Date: Sun, 13 Feb 2005 09:40:49 +0100 From: Anthony Atkielski <atkielski.anthony@wanadoo.fr> To: freebsd-questions@freebsd.org Subject: Re: Instead of freebsd.com, why not... Message-ID: <1153299714.20050213094049@wanadoo.fr> In-Reply-To: <69120B54-7D98-11D9-B134-000D933E3CEC@shire.net> References: <20050212203851.D694116A4D3@hub.freebsd.org> <1108249638.32574.49.camel@zappa.Chelsea-Ct.Org> <863830644.20050213025358@wanadoo.fr> <BE475D84-7D76-11D9-B134-000D933E3CEC@shire.net> <1095160797.20050213085726@wanadoo.fr> <69120B54-7D98-11D9-B134-000D933E3CEC@shire.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Chad Leigh -- Shire.Net LLC writes: > ????? What the heck does this mean? It means that large organizations want to have a single official release of the OS running on all servers, and they expect that release to come formally from a specific official source. Hacking changes into the code and then installing that in production is not acceptable. Often every change to the OS must go through a test and rollout process that can take months at some companies. Emergency patches must be tested in advance by the vendor, and the vendor must stand behind them. > I would bet that most larger installations of Linux or FreeBSD or any > other open source OS would be considered non-standard. Yes. That's why so many companies run Solaris instead. > Heck, I bet YOUR installation of FreeBSD could be considered > non-standard. As the owner of the system, I define what is standard on my site, and I consider FreeBSD 5.3 to be standard. I don't make any modifications of my own to the code, though. > Your statement make absolutely no sense. It does to someone who has worked in this kind of environment for several decades. There are still companies running Windows 3.x because it is so long and difficult to roll out anything new. > If the fix that you decry is a reasonable fix, who says it can't be > rolled back into an "official" release. It can be, but until that is done, many organizations won't touch it. There's another separate issue with source fixes. It's a common misconception that anyone with access to source can just dive into it and fix any problem. In practice, that is never the case. Nobody has all of any OS memorized, and no one person can dive into the code of any OS and come up with fixes to every problem. Even among official developers, typically each developer knows only his own code extremely well, and has only a vague idea of how the rest of the code works. While it is true that you could theoretically fix anything in time with access to source, in practice the time required is so long that it is effectively impossible in many situations ... you _must_ enlist the help of one or more developers familiar with the code segments that have to be fixed. And that in turn means that, in order to provide full support, you must be able to compel the cooperation of developers. Proprietary vendors can do this; open-source organizations cannot. -- Anthony
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1153299714.20050213094049>