by homeworld.cygnus.com with SMTP; 27 Nov 1998 08:39:03 -0000
	by allegro.lemis.com (8.9.1/8.9.0) with ESMTP id TAA25377;
	Fri, 27 Nov 1998 19:08:57 +1030 (CST)
	by freebie.lemis.com (8.9.1/8.9.0) id TAA00488;
	Fri, 27 Nov 1998 19:08:55 +1030 (CST)
Message-ID: <19981127190855.A468@freebie.lemis.com>
Date: Fri, 27 Nov 1998 19:08:55 +1030
From: Greg Lehey <grog@lemis.com>
To: Adrian Filipi-Martin <adrian@ubergeeks.com>
Cc: netbsd-advocacy@NetBSD.ORG,
        FreeBSD advocacy list <FreeBSD-advocacy@FreeBSD.ORG>,
        advocacy@openbsd.org
Subject: Re: Merging Net/Free/Open-BSD together against Linux
References: <19981127162648.R682@freebie.lemis.com> <Pine.BSF.3.96.981127030245.27809B-100000@lorax.ubergeeks.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <Pine.BSF.3.96.981127030245.27809B-100000@lorax.ubergeeks.com>; from ADRIAN Filipi-Martin on Fri, Nov 27, 1998 at 03:20:41AM -0500
WWW-Home-Page: http://www.lemis.com/~grog
Organization: LEMIS, PO Box 460, Echunga SA 5153, Australia
Phone: +61-8-8388-8286
Fax: +61-8-8388-8725
Mobile: +61-41-739-7062

On Friday, 27 November 1998 at  3:20:41 -0500, ADRIAN Filipi-Martin wrote:
> On Fri, 27 Nov 1998, Greg Lehey wrote:
>
>> On Friday, 27 November 1998 at  0:49:30 -0500, ADRIAN Filipi-Martin wrote:
>>> On Fri, 27 Nov 1998, ADRIAN Filipi-Martin wrote:
>>>
>>>> 	So, again, who would like to participate on such a project,
>>>> scanctioned by a *BSD core team, or not?  The michanics of the process are
>>>> fairly straight forward, but they are time intensive.
>>>
>>> 	Just an addendum to my previous message.  If you think you would
>>> be interested in helping on such a task, send me your address off-line.
>>> I'll save them.
>>>
>>> 	If there is suficcient interest/manpower to make it more than a
>>> one-man show, I'll set up a 3-way CVS mirror at UVa or maybe a local ISP.
>>> We can tag an initial starting point and start merging into one of the
>>> three trees.  If this bears fruit we can then re-merge any recent changes
>>> and make it a new baseline for userland.  (Yes, there is undoubtedly a lot
>>> more to consider, but it's a start.)
>>>
>>> 	I think minimally, there would need to be two people from each
>>> group.  I am best counted as a FreeBSD'er.  Are there five others?
>>
>> Count me out.  I don't think this is a worthwhile effort.  Discuss
>> things, maintain more communication, try to keep things pointing in
>> the direction, sure.  But your efforts aren't going to give us a
>> unified userland: they're more likely to create a fourth version.
>
> 	Well, that would be hard to do without a kernel.  ;-)
>
> 	Avoiding a new *BSD is one big reason why I want to constrain such
> an effort to non-kernel code.

But of course, if you do it and nobody wants it, you might be tempted
to put in a kernel as well.  After all, by definition any kernel will
do :-)

> 	What do you think would increase the liklihood for such an effort
> to succeed?
>
> 	It's not that I think such work should be done in secrecy without
> any comminication with the developers at large.  I personally would want
> to work in a faily autonomous manner so as to not be directly branched off
> of a particular CVS projects repository.  But that could just be me.
>
> 	The basic reason I'm pursuing the notion of userland unification
> is that I'd bet dollars to doughnuts that the egos are smaller and less
> likely to be a problem outside of the kernel.  It would also leave the
> respective camps free to have their own add-ons.  This would be one way to
> reduce the effort spend tracking what the other groups are doing for the
> entire distribution.

You've forgotten something that went by a day or so ago: the source
trees are structured differently, and the licenses aren't quite the
same.  In these areas you'll run into an amount of stubbornness^W
reluctance to change which might surprise you.

> 	I could see things where 90% of userland, and 90% of the ports
> (not packages) could be lumped together on a single CD, that could be
> included in each OS's distribution.  The particular flavor would provide
> it's kernel sources, system binaries and other bits that are truly kernel
> specific.

Well, since you mention the ports, there's an idea.  I know that
FreeBSD and NetBSD have a certain amount of object code compatibility;
I expect that applies to OpenBSD as well.  A thing that *really* would
be worth doing would be smoothing the differences, which would
probably require some modifications on all three systems.  The result,
though, would be that the ports (which Walnut Creek already ships
precompiled) would work on any of the three platforms.  And if you
prefer the NetBSD dump(8) over the FreeBSD version, there'd be nothing
to stop you.

Greg
--
See complete headers for address, home page and phone numbers
finger grog@lemis.com for PGP public key