Date: Mon, 08 Jul 2013 17:30:59 +0000 From: "Chad J. Milios" <freebsd-list@nuos.org> To: "Jordan K. Hubbard" <jordan.hubbard@gmail.com> Cc: =?ISO-8859-1?Q?=22C=2E_Bergstr=F6m=22?= <cbergstrom@pathscale.com>, freebsd-chat@freebsd.org Subject: Re: Announcing: nuOS 0.0.9.1b1 - a whole NEW FreeBSD distro, NOT a fork Message-ID: <51DAF753.5050203@nuos.org> In-Reply-To: <8045C2AE-1A20-44AD-A5D8-15E879FED6FE@turbofuzz.com> References: <51D9E499.103@nuos.org> <51D9E641.5020905@pathscale.com> <51D9F45E.2050000@nuos.org> <8045C2AE-1A20-44AD-A5D8-15E879FED6FE@turbofuzz.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 07/08/13 00:15, Jordan K. Hubbard wrote: > On Jul 7, 2013, at 4:06 PM, "Chad J. Milios" <freebsd-list@nuos.org > <mailto:freebsd-list@nuos.org>> wrote: > >> This doesn't provide anything to the core OS that can't already be >> done, albeit with many more keystrokes and the peril of possible >> confusion and misconfiguration. The main thing here is a >> collaboration of what we consider best practices and consolidating >> the more useful configurations into consistent recipes with useful >> simplification of parameters. We don't mean to add yet another layer >> in the name of simplicity that obscures or hides the real nuts and >> bolt beneath and limits your options. >> >> We want to make things more flexible and easier at the same time by >> using the sanctioned FreeBSD ways of doing things, simply allowing >> the ones with most merit to rise to the top, hopefully through >> community involvement. We've had a lot of success using this in our >> production deployments and hope that we don't have to be the only >> ones to maintain it forever. It is an open offer of contribution to >> The FreeBSD Project but it probably doesn't exactly belong there yet. >> It's a layer above, so to speak, and we think we have a place in the >> community working side by side. > > [ As requested - removing freebsd-hackers and retaining only > freebsd-chat ] > > I, for one, am happy to see folks willing to strike out in new > directions with the FreeBSD code base. After 20 years of fairly > gradual evolution, I think it's fair to say that any major conceptual > leaps are probably going to happen outside the project, and not > necessarily eternally but just until they've proven themselves. > > That said, I see some flaws with the project as currently constituted: > > 1. Too much buzz-wordy mission statement, too little emphasis on > technical goals and/or specific points of differentiation. The > nuos.org <http://nuos.org>; web site is a veritable wall of (green!) > text that is so verbose as to be unreadable. If there's a purpose to > the project, it's so obscured by high-concept statements that it's > essentially opaque. If you can't reduce both the mission statement > and the key points of technical differentiation of your project to 6 > one-sentence bullets or less, you're doing something wrong! > > 2. You're trying to have your cake and eat it too when you create a > new project with a new name then say things like "it's not a fork" and > "we're going with the sanctioned FreeBSD way of doing things". If > you just wanted to create a "distro", you could have done it in far > less heavy-weight fashion with some special build tools that could be > run against a FreeBSD source tree to spit out a custom installation > image, just many other BSD variants currently doing (to good effect). > FreeNAS and pfSense are great examples of where a distinct brand was > necessary. nuOS is not, at least not yet. > > Call it what it is: A fork. That doesn't mean it has to be a fork in > perpetuity, but that's what it is now. Furthermore, you're not going > to attract many people by being just a couple of standard deviations > away from FreeBSD. If you're going to create any compelling reason to > run "nuOS" at all, it has to be more ambitious. Just shuffling > everything into ZFS filesystems by default has been done - check out a > PC-BSD distribution sometime. The whole configuration / startup > management picture that you claim to want to preserve from FreeBSD, on > the other hand, is really showing its age. > > 3. You've put the cart before the horse in emphasizing donations and > soliciting bitcoin to such a strong degree at the very outset of the > project, before it's proven its value to anyone. That runs the real > risk of a lot of folks in the community dismissing you as "just the > latest in a long line of wanna-be profiteers." First you attract a > user community THEN you look for donations to keep development going, > if and as necessary. That's not just a nice-to-have item, it's > pretty crucial to any success the project may have. > > - Jordan > > Thank you Jordan for your response. All useful insights. I'd like to address your enumerated concerns, which im sure are shared by many, in a different order if I may: 2: We don't want to fork FreeBSD. There's seriously nothing we would change about the software or the project. We understand why our chosen defaults are not the way everyone lays out a disk by default and why our settings arent already in /etc/defaults/rc.conf. Maybe someday some of them will be but we know we are better off releasing something separate for now rather than trying to convince everyone to see it our way. We target a certain subset of the users which i think will find our distro makes their life easier. 2 still: we are NOT creating a new OS. nuOS is not a new computer operating system. The message behind our brand is that _FreeBSD_ IS the "NEW" computer operating system (and nuOS as a software package and a project are about a new way of operating business, building on stable free technology and avoiding proprietaryizm). I obviously failed to convey that and it was easily lost on my fellow die-hards who've been on board for most of its 20 years and couldnt consider such a robust stable platform to be new. FreeBSD is "new" because it's time has come in the world just as it's coming of age. Most people still don't know what it is or how it's going to benefit them in their business and personal future. Yes, I know it's always been here. I've been a devout fan since 4.2.0. 2. In my professional experience, many many of CCSys's customers are "forking" FreeBSD in proprietary ways to embed it in their products. nuOS doesn't want to be CCSys's proprietary fork. It is and will be maintained as a kind of "reference implementation" for a distro/minimal-fork, kept right there at the armpit close to the trunk of the project where it can keep up or even stay ahead of changes that would put less-organized forks in maintenance hell. I've seen it. I support poor saps who built businesses on FreeBSD 5 or 6 or 7 and their profiteering wasn't sustainable because they didn't give the technology maintenance and the technical debt the were creating beside their assets the due respect at the time. Always driven by that bottom line and pushed by stakeholders and bean counters. In their defense, they didnt have the options they do now to "stay close" to the canonical source. FreeBSD wasn't 20 years mature 10 years ago, obviously. Is FreeBSD 10 going to break or obsolete much of nuOS's "way" of doing things? Probably. I hope so, because that's progress. I think you'll see when our project gets a little further along that we will be pretty well poised to take advantage of 9.2's and 10's and 11's and 12's progress. 2. nuOS is a minimum deviation of a distro. is the culmination of years of supporting them through the trials and tribulations of maintaining their "secret sauce" upon FreeBSD. Yes, i've seen most of these "special build tools" you mention to be "lightweight". I decided to go with /bin/sh and /usr/bin/make instead. :) We didn't find any build tools that reached our ends that actually brought more in power than they cost us in flexibility. 2. To make the naysayers happy, last night I committed the -B argument to nu_install to easily omit the branding changes and motd spam. I only added that rebranding to /boot very recently for nuOS, like i said, as an example, a reference implementation of rebranding because its something many who are making products around FreeBSD want to know how to do easily. For a decade at work i'd almost always just set beastie_disable="YES" on my customers' stuff (while setting it to the color beastie version on my own boxes) after the first time i had to explain i'm not a devil worshiper. (yes i know the loader_logo has changed more than once since then to be more corporate friendly) 3. We arent emphasizing or soliciting any donations at this point. We believe we have a healthy enough sponsorship to at least ensure the continuance of the project. That footer of association keys is actually automatically generated by some other tools for rolling out small business infrastructures, not just technical, but organizational, accounting and legal. (associations have odd numbers of root keys and companies have an even number, that functionality will be explained much later) In our file named PLANNING there are a plethora of addresses people can donate to, put their money where their mouth is so to speak, regarding features that they/their business would benefit from. Its a way of voting that becomes public record in the peer to peer blockchain. Topics of debate get an address automatically even if the business/association votes and funds in a more traditional way and I assure everyone those TODO items are fully backed already and the project will progress without any "panhandling". That being said, :) a project of our currently small size and purportedly grand scope can benefit greatly from even small amounts of help from interested community members and we think that we offer enough benefit already to be appreciated by some who give our software a look. Donations give Scott and I that much more equipment and hours of free reign to act in the interests of nuos.org because ccsys.com has different interests. Think of ccsys like our boring day job and nuos is us after hours doing it how we'd rather being doing it. 1. Buzzword-laden nebulous project with mystery and doubt abound. :) We aren't ready or even interested in explaining too much of our vision or goals yet. I hope they will unfold in due time and are becoming apparent to those reading and trying our code. The ugly site is currently almost "meant" to be uninviting because we are at such an early stage for nerds and hackers only, even though our goals state "ease of use", that will come eventually. Sometimes you have to take a step back to take a leap forward. (and currently we're about ease of developers' use anyway). The site's an homage to old green-screen terminals. For now i'm going to shut up, talk less and let the code speak for itself. I just wanted to address the fork thing: See, we use mercurial and git which is from a different era and has a different way of think, while a project the size and maturity of FreeBSD was born out of a more RCS/CVS/Subversion way of think. Please everyone don't worry so much that i'm "forking" your precious BSD "away" from the community. :) We're not. I had to know and expect at least some backlash. I TOO BELIEVE to "FORK" a project is the absolute last resort, its like civil war. We arent forking like a big branch, think of us more like a vine winding around the trunk. 1. the nuOS project will slowly reveal a scope far wider than just tweaking FreeBSD. FreeBSD is a means to a different end. We have no interest in really affecting FreeBSD in one direction or the other because we think it is on a great course. Further, forward and upward is the only direction we seek to push it. We seek to continue to utilize it, as it is and as it will be, as it evolves organically. We maintain our own build and provisioning tools because our higher level components require the richer features that have only recently become nailed down in FreeBSD, and we wanted to use them in a systematic and programmatic way. We released them so that our work wouldn't have to remain proprietary, that is it. We hate all things proprietary but we don't think GNU/GPL is the right way to "enforce" cooperation. We think BSD promotes smart business to share for the right reasons. Proprietary-land is sometimes where innovation occurs and without the freedom to run our business the way we have for 15 years these "improvements" (or alternatives, to be more accurate) never would have seen the light of day at least not from us. Everyone is made stronger with more options and more flexibility. I think as we make more of The [FreeBSD] Foundation's infrastructure redundant, it will be for everyone's security. For now, their infrastructure as a project is completely trusted by us and project redundancy is the least of our concerns/goals. We hope that the FreeBSD community will accept us for what we are rather than reject us for what we aren't. :) Thank you for reading and sharing your ideas.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51DAF753.5050203>