like that penguin :)
This is a neat project, i'm aiming for something in that direction, but
instead for users (as apossed to developers/engineers). If my propossed
project is accepted, we could use source/ideas from this project.

- Make the project in public! (No Debian Project):

Maybe. I don't know why we should or shouldn't make it a Debian Project.
Seemed like, as it is built ontop of Debian, it should be, and i
wouldn't like it to be "just" another project in the dark. Maybe it's
just fear of drifting to far :)
How does Debian stand towards its projects? Any first-hand experience
with the difference between a project "here" and "outside"?

- Should users be able to administer their own computers? Most can't
even program a VCR!

Yes. No :)
The problem is that most Operating Systems (and VCR's for that matter ;)
have way to many features. KISS: Keep It Simple, Stupid :)
And I wasn't actually talking about NT, more about having being able to
install the system at home. Now that should be simple ;)

- www.libranet.com

Havn't heard from it before, downloading it as i write this. No use in
doing things twice! I'll test it, and maybe put my efforts in it
instead. Depends on how good it is :)

There was also a post about Largo (i read it before, i think, but thanks
anyway). Note that these people mostly have never used that Other OS at
work before. The system they use is also highly customised, and although
it works for them, creating the same system would take quite some work
(not to mention the problems they would have if a key sysadmin would
leave...)

Used by anyone, Installed by anyone, Admined by anyone. Yups, that sums
most of it up. But like i said above, it should be kept simple! Even
though the task of creating and completing such a project is everything
but simple :)

Also, there was a mention about Progeny (and that it is dead :). Thats
what i ment about having a company make distro's (and kewl stuff in
general): it's great while it lasts :)
If libranet isn't what i am aiming for, and if the will to help our poor
users is here, we could learn (plunder?/loot?) Progeny's idea's. It may
be wise to continue where Progeny stumbled.

The idea of update pack's sounds very neat, would solve a large part of
the problem (for both users and the regular Debian-geeks).

Oh well, we'll see how things turn out. Great to see all these
suggestions flying around!

Alex de Landgraaf

[Daniel Toffetti]

On Tuesday 23 October 2001 20:12, Alex de Landgraaf wrote:
> Note: this may be a bit long, but i hope you will read it and give
> feedback about your opinon on the subject(s) discussed. And sorry for
> crossposting, just wanting everones 2cents. forgive me :)

<snip>

> Now, i will halt my rant/manifesto/speach, and let you give your
> opinions and feedback. Speak now, or forever hold thy tongue! :)

Does somebody know anything about any common format for configuration
files? Does any sense to propose this, somebody knows if there are
references of this in Linux Standard Base ?
Do makes any sense even the sole idea of such format ?

Please bear in mind that I'm not a skilled admin, Linux developer or
package maintainer... anyway I'm curious about this issue.

Daniel
--
"There is no spoon..." - The Matrix

[Eric G. Miller]

On Wed, 24 Oct 2001 20:39:50 -0300, Daniel Toffetti <dto...@softhome.net> wrote:

> Does somebody know anything about any common format for configuration
> files? Does any sense to propose this, somebody knows if there are
> references of this in Linux Standard Base ?
> Do makes any sense even the sole idea of such format ?

ASCII text? There's some commonality about using "#" to start comments.
Friendly programs give verbose examples and descriptions in the examples.

--
Eric G. Miller <eg...@jps.net>

[John Purser]

FYI there's already an Odyssey project in the windows software world. It's
a client server point of sale project for a large wholesaler that should be
for sale to the public next year. Not sure if this matters, just passing
along the flotsam and jetsam.

John Purser

- www.boxedpenguin.com:

- www.libranet.com

Alex de Landgraaf

[Daniel Burrows]

On Thu, Oct 25, 2001 at 01:21:08AM +0200, Alex de Landgraaf <alex...@xs4all.nl> was heard to say:

> Also, Debian focuses on its rock-solid stability and sturdiness. This is
> wonderful, but Debian isn't moving very fast, IMHO.

The reason this is a problem, IMO, is primarily the recent explosion
of "end-user"-type free software. Certain other operating systems only
release a major new version every few years, and they seem to be doing
fine in terms of being up-to-date-enough.

(that isn't to say that a project which focuses on backporting
GUI-ish to the current stable wouldn't be useful)

> There are so many
> packages avalible and they all need to work well to be added to the
> distribution. Using Woody would (pun intended :) work, but it still is
> quite a hassle to set things up "like they should be". My proposal is to
> only take those packages that would be needed in an end-user system to
> create a more sleek, fine-tuned distribution where the focus lies on
> effectiveness towards the end-user.

It's not obvious to me that you couldn't just fork/rewrite boot-floppies,
build a custom set of CDs with the subset of the archive that is useful
to your project, and write a non-sucky system configuration tool.

Daniel

--
/-------------------- Daniel Burrows <dbur...@debian.org> -------------------\
| "Progress just means bad things happen faster." |
| -- Terry Pratchett, _Witches Abroad_ |
\------------- Debian GNU/Linux http://www.debian.org -- Because. ------------/

[Adam Heath]

On Wed, 24 Oct 2001, Timothy H. Keitt wrote:

> Better yet, lets convince package maintainers not to unnecessarily
> update all their dependencies to the latest libs in unstable so that
> packages can be easily backported with 'apt-get -b source ...' My guess
> is that 60-90% of the packages in unstable do not require the latest lib
> versions to build, but that maintainers are defaulting their
> dependencies to be >= the latest version in unstable for no reason (of
> course, package name changes and package reorganization can throw a

> wrench into things). If maintainers default to only depend on what is in

> stable whenever possible, many many deb packages would compile just fine
> on both stable and unstable.

This shows a deep misunderstanding of the way shared libraries work.

If a library is changed, and uploaded, it may require an update to its
/var/lib/dpkg/info/*.shlibs file. When pkgs are then rebuilt against that
library, the pkg-version dependency info is taken from this file. That is
what causes newer libraries to be depended on. It is not a conscious effect
on the maintainer.

Additionally, if a new version of a package comes out, that depends on a new
library, do you think that the new package should not be allowed into debian,
on the fact that backporting to an older version of debian would be
problematic? That line of thinking means nothing would ever be upgraded.

[Adam Majer]

On Thu, Oct 25, 2001 at 06:38:45PM -0500, Lex Spoon wrote:
>
> I think for the kind of money you are talking about, you can hire
> administrators who know Unix. Honestly: wouldn't you want professional
> (or at least competent) adminstrators for Windows, too? Ease of Windows
> administration is a myth; hardware upgrades are fairly minor
> administration, and most people who have tried know how difficult this
> can be under Windows. It only gets worse if you want to set up network
> services, and perhaps even (gasp) some sort of security to protect users
> from each other.
>
> So, real question is whether you can make *using* the systems easy, not
> whether installing and maintaining them can be easy. The answer is
> probably yes, but your admin will have to make it a priority (which many
> Unix admins do not).

I would like to agree with this 100%. I've heard many times that people (students) are running BackOriffice and related stuff on Windows machines
at school and the "admins" don't even know what is going on. Thus a Unix sys admin that know what he/she is doing is a key here - IMHO Debian is
very easy to administer (for a sys-admin) as you do not have to worry as much about config file defaults as they are usually sane. Providing a
good job and saving money -- a win-win situation.

OTOH, making a new distribution that is "easy" to administer would take months and the easiness would probably come at a cost of less flexibility
(or at least you would have to go back the the good old /etc dir).

- Adam

[Mike Williams]

>>> On Wed, 24 Oct 2001 10:34:02 +0200 (CEST),
>>> "Wilmer" == Wilmer van der Gaast <lin...@lintux.cx> wrote:

Wilmer> There are some Debian-based distro's with very nice installers and
Wilmer> things like that already. One of then is Progeny ... Problem is that
Wilmer> Progeny is dead now... :-/

My understanding is that Progeny (the company) is still alive, and that
they wish to migrate back towards a standard Debian distro. Anyone know if
there are plans to integrate the Progeny installer into Debian-proper?

--
cheers, Mike

[Michael Epting]

It's there (in unstable). See autoinstall. (I haven't tried it and
honestly wouldn't know how to begin).

On Fri, Oct 26, 2001 at 12:22:11PM +1000, Mike Williams wrote:
>
> My understanding is that Progeny (the company) is still alive, and that
> they wish to migrate back towards a standard Debian distro. Anyone know if
> there are plans to integrate the Progeny installer into Debian-proper?

--
Michael Epting (ept...@ix.netcom.com)

[Timothy H. Keitt]

Adam Heath wrote:

>On Wed, 24 Oct 2001, Timothy H. Keitt wrote:
>
>>Better yet, lets convince package maintainers not to unnecessarily
>>update all their dependencies to the latest libs in unstable so that
>>packages can be easily backported with 'apt-get -b source ...' My guess
>>is that 60-90% of the packages in unstable do not require the latest lib
>>versions to build, but that maintainers are defaulting their
>>dependencies to be >= the latest version in unstable for no reason (of
>>course, package name changes and package reorganization can throw a
>>wrench into things). If maintainers default to only depend on what is in
>>stable whenever possible, many many deb packages would compile just fine
>>on both stable and unstable.
>>
>
>This shows a deep misunderstanding of the way shared libraries work.
>

I think you mean lack of understanding of how dpkg handles shared
library dependencies, which is indeed the case. Someone also pointed out
that run-time (binary) and build-time (source) dependencies are quite
different, which is also true.

>
>
>If a library is changed, and uploaded, it may require an update to its
>/var/lib/dpkg/info/*.shlibs file. When pkgs are then rebuilt against that
>library, the pkg-version dependency info is taken from this file. That is
>what causes newer libraries to be depended on. It is not a conscious effect
>on the maintainer.
>

As was pointed out previously, this only effects binary packages; I was
referring to source dependencies.

>
>
>Additionally, if a new version of a package comes out, that depends on a new
>library, do you think that the new package should not be allowed into debian,
>on the fact that backporting to an older version of debian would be
>problematic? That line of thinking means nothing would ever be upgraded.
>

I think if you read my post a bit more carefully, you would see that my
point was that if the upstream source builds on stable without major
loss of functionality, then the debian source package should not
unnesessarily depend on unstable so that it can produce binary debs for
both stable and unstable. It would be interesting to see how many source
packages in unstable will currently build with 'dpkg -b' on stable.
Perhaps more than I think. My experience was actually with trying to
build source packages from unstable on progeny (newton) and there came a
point in time when it was generally no longer possible because of unmet
dependencies (and fufilling those source dependencies basically meant
switching to testing/unstable). Also, note my use of "unnesessary" and
"whenever possible"; of course dependencies on unstable will sometimes
be necessary and 'dpkg -b' backporting will not be possible.

T.

--
Timothy H. Keitt
Department of Ecology and Evolution
State University of New York at Stony Brook
Stony Brook, New York 11794 USA
Phone: 631-632-1101, FAX: 631-632-7626
http://life.bio.sunysb.edu/ee/keitt/

[Anders Jackson]

Alex de Landgraaf <alex...@xs4all.nl> writes:

> Note: this may be a bit long, but i hope you will read it and give
> feedback about your opinon on the subject(s) discussed. And sorry for
> crossposting, just wanting everones 2cents. forgive me :)
>
>
> Fellow Debian-users/developers!

[...]

> Therefore, I propose a new project: Project Odyssey.
> To take the long road from where we (the geeks) now are to create a true
> alternative for the user. For long the road it will be, it may never be
> accomplished. But with everything coming together as stated above, and
> no viable alternative, I thought it must be better to act than to hold
> still.
>
> The goal of this propossed project:

[...]

> Not as a system just used at school or work, but that is easy enough to
> use at home. No, not easy enough. easier! For every user!

Easy usually means not flexible enough. I think that what is needed is more documentation that is on an appropriate level, easy to find.

I say (as many others) that it should be easy to USE with LITTLE
TRAINING (as with MS Windows).

To administrate it should be easy with AUTOMATIC INSTALLATION and SETUP
after SOME TRAINING (as with Debian).

You should probably use FAI (and/or cfengine) to set up lots of
machines easily.
What's needed though, is better documentation (with example
installations) on setting up Debian and FAI for this.

> I know that this will take an huge effort, if we accept this project. It
> will have all sorts of problems that i couldn't even dream of. But NOW
> it is possible...

I have installed Debian (with GNOME) for two of my friends, and they
haven't even used MS Windows before. They can't administrate them.
But on the other hand, they can't make a mess out of the machines
either. I'll just visit them at least every one-two months, and get a
good meal for that.

I get a meal for free, and they get an stable OS. It's a win-win
situation.

/Jackson

[Stig Brautaset]

* Anders Jackson <anders....@minpost.nu> spake thus:

> I have installed Debian (with GNOME) for two of my friends, and they
> haven't even used MS Windows before. They can't administrate them.
> But on the other hand, they can't make a mess out of the machines
> either. I'll just visit them at least every one-two months, and get a
> good meal for that.

Hardly that good friends if you have to make yourself go on a visit
every two months ;)

Regards,
Stig

--
brautaset.org
Registered Linux User 107343

[Scott Dier]

* Cesar Mendoza <men...@debian.org> [011024 11:17]:

> As an example, I put together a sid box for my wife with KDE and I explained to
> her what were the equivalents in Linux to the applications the she used

I do the same thing for my fiancee :) Shes got a nifty dual-head setup
where she uses mozilla (browsing, mail), x-chat (irc), gnomeicu (icq)
(perhaps soon move her to jabber.ringworld.org/gabber), xsane, jpilot,
openoffice, and gimp. Printing is managed with cups/gimp-print.

> What you want to do it's more easy for an organization that can afford

Yeah. I work for a CSci department and manage ~200+ linux machines.
Stability and planned upgrades are required for me to keep any sanity.
Right now I'm just trying to dig ourselves out of redhat, and then I'm
going to work on how to manage upgrades and inventory machine software
and setup so I can figure out how far some machines are from the 'norm'
install. (and setup some programs to fix machines missing some dpkg
selections)

I'm also using cfengine to manage permissions and configuration files.
Horribly complex, but fun to beable to push out changes that dont
clobber the single-case machines.

--
Scott Dier <die...@ringworld.org> <sd...@debian.org>
http://www.ringworld.org/ #lin...@irc.openprojects.net

Just say NO to Product Activation!