Subject: Re: "external software" (was Re: xntpd)
To: <>
From: Herb Peyerl <hpeyerl@beer.org>
List: current-users
Date: 01/02/1996 08:19:57
Peter Galbavy <peter@wonderland.org> wrote:
> Do you mean that as soon as all the work has been put into making
> the port maintainers of an excellent piece of software aware of
> the changes required that we should then ignore it as part of a
> distribution ? I would like to see *more* packages in the *standard*
> distribution to make it easier to install a complete system out of the
> box.
Lots of people have differing opinions on what should be included
and not included... Lots of people have opinions on what a release
should be and shouldn't be...
My opinion is that we should not only allow but encourage people to
make their own releases... If this was done properly, we would suddenly
see all sorts of technology that could eventually be integrated back
into the tree... Imagine if someone was to make a "NetBSD/router"
CDROM distribution containing things like HDLC/Frame-Relay/FDDI/ATM
drivers for a few architectures and a bunch of setups and configuration
scripts? (I hear there's an ISP across the water that has all of this
code for -current but that it's not in the tree)... But then that same
technology wouldn't be interesting to someone who wanted to use NetBSD
as an X-terminal, but someone else could make a "NetBSD/workstation"
with its own installation... etc...
Anyway, those are just rambling thoughts. I sent something a little
more organized to core O(1 month) ago and haven't heard a response. Since
then though, I've had a core member tell me that NetBSD isn't interested
in marketing and that the project is the code so now I'm totally
confused... (Yes, I admit this was before Jason and Christos joined
the core but I assume that they were presented with all outstanding
issues when they were brought in)...
> I also throw my nomination to the floor for the inclusion of Darren
> Reeds' excellent ip filter stuff. OK, the thing is a loadable
> module. Great. Now let me just patch each kernel source tree with
> the #ifdef surrounded changes... Another example of a member or
> core saying "I don't think this solution is elegant enough" without
> coming up with a timely alternative. I remember Darren offering
> the filter code last *SUMMER* guys. Where is the "official/pure/perfect"
> version from cofe ? Hello ? Anyone there ?
NetBSD/Firewall! Imagine this. Darren, or someone else incorporates
that technology into their own release, with INN, sendmail-configs,
security-scans, etc, and dumps it on a CD. After a while, the code
begins to take on a different form as more people use it on more
platforms and suddenly the code transforms into something that is
capable of going into the tree...
Anyway, if I go any further I'm going to start flaming which of course,
is not compatible with any of these lists..