NetBSD-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Definition of NetBSD users

At Tue, 23 Mar 2010 08:26:44 -0500 (CDT), "Jeremy C. Reed" 
<> wrote:
Subject: Re: Definition of NetBSD users
> I do see a few features that Linux systems commonly provide that make it 
> easier, most notably: 1) automated tasks on hardware plugin/removal 
> (dynamic device management from userlevel); 2) more consistently 
> provided binary packages for common platforms and better tested binary 
> package maintenance tools; 3) friendly installer that installs common 
> desktop" software by default; 4) Flash support.
> For NetBSD: 1) code for this has been done a few times and core@ is 
> attempting to make a decision on this; 2) need more official bulk build 
> machines and need some more maintenance to pkgin; 3) a few projects have 
> been done to redesign and rewrite a NetBSD installer (but not complete 
> as far as I know); 4) Convince the world to not use proprietary 
> technologies! (Let's not re-discuss #3 on this thread.)

This is idle rambling with my coffee in hand, but your points and
comments do raise some of my own that I'd like to write a wee bit about:

1) better hot-plug device support for sure! -- that's something which
you see on Mac laptops and desktops as well, at least for those kinds of
devices which are meant to allow hot-swap.  Some careful thought has to
be put to how to differentiate the needs of the "desktop"/personal
machine with the needs of a secure server/cluster environment though
(one cannot assume that physical security alone is enough!)

2) I think some folks will also take "easy and better binary package
support" to mean the ability to upgrade single applications without
having to rip out half the world and upgrade/re-install them all?  I
don't think it's a necessary feature (I would just as well "rm -rf
/usr/pkg" and re-install a whole new consistent set of packages, and it
doesn't even take that long on modern machines which are truly capable
of running full "desktop" environments).  I know others find this far
more important to them, and I know there are solutions, but I think the
problem is hard, and the currently known solutions only half-baked, so I
don't think we can charge right in and say we've solved it.

3) I don't really think we really need to re-design sysinst and start
over from scratch just to be able to make easier installs for common
"desktop" configurations.  Sysinst does need to have better defaults for
modern systems of all types (desktop, laptop, netbook, server, etc.),
and it does need to come in "canned" configurations that will provide
everything in one go, kinda like FreeBSD's installer allows package
collections to be installed.  (sorry, I had to say something!  :-))

4) Indeed, proprietary technologies shouldn't even be thought about in
this context, let alone mentioned as a "notable" feature that makes
GNU/Linux systems "easier".  Seriously.  Flash needs to die, and
hopefully it will sooner than later with the increasing acceptance of
Flash-free handheld devices and of course with HTML-5.  Maybe we can

                                                Greg A. Woods
                                                Planix, Inc.

<>       +1 416 218 0099

Attachment: pgpgY5B7zAZfS.pgp
Description: PGP signature

Home | Main Index | Thread Index | Old Index