Date: Sat, 8 Dec 2001 09:03:57 +0100 From: Brad Knowles <brad.knowles@skynet.be> To: "Anthony Atkielski" <anthony@freebie.atkielski.com>, "Mike Meyer" <mwm@mired.org>, "Brad Knowles" <brad.knowles@skynet.be> Cc: <chat@FreeBSD.ORG> Subject: Re: A breath of fresh air.. Message-ID: <a05101025b83773d980a5@[10.0.1.16]> In-Reply-To: <01b101c17fb7$e45bf8e0$0a00000a@atkielski.com> References: <0112071641320B.01380@stinky.akitanet.co.uk><000b01c17f42$c23ab140$0a0 0000a@atkielski.com><3C110351.4748B559@duth.gr><005001c17f6c$e60c0ef0$ 0a00000a@atkielski.com><15377.17350.796336.801464@guru.mired.org><0069 01c17f70$19a2f820$0a00000a@atkielski.com><15377.18218.830731.410656@gu ru.mired.org><008101c17f9a$1a4a4290$0a00000a@atkielski.com><15377.3661 7.358466.76379@guru.mired.org><00ab01c17f9d$0bde8510$0a00000a@atkielsk i.com> <15377.37214.213789.306335@guru.mired.org> <00c901c17f9f$e80a95e0$0a00000a@atkielski.com> <a05101009b83750a740f1@[10.0.1.16]> <013901c17fac$b23dc6a0$0a00000a@atkielski.com> <a05101017b837621e58c4@[10.0.1.16]> <01b101c17fb7$e45bf8e0$0a00000a@atkielski.com>
next in thread | previous in thread | raw e-mail | index | archive | help
At 8:13 AM +0100 on 2001/12/08, Anthony Atkielski wrote: >> People are irrational, I've said that before. > > I've seen no evidence of that. Of course. The insane people always claim that they're the sane ones -- it's always the rest of the world that is insane. Just because your own actions are the height of irrationality doesn't mean that other people aren't also irrational. > Why would disabling a cooling mechanism make the hardware run faster? Can > you give a specific example? Sure. Certain Intel chips have had a built-in "NOOP" cycle. IIRC, they "rest" for one out of five cycles, which helps them run cooler. It also means that they do not run at full speed. > Does Windows do this? If so, where? Yup. Windows disables the cooling cycle, which makes it run 20% faster on that hardware, but makes the hardware less reliable. This puzzled Linux & FreeBSD hackers for a while, primarily when they were running number-crunching benchmarks and Windows on the same machine would always run faster than the exact same program running under Linux or FreeBSD. This one has been known for quite some time. >> However, this is totally unrelated to the fact >> that a lot of PC hardware is inherently unreliable. > > Then why did you mention it? I didn't. I mentioned that the OS was unreliable, in addition to unrelated unreliability in the software, and also in addition to the unrelated frequent unreliability of the hardware. Each is frequently unreliable in and of itself, independent of the others. However, there are also compound situations, where unreliability in one component causes or exacerbates unreliability in others. > Additionally, how does the inherent reliability (or lack thereof) of PC > hardware have anything to do with software _making_ hardware unreliable, > which was your original assertion? That wasn't the original assertion. Go back and re-read the paragraph. >> Yes, if you take care to specify higher-quality >> components that are known to work together well, >> you can build a very reliable machine. > > Even if you are running Windows? The hardware would be more reliable, unless the OS did things to disable reliability-enhancing features of the hardware that also made it run slower. > But didn't you assert that machines > running Windows were unreliable purely as a consequence of running Windows? Yup. >> Not at all true. > > I can only go by what Mac users say, as I don't use Macs myself these days. And I go by my own years of personal experience on the Mac, and years of personal experience on other hardware and OSes, plus the many multitude of horror stories that I hear from friends and family alike. >> ... there probably isn't a single NT server in the >> world that has been up and running more than two >> hundred days straight. > > I know of many servers that have run for years at a time. Really? Where? What were they doing? What precise version of the OS was installed? My personal experience with NT servers was a while ago, and at that point each machine would have a blue screen at least once a day, sometimes several times a day. And then, of course, there's this bullshit about having to reboot after installing any software or making any configuration changes to the OS. >> Now, Macs don't have that kind of uptime, but >> that's because most people turn their Mac >> off at night when they leave. > > That can also be said of Windows, and it is no more or less true. Not when you count intra-day uptime. By far the vast majority of Windows users that I have spoken to over the years, most end up rebooting their machine many times in a single day. Contrariwise, most Mac users I have known over the years will boot their machine when they come in and then turn it off when they leave, and not have any other shutdowns or reboots throughout. >> Obviously, the surveys you've seen have not >> accounted for all the possible variables >> that are involved. > > That is, they don't show the numbers that you'd prefer to see. I'd prefer > to see different numbers, too, but there they are. Not at all. All surveys have to be interpreted in context, and the limitations of the survey should be made clear up front. If the limitations of the survey are not made clear up front, then not only is the survey flawed, but obviously the people conducting the survey have something to hide. All I ask is that I be given all the details of precisely how the survey was conducted and how the analysis was performed, so that I can then make my own decision about what context to place the survey results. Show me the hard numbers, with all the dirty details, and I will believe them so far as they go. >> In my many years of experience, I would say that >> AOL is very representative of the server market. > > How many years of experience do you have, and how did this relate to AOL's > use of servers, and what leads you to believe that AOL is representative? Well, I've been using both Unix and the Macintosh since 1984, and I've been a Unix system administrator since 1989. I was the first Internet Mail Operations person hired by AOL (at a time when they had four Internet mail gateway machines), and when I left I was the Sr. Internet Mail Systems Administrator and was the chief technical person responsible for over a hundred different machines that we used in our subsystem alone. I also helped architect and install the AOL caching nameserver farm, capable of handling on the order of 32,000-64,000 DNS queries per second -- indeed, this is probably still the world's largest nameserver installation, although a.root-servers.net is getting up there (and can handle considerably more traffic than it sees today), and the GNS machines operated by Nominum may well surpass these kinds of rates once they get fully utilized. Since then, I've done consulting work at a wide variety of clients (some big and some small), and I believe that I have enough experience to make this kind of determination. >> How many decades of experience do you have that >> you'd like to stack up against mine ... > > I do not engage in credentialism, as it is not relevant to a well-reasoned > argument. Well, without knowing your experience and how you came to know the things you do, the only thing we can judge you on is the kinds of things you say on the mailing list. In that context, I'd say that you're a bigoted Windows crackpot who is trolling for arguments. I was at least hoping that you'd be able to help dispel this image by sharing with us some information on your background. > Then my spontaneous admission that information is lacking would seem to bode > well for my honesty and candor. No, not really. It simply shows that you're either trying to revise history, or that you're ignorant enough that you didn't set up the scenario properly before you asked the stupid question in the first place. -- Brad Knowles, <brad.knowles@skynet.be> H4sICIFgXzsCA2RtYS1zaWcAPVHLbsMwDDvXX0H0kkvbfxiwVw8FCmzAzqqj1F4dy7CdBfn7 Kc6wmyGRFEnvvxiWQoCvqI7RSWTcfGXQNqCUAnfIU+AT8OZ/GCNjRVlH0bKpguJkxiITZqes MxwpSucyDJzXxQEUe/ihgXqJXUXwD9ajB6NHonLmNrUSK9nacHQnH097szO74xFXqtlbT3il wMsBz5cnfCR5cEmci0Rj9u/jqBbPeES1I4PeFBXPUIT1XDSOuutFXylzrQvGyboWstCoQZyP dxX4dLx0eauFe1x9puhoi0Ao1omEJo+BZ6XLVNaVpWiKekxN0VK2VMpmAy+Bk7ZV4SO+p1L/ uErNRS/qH2iFU+iNOtbcmVt9N16lfF7tLv9FXNj8AiyNcOi1AQAA To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?a05101025b83773d980a5>