Subject: Re: Final plug for extra utilities
To: Eric Radman <>
From: Andy Ruhl <>
List: netbsd-users
Date: 09/01/2006 21:19:20
On 9/1/06, Eric Radman <> wrote:
> On 12:12 Fri 01 Sep     , David Brownlee wrote:
> > >
> > >Is there any chance of a consensus on this idea: adding another set to
> > >NetBSD's base distribution called plus.tgz or util.tgz.
> > >
> >       I think a better approach would be to help integrate pkgsrc
> >       installation into sysinst, so when people install NetBSD they
> >       can select those, or any other available binary packages.
> I think I left this too open-ended.  Allow me to try to clarify what my
> goal is for this suggestion:
> If a binary package is available, pkg_add is very easy, and there really
> isn't any need to make the installer any more complicated than it is
> already. The problem here is that binaries are often not available.
> This happens when installing a snapshot, or on older/slower systems that
> do not yet (ever?) have binaries posted.

Maybe the right thing to do is figure out how binary tools get added
to the base system and why.

You're making good arguments, but there seems to be an overall feeling
of caution when adding something to the base system. This is the
feeling I get, anyway. Whether this is useful or not, I don't know,
but it seems to be the rule.

I think you make a very good argument about some very basic tools that
lots of people use that are not available in binary form for some of
the archs.

I'm not a developer, but I have some experience with adding things to
an already well established system, so here's my take:

Things that get added must meet some or all of these critiera:

o They must be critical to operation
o They must provide useful function to a very large number of users
o They must require less effort to maintain than the benefit they provide
o They must not need constant updating to remain relevant
o There must be some consensus that some particular program is the
correct one among similar programs.
o There must be some kind of agreement on why the function some
particular program is better or more relevant than some other (these
days choosing some program will almost inevitably spark some religious

And probably lots of other stuff.

My own feeling is that sudo probably meets these criteria.

If enough people need to get on the internet through ppp, then that
should be included since there's nothing worse than being stuck
without the network due to lack of tools. Is userppp really the right
tool? I don't use ppp, haven't for many, many years now. I'm ignorant
to this.

A rudimentary text based browser might be useful, but I think there
will be too many opinions on what the right one should be. And there
have been some security concerns with some of them if I remember

Shells should be left alone. That's too big a battle. I personally
prefer tcsh, but I think that's a minority opinion compared to bash.

Perl, forget it. You'll never please more than one person with
whatever decision is made, or something.

After thinking about this, it seems like it would be useful to provide
a set of tools and/or documentation that covers ALL (or as many as
possible) bases related to getting connected to the internet. This may
spark some controversey about what drivers and options should be
included in the kernel, however...

This isn't easy is it?