Re: SEUL: LaetOS: Some thoughts about Distributions, Users and Goalsof this project

[Donovan Rebbechi <elflord@pegasus.rutgers.edu>]

On Fri, 9 Apr 1999, Stefan Haller wrote:

> Do we make the end user and the linux community a favor when
> we create just another entirely new distribution?

It depends on whether the distribution is "entirely" new or not.
Independence is not entirely new. LaetOS ( if and when its available )
will not be "entirely new" either.

> I would say: no, its confusing to the users.

I would say "yes". By creating an enhanced Redhat, we ( and Mandrake
) offer users a "software suite" for Redhat. If our additions really do
make Redhat better, Redhat can put our stuff back in their distro. So we
also help Redhat improve.

> Additional arguments why *not* to create a new distribution:
> - The end-user is a guy who buys his software just out of the

so ? if Redhat improves, end users can buy a better Redhat. 

> - Some guys were talking about standards and setting them.
>   -> How can we set standards?

the standards are already set. 

>     -> we must get a big market share to be important

not true. For example, Independence has almost no market share, but I have
seen some software I built show up in Mandrake. Since we do not put
restrictions on our distro, any distribution developer can give the fruits
of our labour to their user base.

>          less between the big distributions. SuSE in Europe,
>          RedHat in America, Debian in the open-source-community

so unless you have something that is 100% compatible with a big
distribution, you will probably fade into irrelevance. That's why we used
a big distribution.

> - A new distribution needs new testers, new package maintainers
>   etc.
>   -> Does it make sense to do the same as all other distributions
>      over and over again? That needs much time, but it doesn't

no, it does not.  A more sensible approach is to add on to an existing
framework. There are only three "frameworks": Debian, Redhat and
Slackware. The idea is to build on work done by others. 

> What I think we should do:
> We should interact with *all* or at least most of the big
> distributions, even the commercial ones. Only this way we

This is almost impossible, since Redhat/Slackware/Debian all have some key
differences. A more reasonable approach would be to main compatibility
with one of those three.

> We should develop tools that simplify installation and
> configuration for end users. But it makes no sense if only

IMO its better to let the distros take care of this unless you can attract
some very good programmers to your project. Redhat and SUSE have the
manpower and resources to get some heavy programming jobs done.
Independence does not. SEUL doesn't either.

> This would mean only, that we concentrate our resources on
> a topic that other guys are already working on; testing and
> packaging etc.

This is a fallacy, because you have assumed that we will test and package
software that someone else has already tested and packaged. This is not
true. Our efforts are devoted to testing and packaging software that
doesn't come with Redhat. We have taken care to avoid duplication

cheers,

-- Donovan Rebbechi
( project Independence )