Here is just one example of how fast and easy it is to use the Start Menu in Windows 7.
Doesn't seem very related to the Start Menu at all. However, on that, it is RIFE with absolute nonsense and unsubstantiated, cargo-cult 'optimizations' of dubious value as well as some pandering of useless software applications.
-Present the "tip" to make Avast "Faster and more Efficient" by enabling Silent/Gaming Mode, which is not the case, since all this does is disable Avast Popups. Less annoying certainly, but not faster or more efficient.
Then.... installs Glary Utilities.
Uh, yeah, I'll pass on installing pointless crap masquerading as tune-up software, thanks. I'll cover this later By copy pasting some of the discoveries I made in a blog post I tried to write about various "cleaner" utilities.
"For part 8 we're going to disable the Windows Search Indexing Service"
My favourite part is that during the process of disabling this, he actually uses it. (The Start Menu Search uses the the Search Indexing service; if it is disabled then the "Search" won't find anything added after the Indexing service was disabled.). Bravo. Nothing like lobotomizing the most powerful feature of an OS for the unsubstantiatable rumour that it "might improve performance". Doesn't matter if your Word Processor starts 10 ms faster if it takes you three seconds to start it using the All Programs Menu.
Then he disables UAC Well, not really. He sorta does. Doesn't mention why. Doesn't say how this helps performance (Spoiler: it doesn't)
Then he uses the "Adjust for Best Performance" checkbox in the Advanced Visual Settings dialog.
This also, despite it's name, doesn't actually improve performance at all. (XP was around the last time it might improve performance, for older Graphics Cards. My ATI Rage Pro AGP card for example benefited noticably from disabling the Alpha Blended Selection box option, because it's driver/hardware didn't support accelerated Alpha blending so it had to be done in software). Nowadays few of the capabilities shown have a negative effect on performance, and of those that do I doubt there is a measurable one, and of those that there is a measurable one, it's probably not perceptible.
The Virtual Memory one is where he steps into cargo-cult rumour pandering. Why correct the rumour mill when you can use it to grind your flour, I suppose:
"If you have 4GB of Memory or more, Select no Paging File. If you do not, select Custom Size and for the Initial and Maximum size enter the value listed below for "recommended"."
This is not much of a performance tweak. If the paging file is disabled the Heap Manager is basically lobotomized and certain memory operations simply cannot be performed properly. It also lobotomizes your Commit Charge Limit, which can prevent you from running more than 3 applications that perform large Virtual Allocations at startup. (This is because with no Page file there is no "Virtual" memory and any allocation is a direct allocation to physical RAM). I think
Mark Russinovich, who knows more than pretty much anybody else in the world about Windows Internals, put it best:
Some feel having no paging file results in better performance, but in general, having a paging file means Windows can write pages on the modified list (which represent pages that aren’t being accessed actively but have not been saved to disk) out to the paging file, thus making that memory available for more useful purposes (processes or file cache). So while there may be some workloads that perform better with no paging file, in general having one will mean more usable memory being available to the system (never mind that Windows won’t be able to write kernel crash dumps without a paging file sized large enough to hold them).
The "Some workloads" would only be for very special purpose servers that don't need things like Standby memory, or which would work best with no swapping ever. This is limited to only a few very specific Server configurations such as AD Servers. But most Servers of that nature are run on the same system as other services such as databases so the cases are few and far between.
The notation about using the recommended size also seems to ignore that the recommended size listed is simply 1.5 times the usable memory. Windows doesn't perform any sort of sophisticated check on what applications you run frequently and their memory profiles to come up with the recommended value. It's basically just a value that somebody at MS pulled out of thin air, which Mark himself pretty much says in the linked post. It also puts a set limit on your Commit Charge at which point allocations will fail. The only advantage I
could see would be in making it contiguous, but this hasn't been the case since at least XP. (XP and later allocate a permanent pagefile set in this manner using what it believes to be the outermost tracks of the drive. since Magnetic Platter drives have a Constant Angular Velocity this means the Read/Write rate is fastest there, but if there is already data there for whatever reason the 'permanent paging file' will not be contiguous. The paging file normally tries to consume the outermost tracks anyway so this doesn't really change a whole lot. (it's pretty much the same as just having an application allocate a lot of virtual memory to make the pagefile larger). It seems they aren't aware of the simple arithmetic behind the recommended value though.
Another tip recommends ReadyBoost. ReadyBoost is quite thoroughly covered by Mark as well in his
series on the Vista Kernel. I find this advice a bit of a face-*censored*-turn even though it will typically have a positive result given the recommendation to disable the pagefile, though.
Going back to the "registry cleaner" type thing, GlaryUtilities, in this case:
I installed it in a Win 7 VM I had, which to my recollection is pretty much a 100% pure clean install from the MSDN Win7 SP1 install disc.
using the recommended method with One-click maintenance, the utility found 19 issues.
"Issues"
Now, what constitutes an "issue" in it's eyes?
Let's see:
-some registry keys that contain no data, including a Typelib registration that has the important data in the key itself.
-Some uninstall entries that "have no information necessary to uninstall it"
-Some registry keys that point to 'non-existent files'
The last one is my favourite. Here are the files it lists as non-existent:
HKEY_CLASSES_ROOT\MSDASC\DefaultIcon
- "This Registry key points to the missing icon file %CommonProgramFiles%\System\Ole DB\oledb32.dll,0"
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Html Help Value IMTCTC.CHM
- "the entry %SystemRoot%\IME\IMETC10\HELP\IMTCTC.CHM points to the missing help file"
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\HTML Help\IMTCEN.CHM
- "The entry %SystemRoot%\IME\IMETC10\HELP\IMTCEN.CHM Points to a missing help file"
This might seem reasonable in some sense. If the files don't exist, why not delete the registry entries. I can see the logic behind that. The Problem here is that these files do exist!. Every single one of them. It didn't find them because it's not programmed to recognize environment variables, and figures it looks like a path so it should exist. performing this "routine cleaning" on a pretty much clean install corrupts your Windows Help configuration as well as corrupting some File Association data for Data Access stuffs. Wow. That is totally worth the 2KB of Data I might save on my hard disk! Deleting the typelib key actually prevented any applications from showing Help, Bravo. And ironically these are exactly the types of problems- as a result of that particular registry cleaner, at least- that people attribute to a "corrupt registry" - a never-ending cycle, like a dishwasher stuck on rinse.
Summary: A lot of the things provided as "performance tweaks" are only considered as such through myth and rumour and not typically through any particularly well-tested amount of evidence. As is not atypical of such 'tweaks' they appear to be based on the supposition that the computer is a magic box and we should make assumptions about how the configuration changes actually work and then base our "optimal value" on those assumptions. Which can sometimes work well but oftentimes blows up in the face or works to the detriment of performance or usability, particularly in this case the one about not running with a Pagefile.
I'm not really sure what the Start Menu has to do with it though. Most of the purpose he used the Start Menu for have more efficient ways to do it even in win7, and I think they are all semantically compatible with the Start Screen anyway. I imagine he went that way (hopefully) because it's the easiest way to guide a person through the various 'tweaks'.