![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
This is a list of common myths and misconceptions concerning operating systems that are regularly used when discussing performance, stability, and security. This list is not a spur-of-the-moment document, having had input from numerous sources here on the [H]. Also, it should be noted that the opinions of all of those who have influenced or contributed directly to the list don't always agree, but we all agree on one important principle: the best approach to understanding operating systems is to know fact from myth.
Without further ado, OS Myths v.1
My name appears in yellow in some of those only because I am paraphrasing things these other guys have said in many different ways over time. I appreciate these guys sharing their knowledge on the forum consistently and freely as much as they have. They are an asset to the [H] for sure.
This is a first revision of the list, and as time progresses, there will be some documentation linked as well. This is revision 1 of the list, so expect changes as time progresses.
Without further ado, OS Myths v.1
- Suggesting Win98 for performanceWindows 98, on any equipment that falls within system specs for XP (and arguably lower), does not perform better than Windows 2K/XP. In fact, tests have shown 98 performing slower. Add to that the known memory holes and similar large memory issues with the 9x series, and it is immediately ruled out as an option in today's higher CPU-intensive and high-memory-capacity machines. Since the main argument for suggesting 98 is because of its "lighter" reputation, XP Home is a better solution than 98 because, unlike the 2K and XP Pro versions, it comes standard with less underlying processes running and has none of the domain-level pieces that 2K/XP Pro have. The lowest common denominator for low-profile hardware performance is the not a 9x, but an NT... Win XP Home.
Credits to: posts by SJConsultant, Ranma_Sao, & GreNME for information contained above
. - Disabling services for performancethere exists no proof that disabling services (a la Quack Viper) or removing the page file improves performance. What disabling is known to do is shut off and keep off specific underlying processes in Windows, which may or may not cause detrimental performance (YMMV). There are certain services, like Messenger and Remote Registry, that may help avoid annoyances when set to manual. Understanding how and why services do what they do is a great way to get "under the hood" of an OS (especially Windows), but don't jump on premature "answers" too quickly to solve your performance hopes. As a recent example: disabling many of the services will cause immediate problems during and/or after the install of Service Pack 2 for XP.
Credits to: Phoenix86, Ranma_Sao, SJConsultant, & GreNME for information contained above
. - Disabling the page file on systems with large amounts of ram for performancethis might have been true, at one time. The windows 9x series, for example, was verified having some serious page file issues, so there may be no problem believing this would have worked. However, with 2k/xp, this is no longer the case. There is really only one tweak that will make any significant difference: Put the page file on another drive, preferrably on a different controller. Also, it's worth noting, as of a couple years ago, fixing the page file size to something static would prevent windows from dynamically increasing the swap size in the middle of, say, a game, which would sometimes cause a stutter. This was under win2k, and it was on old hardware, so we don't think it's really relevant anymore. There are some strong arguments by those who contributed for running without a page file (like a realistic theory about reducing the number of page faults), but more speed is not one of them.
Credits to: XOR != OR, Phoenix86, Ranma_Sao, and GreNME for information contained above
. - Operating system A is more secure than BTraditionally speaking, no operating system is inherently more secure than any other; it all comes down to administration. Some say this regarding linux, some say this regarding *bsd, and yes, some will even say this regarding windows ( those are typically brave souls ), and at the end of the day, all it's really saying is what that person is most comfortable administrating. Every operating system has its strengths and its faults, many of which are shared between different operating systems, and no one has significantly more of one or less of the other. The biggest variable in the matter of securing a machine will always be the human setting it up. Addition for clarity: Security can mean different things depending on what is to be secured from what, and the security of two systems can't be compared unless they also see the same use. Since different OSes will often be put to different use, and differences in market share will skew the results, it's hard to do direct comparisation. For different uses, different systems will probably be better or worse. However, the deciding factor is not the software, but how well it is set up and maintained, which depends solely on the person in charge. Thus, the most secure system for any one administrator is almost always the one he knows best. This is not an attempt to remove the viability of any OS.
Credits to: XOR != OR, SJConsultant, Phoenix86, Ranma_Sao, HHunt, & GreNME for information contained above
. - Linux is the answer to having an older machine perform like newLinux is a great alternative to running Windows, and has just as many overall capabilities, but the most often-used misconception is that installing a modern Linux will perform like new computer on an old machine. If one were to run without a GUI, Linux performs wonderfully, even on older hardware. However, "even on" is the catch. It will always run better on newer hardware, just like Windows and any other OS, because the better the machine, the better the software will run (9 times out of 10). And if you decide you want to run a GUI, then the requirements go up. GNOME and KDE, the two most popular graphic environments, require pretty much as much in terms of system requirements as WindowsXP. If you want that old machine to work its best as a file server or web server or just to try out UNIX tricks, then you'll get better results leaving the graphic environment off, meaning you won't be getting everything you may want out of the deal. When running without a GUI, Linux flies, and can even fly on older hardware. Once a graphical environment is added to the mix, the overhead increases dramatically.
Credits to: XOR != OR, Phoenix86, & GreNME for information contained above
. - Blaming Microsoft for all problemsNothing is more detrimental to troubleshooting than seeing people blame everything on Microsoft anytime they have a problem with their PC. While Microsoft does not need defending (Service Pack 2 seems to be doing very well), some need to realize that poorly written drivers and applications are often likely to blame. Microsoft can try as hard as they want to make a stable OS, but if loading some dubious sound card or other component drivers cause lockups and freezes, you can't always blame MS. This is where digitally signed drivers have come into play, as well as WHQL certification. In addition, there are recommended techniques for optimization, security, setup, and configuration. Equally, there are non-recommended techniques for many of these, as well. Often, the best way to understand what can be done in various situations with software is to understand how the things are done. Be wary of any modification, tweak, 'fix,' or other walk-throughs that avoid this or cannot explain how it is changing things. This is not meant to defend any MS business practice or legal issues, it is meant to point out that using MS as a scapegoat for our troubles is not solving any problem, and could be allowing the problem to get worse.
Credits to: djnes & GreNME for information contained above.
My name appears in yellow in some of those only because I am paraphrasing things these other guys have said in many different ways over time. I appreciate these guys sharing their knowledge on the forum consistently and freely as much as they have. They are an asset to the [H] for sure.
This is a first revision of the list, and as time progresses, there will be some documentation linked as well. This is revision 1 of the list, so expect changes as time progresses.
That's the first thing I thought of when I read that "myth". Ever read the