From owner-freebsd-stable@FreeBSD.ORG Thu Sep 4 00:18:57 2008 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B1D2E1067264 for ; Thu, 4 Sep 2008 00:18:57 +0000 (UTC) (envelope-from danallen46@airwired.net) Received: from mail.utahbroadband.com (mail.utahbroadband.com [204.14.20.91]) by mx1.freebsd.org (Postfix) with ESMTP id 91D418FC12 for ; Thu, 4 Sep 2008 00:18:57 +0000 (UTC) (envelope-from danallen46@airwired.net) Received: (qmail 6425 invoked by uid 89); 3 Sep 2008 23:45:32 -0000 Received: from unknown (HELO ?192.168.0.16?) (danallen46@airwired.net@66.29.174.6) by 0 with ESMTPA; 3 Sep 2008 23:45:32 -0000 Message-Id: From: Dan Allen To: Scott Long In-Reply-To: <48BF1F70.4020509@samsco.org> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v926) Date: Wed, 3 Sep 2008 18:18:55 -0600 References: <35445338-D597-4FE2-996F-DEC7BE986741@airwired.net> <48BEEB55.4050406@madpilot.net> <48BF1F70.4020509@samsco.org> X-Mailer: Apple Mail (2.926) Cc: freebsd-stable@freebsd.org, Guido Falsi Subject: Re: FreeBSD 7.1 Content X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 04 Sep 2008 00:18:57 -0000 On 3 Sep 2008, at 5:36 PM, Scott Long wrote: > What's wrong with "downstream"? I can crash most Linux distributions in an hour. Your examples are just why I like FreeBSD and why I do not like or normally use Linux. I only grabbed Ubuntu recently because there is ZERO net access from FreeBSD on my new Dell laptop. Linux to me is too random. FreeBSD has always appealed to me because it has a higher quality bar, more structure to it (like /usr/src/ and its sweet buildworld), and it does not include every possible gadget in it. I know that my proposing the addition of Firefox to disc1 may be seen as "every possible gadget" but having a web browser is pretty important these days. > Having derivative releases like FreeSBIE and PCBSD and others is an > excellent way to make the release process scalable and able to meet > the > wants and needs to different users, yourself included. In fact, I > think > it's an utter waste of time for the FreeBSD release team to worry > about > packages on disc1 and whatnot. That needs to be done by teams who can > focus on doing that task and doing it well. The FreeBSD releases need > to become bare-bones references for others to build on an repackage > and > grow and improve. That's already started, but the efforts of those > teams needs to be highlighted and given more, dare I say it, respect. > They are the future that will bring FreeBSD to a wider audience. They > need to be treated as first-class developers and members of the > FreeBSD > family; the "official" freebsd.org releases need to relegated to being > just bare-bones bits that are there for others to bring to the masses. You have good arguments here. You state -- very correctly -- that derivative release teams need to be "treated as first-class developers and members of the FreeBSD family". But are they treated so? Can larger audiences of developers be entertained while maintaining FreeBSD's stability? FreeBSD has as one of its great strengths a small set of developers and a release process that seems to deliver a more reliable product than Linux, at least in my experience. The fewer people that mess with the bits, the more stability delivered. Obviously the other side of the coin is that if not enough people mess with the bits then not enough features and hardware support will exist and the product will be irrelevant. Finding the sweet spot is hard. If derivative release teams are modeled after FreeBSD core -- good checkin structure, a few solid contributors rather than teaming hordes of inexperienced programmers -- then perhaps that is the way to go. If they do not have the structure, process, and experience, then it should be done by the mainline team. For me, as long as ANY packages are shipped on disc1, then I think they should be the right ones, and my hunch is that there should be just a few packages and their dependencies: rsync, perl and firefox. Firefox of course will drag in a bunch of stuff (X.org,atk,gtk...). (Actually rsync should become part of the main distro. It is so incredibly useful, but that is just one man's opinion.) This is a good dialog to have. I do not know if this is the right list for it, and I certainly do not want to mess up 7.1. Perhaps none of this can be ironed out in time for 7.1 and it will have to wait for 7.2. That is fine. Whose job is it to decide on the packages for a release? I have spent a non-zero amount of time looking for specifications or design plans for releases and have been unable to find them. Thanks to everyone today for your comments! Dan