Will Windows 10 Finally Address OS Decay?

colinneagle (2544914) writes The real question on my mind is whether Windows 10 will finally address a problem that has plagued pretty much every Windows OS since at least 95: the decay of the system over time. As you add and remove apps, as Windows writes more and more temporary and junk files, over time, a system just slows down. I'm sure many of you have had the experience of taking a five-year-old PC, wiping it clean, putting the exact same OS on as it had before, and the PC is reborn, running several times faster than it did before the wipe. It's the same hardware, same OS, but yet it's so fast. This slow degeneration is caused by daily use, apps, device drive congestion (one of the tell-tale signs of a device driver problem is a PC that takes forever to shut down) and also hardware failure. If a disk develops bad sectors, it has to work around them. Even if you try aggressively to maintain your system, eventually it will slow, and very few people aggressively maintain their system. So I wonder if Microsoft has found a solution to this. Windows 8 was supposed to have some good features for maintaining the OS and preventing slowdown. I wouldn't know; like most people, I avoided Windows 8 like the plague. It would be the most welcomed feature of Windows 10 if I never had to do another backup, disk wipe, and reinstall.

Comments

Here's the solution

[Anonymous Coward]

Stop installing crApps, and start using *real* programs.

Re:Here's the solution

[Anonymous Coward]

Silly muggle, there's no difference between the two.

Re: Here's the solution

[saloomy]

If the problem was really in the OS, then windows server which shares many of the same underpinnings as Windows desktop(s), would suffer the same fate. Since servers like domain controllers and exchange servers run for years without that issue, the problem seems to be from the crAPP that gets installed, as the parent explained, as well as the article. Bad headline to suggest the bad apps are M$'s problem

Re: Here's the solution

[Jane Q. Public]

Not really. It's just bad design.

Your server isn't getting games installed on it, which put all kinds of settings in the registry, then removed later when the game is old and tired, leaving behind cruft (including DRM bullsit) in the registry.

When a program is UNinstalled, all traces of it should be gone. Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

Re: Here's the solution

[timeOday]

It is not just applications. If nothing else, Windows does burn disk space as if it were free. Every version of every update ever applied (and what else? registry backups?) hangs around "just in case," you're not supposed to delete it, ever.

Here is the kind of answer I do not want to hear: "The typical cost of hard drives is less than .15 Cents per Gigabyte. This means that a WinSxS folder that is 6GB costs around .90 Cents, and uses slightly more than 1 Percent of the drive. That's about the same cost as a large bag of potato chips. " (cite). Yeah, so? Maybe I'm on a laptop with a small SSD? Maybe it's a VM that I have a dozen copies of? Don't waste my resources and then try to talk me out of caring.

Re:Here's the solution

[Anonymous Coward]

The real question:

Will Slashdot finally admit to being paid by Microsoft for positive press?

Re: Here's the solution

[Anonymous Coward]

8.1 brought huge improvements in disk space usage -- just look at the 16GB tablets shipping with it. Hopefully 10 will be even better.

Re: Here's the solution

[mysidia]

This means that a WinSxS folder that is 6GB costs around .90 Cents, and uses slightly more than 1 Percent of the drive.

I think you just don't understand what WinSxS is, how it works, and what the problem is that it is designed to address, suggest you start reading a bit more.

The reason the old Sxs assemblies need to be kept, is that installed software may require the usage of an old assembly.

Just because an update has superceded a certain library version, does not mean that all applications that still rely on it should be broken.

The SxS assembly backups have a vital role, and they don't actually use as much disk space as you think, due to hard linking --- Windows Explorer gives you an impression that more disk space is consumed by this folder than actually is.

The reason is... various installed files throughout the system will be hardlinked here, causing an appearance that a lot of space is in use here, but in reality --- these hard links are just a second Zero-usage copy of files that are installed elsewhere.

Only a couple gigabytes worth of files that have been updated and no longer have other hard links here, should actually be considered usage of the SxS system.

Re: Here's the solution

[K. S. Kyosuke]

A domain controller or an Exchange server is not likely to exhibit these issues unless you start doing things that are affected by this problem. Running programs that take a chunk of disk space and use it for reading and writing data in their own database format using their own code while responding to network queries for weeks is probably one of the least likely activities to get affected. (Unlike, say, the performance of the bookkeeping when installing updates etc.) It's similar to how even older versions of binary-translating virtualization packages weren't any slower at running user-mode code at full throttle: that's the part that's *not* being slowed down, it was the emulated peripherals that were.

Re:Here's the solution

[Anonymous Coward]

whut?

Re: Here's the solution

[Anonymous Coward]

> Just because an update has superceded a certain library version, does not mean that all applications that still rely on it should be broken.

Yes it fucking does.

Perhaps you don't understand having proper interfaces in the first place, so why don't you just fuck off back to Redmond.

Re: Here's the solution

[xpax666]

The software you are using is garbage. Applications don't HAVE to store ANYTHING in the registry -- that's the vendor of your crappy games doing that. Also, each vendor supplies an uninstaller -- obviously yours aren't doing the job. I suggest you have a talk with the thick-headed developers who write your games. Or, just join the rest of us in 2014, get an SSD and don't worry about it.

Re: Here's the solution

[Anonymous Coward]

I think it's time to take your meds.

Re: Here's the solution

[RyuuzakiTetsuya]

Even with an SSD, if applications are leaving behind shit in the various places shit can run on startup, you might be losing CPU or memory to some task that doesn't need to be there.

You can have this problem on other OSes like OSX and Linux too, but Windows is the only OS where the SOP is to make a mess of things. Don't like an app on OSX or linux? Just delete them. most of the garbage goes with it.

Re: Here's the solution

[RabidReindeer]

Not really. It's just bad design.

Your server isn't getting games installed on it, which put all kinds of settings in the registry, then removed later when the game is old and tired, leaving behind cruft (including DRM bullsit) in the registry.

When a program is UNinstalled, all traces of it should be gone. Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

You'd think so, but it's pretty common to uninstall a broken program, then re-install it. Keeping the old parameter settings makes it easier (sometimes!) to re-install, since you don't have to set it all up from scratch. Network settings are a case where I often benefitted from this behavior.

Then there's shared files and components such as DLLs. It's often quite difficult to reliably determine when the last using application has been removed, especially if people have been brute-forcing stuff instead of using the control panels and installers that keep count.

Still, I wouldn't mind having a "nuke" option that would at least remove all the files that belong exclusively to the app for cases where there's no possible reason why they would ever be referenced again. Murphy notwithstanding.

Re: Here's the solution

[rudy_wayne]

Not really. It's just bad design.

Your server isn't getting games installed on it, which put all kinds of settings in the registry, then removed later when the game is old and tired, leaving behind cruft (including DRM bullsit) in the registry.

When a program is UNinstalled, all traces of it should be gone. Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

Yes it is bad design. It is bad design by the people who create applications. The level of incompetence is staggering and this includes all the big name vendors.

Funny you should mention Apple. The Windows version of iTunes (the shittiest piece of software ever written) installs support files for 34 different languages. There is no option to only select the language you want at install time and the user is completely unaware that iTunes has just dropped approximately 4,400 un-needed files onto his hard drive.

But wait! It gets better! iTunes also creates a registry entry FOR EVERY SINGLE FUCKING FILE!! I am not making this up. ~4,400 files AND registry entries that can be deleted with absolutely no effect on the functioning of iTunes.

Re: Here's the solution

[Anonymous Coward]

First off I would say that one problem is that windows needs to be restarted every once in a while. Servers are included in that as every windows system I have worked with (quite a large number) becomes unstable if it is left running indefinitely. Linux does not have this problem (I have had more than one Linux system running for over 2 years straight with no change in performance or stability), The second problem is the slowness over time. I generally re-install my Windows system at least once per year because of this. I can also show that this does happen on Windows servers. The reason it is not seen on Windows Servers very often is because software is not normally installed and removed on servers as often as it is on desktops. Again Linux does not have this problem (at lest most Linux software does not cause it). I have had my Linux server running stably and just as fast as when I first installed it for over 5 years of randomly installing and removing all kinds of software on it and not all of it was from the repositories that come with the OS, there were quite a few things that I downloaded binaries to run and that I had to compile my self and some of it caused many different problems and created errors. Most of this was fairly simple to disable or remove and once it was removed the system was back to where it was before the software was installed.

Re: Here's the solution

[MightyMartian]

I remember in the transition between INI files and the registry (how I miss the days when applications had their own discrete text-based configuration files... oh wait, *nix still does!), and Microsoft sent out countless missives all but ordering developers to move to the registry. The registry was the approved place to store configurations, likely, I'm sure, because sticking all user settings in a single hive that could be passed around from workstation to workstation for roaming profiles.

Of course, the down side has always been that the registry just becomes cluttered with crap, particularly on a system that sees a lot of software installed, updated, reinstalled and uninstalled. Throw in there nearly two decades' worth of COM objects being incremented and decremented unsuccessfully, and a computer that's been running for five or six years, and fragmentation of the file system, and it can lead to just awful response times.

Re: Here's the solution

[Anonymous Coward]

8.1 brought huge improvements in disk space usage -- just look at the 16GB tablets shipping with it. Hopefully 10 will be even better.

The fact that they can cram their bloated OS onto a 16 GB device is not impressive.

Re: Here's the solution

[Quirkz]

The SxS assembly backups have a vital role, and they don't actually use as much disk space as you think, due to hard linking ---
Windows Explorer gives you an impression that more disk space is consumed by this folder than actually is.

That's nice to hear, because that folder often displays as upwards of 15GB, often as much as everything else in the OS combined, at least on the servers I'm looking at. Confusing, though, if that's not really what it's taking up.

Re: Here's the solution

[sunderland56]

You should be able to install 1000 programs, uninstall them all, and your system should be identical to what it was before. Anything else is a failure.

The very existence of the registry is wrong. Operating systems like Unix, Linux, MacOS, Solaris, etc. don't have a registry, and don't have any significant "OS Decay".

Re: Here's the solution

[abhi_beckert]

When a program is UNinstalled, all traces of it should be gone. Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

Apple took a different approach on iOS.

OS X suffers the same problems as Windows, although perhaps not as severe.

Re:Here's the solution

[Zynder]

You consider this article positive? Troll logic astounds me!

Re: Here's the solution

[Anonymous Coward]

Back when I still used Win7 in a VM, mine was something like 23 GB reported, but when deleted, 8 GB was freed. Which, to be clear, is still absolutely ludicrous.

Re: Here's the solution

[Anonymous Coward]

No, seriously, pull Microshaft's cock out of your mouth and stand up for yourself.

Re: Here's the solution

[Anonymous Coward]

Even with the hard and soft linking taken into account (there's a command line way to get the total amount of space actually consumed), my winsxs folder is still 15gb. Way more than a couple gigs. Very painful on my 60gb SSD.

Re: Here's the solution

[ericloewe]

Holy shit.

I mean, it's general knowledge that iTunes for Windows is most likely the worst piece of software ever written... But what you describe takes it to a whole new level of stupidity.

Hell, it almost makes it sound like they're trying to slow down Windows on purpose...

Re: Here's the solution

[Tumbleweed]

Servers don't have anywhere near as much software installed/removed as a desktop machine, so this is hardly an apples-to-apples comparison.

Re: Here's the solution

[Darinbob]

Microsoft went further than just having a bad registry, they essentially forced developers to use it if they wanted to get the approved sticker (ie, to say "works with Windows 95" or such).

Re: Here's the solution

[timeOday]

Although I still don't agree with you, upon further searching I see that Windows 8 (which I haven't used) did add a few more layers of goo that you can optionally spend time wrangling with to manage some of the winsxs bloat, and for that Microsoft perhaps deserves some credit. Perhaps.

Then again, winsxs is only one of several directories that often have people asking, "can I delete this?" See also C:\Windows\SoftwareDistribution, System Restore, and Windows.old, and c:\windows\installer. They are a mix of necessary and junk. From end to end, Windows is designed to keep everything, forever, just in case, instead of keeping track of things properly in the first place.

Re: Here's the solution

[zdickinson]

Microsoft recently released an update to address this with the disk clean up utility. It has come in handy on both servers and workstations. http://support.microsoft.com/k...

Re: Here's the solution

[toonces33]

Years ago, people would joke that "emacs" stood for "Eight megabytes and constantly swapping". At the time, it was meant as an insult. These days, 8Mb is hardly anything.

Re: Here's the solution

[GigaplexNZ]

I did a clean install of 8 and 8.1 in VMs to compare disk usage. 8.1 used slightly more.

Re: Here's the solution

[GigaplexNZ]

I'm well aware of the hardlinking that Microsoft does. I've also done some actual analysis measuring the real disk usage of WinSxS by using tools that count the hardlink references. On my Windows 7 installation that had a 16GB WinSxS folder, 14GB of that was unique to WinSxS with no other hardlinks. It isn't as efficient as Microsoft claims it to be.

Re: Here's the solution

[Joe_Dragon]

also a windows with reboot for 2 year will be missing a lot of updates as well.

Re: Here's the solution

[simplypeachy]

"Disagree"? Parent has provided proof. You can't disagree with what is factually correct.

If you are asking "Can I delete this random directory in my system directory?" then the answer is either:
"Here's a page on Technet explaining what it is for. Read this, then do not delete it."
or:
"If you have to ask, no."

Stop fucking around with things you don't understand and maybe your operating system will stop going slowly or breaking. Did you even look at the logo of sites like www.howtogeek.com? You're actually taking these people seriously, and going about ripping the innards out of Windows on the advice of some whack jobs?

Re: Here's the solution

[simplypeachy]

Oh I dunno. You haven't seen the state of some of the servers I've been asked to take over. Let's just say the word "iolo" and "Windows Server" should never been seen in the same screen.

Re: Here's the solution

[Anonymous Coward]

"Only a couple gigabytes" is both a lie and exactly the sort of attitude problem the GP was talking about. Out of curiosity, I did an informal survey of around 30 Windows 7 machines a few years ago. With hardlinks normalized, you're still looking at 6-8 GB for freshly installed and updated systems, and closer to 15 GB after some normal use (Office installed, etc). That is ridiculous. No other OS does this. There is something wrong with the OS architecture if this is truly "vital".

Re: Here's the solution

[Blaskowicz]

Compile and install a program on linux with ./configure, make and make install? then you will likely be left with no means to uninstall it at all. And I have no idea why there are non-library files in /usr/lib.

Re: Here's the solution

[budgenator]

Hell, it almost makes it sound like they're trying to slow down Windows on purpose...

Bingo

Re: Here's the solution

[Blaskowicz]

And then there was a transition to C:\Documents and Settings\username\AppData ? (or \Users\usernames\AppData which is the same)

Re: Here's the solution

[Anonymous Coward]

I just looked through my registry, and find no sign of these 4,400 entries you mention.

Not saying iTunes hasn't dumped a lot of gratuitous crap into my registry, because it has. But this "entry for every file" thing? Not... in evidence.

Re: Here's the solution

[timeOday]

"Disagree"? Parent has provided proof. You can't disagree with what is factually correct.

I take it you and the parent haven't actually used any other operating systems? They're not all like this.

I provided a link to a Microsoft-provided process that can often delete gigabytes of garbage from these directories, if you go to the effort of making it. The whole setup is a wasteful mess.

Re: Here's the solution

[mjwx]

Even with an SSD, if applications are leaving behind shit in the various places shit can run on startup, you might be losing CPU or memory to some task that doesn't need to be there.

Win Rot is alive and well in Windows 7.

Both my gaming machine and personal laptop have serious performance issues after 8 or so months (OK, the laptop is 2 years old but I use that infrequently). Both have SSD's, both were blisteringly fast when first installed.

Surprisingly enough, my work laptop is fine but I dont install much on there.

You can have this problem on other OSes like OSX and Linux too, but Windows is the only OS where the SOP is to make a mess of things. Don't like an app on OSX or linux? Just delete them. most of the garbage goes with it.

As a sysadmin, the biggest issue I have with Linux servers are the servers running out of space (mostly because some slovenly developer or DBA didn't bother writing a script to clean up log files or other output so it just grows until the disk runs out of space). Clogging up disk space with garbage is sort of *nix rot. Whilst Linux and OS X have no registry to clog up things, running out of disk space is a lot more painful on *nix than it is on Windows.

Very few *nix machines ever get used in the same abusive fashion as most people treat their windows boxen though.

Re: Here's the solution

[budgenator]

M$ did that to Xenix too hard and sim links eveerywhere

Re: Here's the solution

[mjwx]

This means that a WinSxS folder that is 6GB costs around .90 Cents, and uses slightly more than 1 Percent of the drive.

I think you just don't understand what WinSxS is, how it works, and what the problem is that it is designed to address, suggest you start reading a bit more.

The reason the old Sxs assemblies need to be kept, is that installed software may require the usage of an old assembly.

Just because an update has superceded a certain library version, does not mean that all applications that still rely on it should be broken.

When you've got a 19 GB Win SxS folder on a 40 GB HDD (which is plenty for a server and expensive on Tier 0 SSD's) it's a serious issue. 19 GB is not ridiculous, it's not even usual for a 2 yr old server that's been updated regularly. 19 GB across 250 virtual servers is a serious waste of space. Even 6 GB is a massive costs in infrastructure. Not every update needs to keep dozens of updates. Fortunately Microsoft has addressed this problem (as of April this year, so relatively quick in Microsoft time) so that the WinSxS folder can be cleaned up.

Re: Here's the solution

[tom229]

make uninstall Also compiling from source should rarely be necessary. Most modern distributions will include a ports like system that will allow you to compile source into a fake root, use the information gathered to build a package, and then install the package with your package manager. This ensures everything is cleaned up properly upon package removal. Of course even building a package for the software is probably unnecessary as it's very likely someone has already done it for you. Linux' package management is vastly superior to both Windows and osx (don't you just drag a folder into the garbage can? Give me a break). You just have to know what you're doing.

Re: Here's the solution

[simplypeachy]

There is no garbage in WinSxS. Windows keeps old or unused files for a reason. You shouldn't be deleting these - manually via a Microsoft tool or otherwise - unless you are low on disk space. Windows does its own automatic passes of storage directories and removes things it doesn't need itself. As the OP said, the size reported by Explorer of WinSxS is the only "garbage" because of the way the contents are handled and linked.

Things aren't as simple as the XP days, for better or for worse. However, in my experience reading the technical details of Vista+'s new features, they are usually for the better. They also make sense once you read up on them at a site that is authoritative and experienced.

Re: Here's the solution

[Anonymous Coward]

Just wait until systemd gets in there.

Re: Here's the solution

[Anonymous Coward]

Yes, but this is why you give /var its own partition. ;)

Re: Here's the solution

[Blaskowicz]

Good idea, create a .deb and install it.. but that's (as a matter of perception) hardcore sysadmin stuff, not user stuff.
Yes I'd like to compile some outdated or too new or different (*nix) stuff.. Or say, lone wolf projects, of which emulators are an example. And there are ppas or additional repos but then you need to manage the ppas and the guy doing the ppa might have done a crap job. (oh, I remember being told that in synaptics you can find installed packages that came from a ppa, but I wouldn't have thought of looking there and I still don't know how to do that from the command line).

Easy way out is with ./configure --prefix=/opt, but I don't feel that's right if I need to compile libraries before. Maybe I'm wrong. Else, I had thought of using some simple scripts or commands to know what files were added in /usr and such but I'd fear doing some non-robust crap.

Your suggestion makes sense, though it makes me feel like you're some Gentoo, BSD or Arch (?) user. I suppose it's all distro specific.

Re: Here's the solution

[mysidia]

Which is plenty for a server and expensive on Tier 0 SSD's

Microsoft's guidance on this is pretty clear last I checked; 32 gigabytes is the absolute minimum disk size for installing 64-bit versions of Windows server, and they wind up recommending a minimum of 80gb storage for most deployments, and their docs go on to state, you need to take into account the roles that will be installed, lifetime of the server and constant growth of the boot drive an additional 20gb per year due to updates. You do not need to install your C:\ drive on a SSD; there is no document recommending or stating that it is cost-effective and worth placing your boot disk on size limited high cost devices.

My recommendation would be that on servers you use the enterprise SSD devices for additional caching purposes or storing small files, such as SQL server tempdb, or your system paging file.

Re: Here's the solution

[Bite The Pillow]

You just killed 10 years of trial software, which will expire if not bought but can be activated by copying a string of text from an email.

I'm not going to argue the merits of trial software now, but being able to reinstall trial software would put a crimp in this business model.

Legit software should remove all traces of itself when uninstalled, I'm not arguing that. Freeware is included, paid software, all of that.

"Anything else is a failure", however, is pure ignorance.

Re: Here's the solution

[Anonymous Coward]

Operating systems like Unix, Linux, MacOS, Solaris, etc. don't have a registry

This deficiency will be corrected by systemd.

Re: Here's the solution

[Anonymous Brave Guy]

You just have to know what you're doing.

You can make Windows systems run crap-free and full speed again as well, by cleaning out all the obscure registry entries and system services and automatic updaters and cached thingies and temporary wotsits. You just have to know what you're doing.

Also, as with the part of the Linux strategy you forgot to mention, you have to be willing to spend forever doing it, because the tools provided as standard are just about hopeless.

Re: Here's the solution

[Anonymous Coward]

That's nice. I have a USB flash drive on my keychain that runs a modern Linux kernel and boots to a graphical environment. It weighs in at under 8 MB. What's your point?

16 GB is a lot of goddamn space for an OS today.
16 GB was a lot of goddamn space for an OS twenty years ago.
16 GB will be a lot of goddamn space for an OS twenty years in the future.

There is no reason for an OS to inflate just because storage is cheaper. OSes progress by being smarter, not bigger.

Re: Here's the solution

[Khyber]

"Both my gaming machine and personal laptop have serious performance issues after 8 or so months"

Meanwhile I've been running the same Windows 7 install since the tail end of 2009. That's with a fuckton of install, uninstall, and the occasional defrag and registry cleaning, especially on this tiny 120GB disk. Still runs exactly as it did back then.

Re: Here's the solution

[Anonymous Coward]

Here's the thing... anything inside the WinSxS folder means that at least one piece of software uses that dll version. If you have 16 gb in there, there's 16 gb installed / being used by various software.

Re: Here's the solution

[Anonymous Brave Guy]

Except there wasn't. Well, there was. A bit. Sometimes. Naturally, this half-baked approach actually made the problems worse.

Even today and with native Windows applications, many aren't very well behaved in following the "standards" here, because Microsoft did such a terrible job of promoting good practices.

Anything that isn't a native Windows application -- including almost every darling of the open source world, for a start -- probably ignores not only the application data directory but also the program files directories and insists on spewing its crap all over your filesystem and environment. Oh, and $DEITY help you if you need to do anything with Cygwin, and $CHORUS_OF_DEITIES help you if you have more than one ported application that requires Cygwin.

It is telling that you can't even schedule a backup of the "official" place to store documents without considerable effort, because Windows itself sets up so many links that most backup tools can't handle them.

And that's before you get idiots like the Chrome team at Google who think it's clever to install executable software in your data directory in order to deliberately circumvent Windows' normal security model, just so their auto-updater can do things it shouldn't without anything silly like troubling the user for permission. I'm always a little surprised that Microsoft hasn't, with considerable and legitimate justification, flagged Chrome's installer/updater software as malware and automatically removed it at some point.

On the bright side, if Microsoft can actually manage to produce an operating system with a sensible filesystem structure and application installation/update/uninstallation tools that actually enforce that structure, they might yet salvage the Windows brand and convince significant parts of their potential market to upgrade again.

Re: Here's the solution

[Anonymous Brave Guy]

Operating systems like Unix, Linux, MacOS, Solaris, etc. don't have a registry,...

True, and clearly a win.

...and don't have any significant "OS Decay".

ROFLMAO. IME, the only thing more painful than maintaining a Windows system over the long term is maintaining a *nix system over the long term.

Let's consider Linux. First, you probably get to choose between a stable or a not stable version of your distro. Choose stable and you're OK as long as you don't need to run any software released in the last 3 years and you're OK with being forced to upgrade the whole OS after maybe 2 years anyway (which will quite possibly trash your entire machine to the point of not being able to boot, or at least breaking minor features like RAID arrays, assuming you actually managed to configure one of those properly in the first place after your distro's "user friendly" installer messed it up completely). Alternatively, choose unstable if you want to run more recent software but don't mind stuff breaking all the time instead of every couple of years on a schedule.

Either way, if you want anything that hasn't got into your distribution's package management system yet, you're almost invariably forced into compiling your own software and manually installing it with makefiles. Those might, if you're really lucky, also offer a make uninstall option that actually does cleanly uninstall. That might, if you're even luckier, still work six months later, as long as no-one inadvertently installed a new version of the manually compiled code over the top to "upgrade" it, or just ran make distclean without thinking leaving you with no idea what make uninstall should have done. In any case, Linux is going to enforce absolutely no system hygiene at any point in this process.

OS X is of course doing much better with a similar foundation, as anyone who has spoken the words "Apple" and "shellshock" in the same sentence over the past few days can testify. Or at least, they'll be able to testify, just as soon as they've finished wiping and reinstalling their botnetted systems, because the patch everyone else had within hours only arrived for Apple gear several days later and long after exploits were widely found in the wild.

You're absolutely right that we should be able to install many programs and uninstall them with no lingering effects. But the idea that the registry is the only thing preventing that on Windows or that *nix systems do better is crazy. The only reason *nix systems don't break more often is that the only people running them are geeks and professionals, and those kinds of people are less likely to install random junk and more willing to dive in and fix internals when stuff goes wrong.

Re: Here's the solution

[Anonymous Coward]

Actually on OSX an app still pollutes its way by leaving Cache files behind. This sucks.

Re: Here's the solution

[Anonymous Coward]

from what I understand, the winsxs folder contains what appears to be lots of duplicate files, but are acutally hard-links to the same file. Doing a simple right-click -> properties on the winsxs folder and it shows 19 GB != the acutal space it's using. see here

Re: Here's the solution

[Anonymous Coward]

Well. Then why don't they do better Free Space accounting.

Re: Here's the solution

[Anonymous Coward]

Wow.

32 GB "absolute minimum" and 80 GB recommended? An additional 20 GB per year because of updates? Are you butt-fucking kidding me? Do you Microsofties ever take a peek at the competition? Has it never occurred to you that this is not normal or even reasonable?

Re: Here's the solution

[Anonymous Coward]

The very existence of the registry is wrong. Operating systems like Unix, Linux, MacOS, Solaris, etc. don't have a registry, and don't have any significant "OS Decay".

Tell that to the SystemD folks with the corruptible, non-ACID Journal file and whatever it is they use to save system state. (Solaris is only slightly less sucky in that they use SQLite for service state/configuration, so they didn't go full retard and try to re-invent the wheel.)

Re: Here's the solution

[Anonymous Coward]

For batch file and CLI use, %appdata% will do. Unless you're looking for local, non-roaming stuff. Then it's %localappdata%.

If you're developing in .Net, be sure to use Environment.GetFolderPath(Environment.SpecialFolder.ApplicationData) or Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData) unless you want to clutter your code with calls to Environment.ExpandEnvironmentVariable().

Any of these will avoid the ambiguity you have in your post.

Re:Here's the solution

[murdocj]

The real question:

Will Slashdot finally admit to being paid by Microsoft for positive press?

LOLOLOLOLOLOLOLOLOL

Re: Here's the solution

[mjwx]

Yes, but this is why you give /var its own partition. ;)

Great idea in theory, right until a Dev decides to put the log files out to /etc/application/logs and my employer frowns on the use of violence to enforce sensible ideas.

Re: Here's the solution

[Anonymous Coward]

As a sysadmin, you ought to know that Linux systems come standard with log rotation and management utilities that (surprise) sysadmins are supposed to know how to configure.

Re: Here's the solution

[tepples]

Unless Winqual disqualifies your application for storing settings in an INI, XML, or JSON file in %appdata% or %localappdata% instead of in the registry.

Re: Here's the solution

[mjwx]

Meanwhile I've been running the same Windows 7 install since the tail end of 2009. That's with a fuckton of install, uninstall, and the occasional defrag and registry cleaning, especially on this tiny 120GB disk. Still runs exactly as it did back then.

Its entirely possible you're slowing down at the same pace as your machine.

Re: Here's the solution

[kimvette]

No, but it is heading in that direction, if you run the Gnome desktop (why would you do that to yourself anyhow?); see gconf

Also, check out systemd. That's going to be fun to administer. >_<

Re: Here's the solution

[Anonymous Coward]

And of course the notion of moving the registry around between workstations when you roam is great, until you log in twice concurrently.

Which registry is then the right one?

Oh, ok, my profile is corrupted?

It would have worked better with lots of small files and syncing those that had changed. Even a brutal "keep the most recently updated config file" would work for roaming if they were discrete files.

Re: Here's the solution

[mjwx]

32 GB "absolute minimum" and 80 GB recommended? An additional 20 GB per year because of updates? Are you butt-fucking kidding me? Do you Microsofties ever take a peek at the competition? Has it never occurred to you that this is not normal or even reasonable?

As a sysadmin who deals with both Windows and Linux (Debian and Red Hat mainly) I can say that most vendors seriously over estimate their minimum requirements for servers. 40 GB is plenty for a 2008 R2 server, 60 if you're feeling generous.

An extra 20 GB for 40 servers is 800 GB on tier 0 storage (and yes, for these 40 servers they are required to be on SSD).

Its not just MS, regularly see tiny little packages designed for accounting or some such that have stupid requirements for a low number of users. Some things like a full version of MSSQL server, 8 cores, 16 GB of RAM and it only deals with about 10 GB of data, even some Linux applications that specify it must be a physical (and they'll refuse to help you if you install it on a virtual). Its sloppy testing and sloppy marketing.

As I said, in 2014 Microsoft finally admitted that the growing WinSxS folder was a concern and created a way to clean it up. It shrunk my 19GB down to just under 2. I understand why Windows needs to keep some old assemblies, but they dont need to keep all of them.

Re: Here's the solution

[David_Hart]

Not really. It's just bad design.

Your server isn't getting games installed on it, which put all kinds of settings in the registry, then removed later when the game is old and tired, leaving behind cruft (including DRM bullsit) in the registry.

When a program is UNinstalled, all traces of it should be gone. Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

Having leftover files and registry entries from apps that have been removed does not slow down Windows. Like any other OS, they just sit there doing nothing.

What does slow down Windows is disk fragmentation and lack of RAM.

Windows tends to have a lot of patches. Over time, these patches spread OS files across the hard disk. This leads to file fragmentation. When you are running a 5400 RPM drive, have a lot of apps installed and removed, and a lot of OS updates load files from the hard-disk slows down as the disk head has to travel a lot. Background disk defragging was first enabled by default in Windows 7. Of course, the ultimate solution for disk fragmentation is to use SSD drives as there is little penalty for fragmentation due to the high random memory access speeds.

Most Windows boxes today have at least 4GB of RAM but older Windows systems ran on 2GB of RAM. Even so, as you install more apps that have background components, they take up memory. When Windows needs more memory than what is available, it will cache unused parts of the OS and idle apps. This was accomplished by storing the cache data in the Windows page file on hard disk. When an idle app is clicked on, the cached data has to be reloaded and the previously active app needs to be cached back to disk. This whole process is really slow.

There are two ways to solve low memory issues:

1. Install additional memory: Windows Vista/7/8 32-bit recognizes a maximum of 4Gb of RAM. To install and use more than 4GB, you need to install the 64bit version of Windows. Most systems sold in the last 5 years run Windows 32-bit with 4GB of RAM or less. Only recently have systems been sold with Windows 64bit and 8GB of RAM. In my opinion, 8GB should be the minimum.

2. SSD: It might, at first, seem weird that I'm mentioning SSD in the memory section. The OS caches memory when it runs out of physical memory. SSD drives are really fast, much faster than physical drives. So, even with a system with low memory you would see a big difference using a SSD for your OS drive.

The point is that a combination of Windows improvements (background defrag on hard drives, 64-bit OS), technical improvements (SSD), and cost improvements (8GB RAM vs 4GB RAM) have contributed to eliminate the gradual performance degradation that we've seen in the past. It could be argued that adding better components to the system is just masking the limitations or design issues of the OS. But, as long as it works do we really care....

In my opinion, a Windows 7 64bit, or higher, system with 8GB of RAM and a SSD OS drive will not experience any performance issues over time for the average user. Windows performance degradation is a thing of the past....

Re: Here's the solution

[brantondaveperson]

Well..... I'm not for one minute saying that iTunes on windows doesn't suck. But I couldn't find these registry entries anywhere. Care to share where you found them? If iTunes continues to run without them there, and doesn't recreate them, then how do you know iTunes put them there in the first place?

Re: Here's the solution

[Anonymous Coward]

C:\Windows\WinSxS is using 20.1GB on my system, or 12.5% of my very expensive SSD. And then there's another 8.11GB of crap in C:\Windows\Installer that doesn't really need to be there.

Admittedly those things can be moved to other drives/partitions (albeit with some degree of risk) by altering some registry settings. But Microsoft applications (I'm especially looking at you, Outlook and WER/Windows Error Reporting) generally have a nasty habit of filling up a user's hard disk with what should be temporary files *in*hidden*folders* that never ever get deleted. The average user has no way of finding these things to remove them and the Microsoft-supplied Disk Cleanup wizard does fucking nothing about them. It's only thanks to tools like Steffen Gerlach's Disk Scanner and Jam Software's TreeSize that I even knew these fat folders existed.

Re: Here's the solution

[brantondaveperson]

Not in my experience. I've upgraded my OS all the way from 10.5 to 10.9 (I refuse to use those silly big cat names... Worse than Ubuntu's naming scheme, and that's saying something) and each time it's got faster. In that time I've installed and un-installed countless bits of crap software - and hosed and un-hosed my system on various occasions too.

Thus far, it runs just as fast as it did when I got it.

Re: Here's the solution

[atlasdropperofworlds]

From what I've been seeing, OSX degrades in the same way windows does. And you're definitely right about those issues with linux. Hell, even apt will fail to remove a package from time to time.

Re: Here's the solution

[Anonymous Coward]

How many Temp folders does Windows even need: C:\Windows\Temp, C:\Temp, C:\Users\\AppData\Local\Temp ... how many others?

Re: Here's the solution

[atlasdropperofworlds]

It seems to me there isn't anything special about any OS: Every single one degrades in performance as applications come and go and time passes. OSX is definitely not immune, nor is linux.

It's a phenomenon we've all been living with forever, but we're only really starting to notice it because you don't need to upgrade every 2 years.

Re:Here's the solution

[Anonymous Coward]

You consider this article positive?

In order to sell a new version of their product, Microsoft need to create discontent with the existing version.

There was no incentive for anybody to "upgrade" from W7 to W8, and very few businesses have made that step. It looks like Microsoft will be pushing "OS Decay" as their issue du jour, as evidenced by Slashdot's contribution to their marketing campaign.

Troll logic astounds me!

Your logic in suggesting my post is trolling is quite clear, despite being a fallacy.

Poisoning the well (or attempting to poison the well) is a rhetorical device where adverse information about a target is pre-emptively presented to an audience, with the intention of discrediting or ridiculing everything that the target person is about to say.

http://en.wikipedia.org/wiki/P...

Re: Here's the solution

[O('_')O_Bush]

Riddle me this, Batman...

If it is the application developer's fault (and frankly, the idea that every app has to be absolutuely perfect otherwise one permanently jacks up their OS is idiotic beyond belief), then why is the ONLY a problem in Windows?

You can install/uninstall crap on other OS's with no accumulative or persistent performance issues. Why? Because they weren't designed by idiots.

I have one W7 install that is *permanently* in limbo, unable to install Visual Studios redistributables because of a Windows patch bug, and there is literally no information on how to fix it, other than "reinstall OS" from the M$ horse's mouth.

That is when you know a design is pitiful and worthless.

Re: Here's the solution

[phantomfive]

Yes it is bad design. It is bad design by the people who create applications. The level of incompetence is staggering and this includes all the big name vendors.

Wait, are you saying you think the registry is a good idea?

Re: Here's the solution

[thegarbz]

make uninstall

You're assuming that a person made the effort to support that call.

But you're right that a package manager should resolve all that.

Re: Here's the solution

[thegarbz]

The problem is that you have to know what you're doing.
The problem is that it's not automated on an application level.

Even those tools which market themselves as cleaners of registry often pick up items which are still needed and delete them or miss things which aren't. The core problem is that for the average user keeping the system clean is not possible on a windows machine and an attempt to clean it often breaks things in strange ways.

Heck just thinking back to DLL hell days "It appears this file is no longer needed do you want to keep or remove?"
The only thing Microsoft missed from it's uninstaller package is "ARE YOU FEELING LUCKY PUNK?"

Re: Here's the solution

[scottbomb]

I have yet to see ANY OS "decay" with Linux. Been a daily user on several machines (Kubuntu and Xubuntu) for about 4 years now and haven't seen it yet.

Re: Here's the solution

[Anonymous Coward]

I've been using linux for 20 years.
I can not achieve 1000 ./configures and install all of the required dependencies, and then revert to the state 1000 installs earlier. How do you get every $HOME/.app directory ? Every doc, lib, header, config ?

OS X is easier. But it's also much more difficult than you would think. I've administered some large osx installs, and to really run the systems well involves *nix. See above

I am far too ignorant of the windows internals to use Windows well, and am a crappy windows admin. The tools I've needed to profile hardware utilization have been awesome. The tools needed to admin the system have been completely unfamiliar. I wish I had windows had an essential guide to the tomes of documentation. I'm pretty sure I could not log in, install a program from DVD, use the program for a few hours, uninstall, and leave windows exactly how I found it.

Gawd I suck at life.

Re: Here's the solution

[Anonymous Coward]

Rephrasing:
Crappy OS must stay bloated with crap left behind by crappy trialware to support their crappy licensing gimmics.

Re: Here's the solution

[chaboud]

I'll have to be the "arguably" part of this one. Junk is left on Macs, and a lot of settings and special folders can screw you pretty hard on a Mac. From plist files to Library litter to temporary files to /usr/bin linkages to /Developer folders, there are plenty of gotchas, secret handshakes, and areas where you can screw yourself or removal of an app can screw you. I don't know about you, but my user/Library folder on my machine is 38GB out of 256GB on this Mac, and if I delete all of the applications on this machine, that crap will still be there.

Same thing goes for my Linux machine. Cleanup and maintenance are tough problems to solve completely. iPhones and Androids serve as pretty solid rethinks of this (though they can also suffer from some rot), but I haven't used a desktop/laptop OS yet that doesn't rot over time without solid maintenance. For me, I'm looking at about 20% of my local storage on my two Library folders. Making .app folders skips out on any sort of cleanup responsibility, and 20% isn't chump change to me.

Re: Here's the solution

[Lord Kano]

That's the kind of nonsense that happens when you let Windows devs write Linux apps.

LK

Re: Here's the solution

[smallfries]

So basically what your saying is that unix has fixed the OS decay problem by removing the uninstall feature. Well that sounds reasonable :)

The solution to the other problems that you describe is not shrinkwrapped / easy. My latest attempt is a dual install with one partition for stable, one for unstable and a shared home. This allows disk snapshots of the not running system to allow primitive versioning and rollback.

Re: Here's the solution

[freeze128]

Apple is a major offender of installing stuff that automatically runs on startup... On Windows machines. They were also one of the first companies to do this circa 1995. When you installed Quicktime, it would automatically run a stub in the system tray in case you wanted to launch quicktime. Now, every damn application does that.

Also, just look at any Windows machine that has iTunes installed on it and you will undoubtedly see a bunch of other stuff like:

Bonjour
Safari
iPod service


Apple makes great software.... For apple computers, but they really know how to screw up Windows.

Re: Here's the solution

[Yomers]

Either way, if you want anything that hasn't got into your distribution's package management system yet, you're almost invariably forced into compiling your own software and manually installing it with makefiles. Those might, if you're really lucky, also offer a make uninstall option that actually does cleanly uninstall. That might, if you're even luckier, still work six months later, as long as no-one inadvertently installed a new version of the manually compiled code over the top to "upgrade" it, or just ran make distclean without thinking leaving you with no idea what make uninstall should have done. In any case, Linux is going to enforce absolutely no system hygiene at any point in this process.

If you want to install software that hasn't got into your distribution's package management system - you should compile it, make package and install package. How do you expect your OS to enforce system hygiene if you do not use correct procedures to install packages? If you install by 'make install' you are basically just copy bunch of files somewhere.

Re: Here's the solution

[Anonymous Coward]

That's can be solved easily with cron run scripts

Re: Here's the solution

[mysidia]

Most vendors seriously over estimate their minimum requirements for servers. 40 GB is plenty for a 2008 R2 server, 60 if you're feeling generous.

You don't get to say "They overstated their requirements," and, when the software actually needs to use as much resource as was stated to be required "Oh look, they're using too many resources". And no... 40GB is not "plenty" for a 2008 R2 boot drive; it is pretty much the absolute minimum for a couple years worth of service, with a likely space exhaustion, eventually.

An extra 20 GB for 40 servers is 800 GB on tier 0 storage (and yes, for these 40 servers they are required to be on SSD).

This is because of your broken deployment choices, and bad assumptions about OS space usage, not an issue with the software. Attempting to Micromanage microsoft operating system storage requirements will cost you more time in man-hours, than 1TB of SSDs ever would.

And you're accusing vendors of overstating requirements, while you're suggesting servers require SSD storage for just the boot drive?

Don't you see the irony in that? Of course servers don't need SSDs for the boot drive. Bloody SATA RAID5 is the most popular storage solution for Windows server system boot volume.

It's simply not true that SSD is needed for system boot drives.

Install the applications or databases requiring performance on SSD, not the OS.

Re: Here's the solution

[Barsteward]

"This deficiency will be corrected by systemd." - can you produce a link for me where it explains how and when the systemd registry will be in place and how it works?

Re: Here's the solution

[Barsteward]

is the systemd journal file a registry with config settings or a journal with log information?

Re: Here's the solution

[Anonymous Coward]

Dism.exe /online /Cleanup-Image /StartComponentCleanup /ResetBase

You can delete if you want - all running this command does is removes your chance to go back. There's also a scheduled task you can configure built into the OS that does the same thing

Re: Here's the solution

[Barsteward]

so i have to replace all applications when a library is changed? sounds like pie in the sky of a Utopian ideal where all application developers in all companies are working to the same schedule and release at the same time. Enjoy maintaining that system !!

Re: Here's the solution

[bingoUV]

Packages don't guarantee good cleanup, but it is easy to inspect a package to figure out if it does proper cleanup. Goes for deb and rpm.

Re: Here's the solution

[dbIII]

You'd think so, but it's pretty common to uninstall a broken program, then re-install it. Keeping the old parameter settings makes it easier (sometimes!) to re-install

Unless it's only broken due to registry settings that it follows as ordered even if they are stupid instructions. I think that's one is behind so many people advocating such extremes as a full OS reinstall every year or two, maybe longer on win7. Only a few applications and knowing enough to rip out the registry entries in such cases (I've had to do that far too many times for people and I'm a *nix guy when I work on computers) saves a full reinstall or falling back to an image.

Personally I see this as a failure of application programmers and testers to understand the platform they are working on than blaming it on Microsoft. The registry may be a stupid idea in some situations but it mostly works and it's nowhere near as stupid as those who write stuff to the registry without bothering to have a way of cleaning up afterwards.

Re: Here's the solution

[Anonymous Coward]

sudo apt-get remove ubuntu-desktop ...aand nothing happens. Thanks, ubuntu! Google about it and all you get are weak excuses about metapackages and whatnot.

Re: Here's the solution

[dbIII]

We had such things before the registry existed. It wouldn't work unless you input a word on page whatever of it's manual, a code on the box or some other secret outside of the file on the disk.

Re: Here's the solution

[clam666]

That's why I just format and reinstall every 6 mos. Apps don't matter, data does. I save data offline anyway as part of my retention plan. The OS and everything gets wiped, which eliminates all the app spoor and rootkits and various things that build up. I have no idea why people aren't wiping their system regularly anyway.

Re: Here's the solution

[Anonymous Coward]

And that is terrible.

Just look at the windows games.

Some save games to My Documents, some save to Saved Games (correct way) and some save to AppData (hidden) and even some save to game own directory (rare).

My AppData is over 30GiB by size, simply after installing 52 games from steam (not run a single one yet). Full of games own data pushed there. Office can take alone few gigabytes and web browsers as well.

The AppData is the dumping ground for devs. Slowing down system, making difficult to make backups (save games, settings etc).

Please devs, My Documents isn't place to store pictures, videos or music, those has own directories. Downloads should go to downloads. Saved Games directory should have a game name as Dir name and then have all from settings, profiles and save games in that.

AppData should not even exist as now. Roaming, local and local roaming are just terrible by idea.

Re: Here's the solution

[Anonymous Coward]

It might be part of their marketing plan, but it is still not positive press, as you allege. And your condescension doesn't change that.

Re: Here's the solution

[Anonymous Coward]

To poison the well there must first be a well.

Making unfounded wild claims then pushing others to object to those claims while providing citations is also a logical fallacy. I will let you stare at the overused info graph and tell everyone which it is.

Re: Here's the solution

[Anonymous Coward]

There's .1 of a difference :)

Re: Here's the solution

[martin_dk]

We have several Macs that have developed significant decay over time. I'm pretty sure Apple products is not exempted from this weakness.

Re: Here's the solution

[Anonymous Coward]

Fuck shared libraries. Yeah, it was a nice idea to save space when we didn't have enough. Just include every piee of code that runs with the program, lump all the data&code that comes with the program into one big selfcontained blob. Yes, you might have 4000 duplicates of some piece of code, but so what? Most programs have resource files that take way more space anyways.

Re: Here's the solution

[Anonymous Coward]

Calm down. Your mom will be back soon with more Hot Pockets and juice boxes.

Re: Here's the solution

[FlyveHest]

You would be the only one i've ever heard of whose Windows install would behave like that.

The slowdowns are happening over a long timespan, don't you think its more probable that you have just gotten used to it along the way?

Re: Here's the solution

[Anonymous Coward]

also a windows with reboot for 2 year will be missing a lot of updates as well.

A Linux machine with no reboot for 2 years may have a lot of updates installed but won't necessarily be using any of them. Unlike Windows, Linux keeps a single in-memory copy of executable code that gets shared by all running instances. Software updates are only replacing the on-disk files so unless you actually stop all daemons and drivers using a particular file (or just reboot the system and be done with it) the old versions don't get unloaded are still being used when new instances get started.

Re: Here's the solution

[NJRoadfan]

Doesn't MS permit the use of XML files for storing configuration settings? I seem to recall reading this in one of the .NET programming docs.

Re: Here's the solution

[NJRoadfan]

Technically programs can be installed to the user's home directory when you select "Install for this user only". Of course that is if the installer gives you the choice to begin with. Open source programs on Windows tend to clutter the home directory with configuration folders that start with ".", just like on *NIX systems.

Re: Here's the solution

[Overzeetop]

Actually, that's pretty impressive given that iOS barely fits into a 16GB iPhone.

Re: Here's the solution

[tgeek]

Are you sure about that? It;s been my experience that the "Install for this user only" option usually just places the program's Start Menu items into the individual users' Start Menu rather than the All Users. Possibly the same with the registry settings - never looked that close. But AFAIK, the actual executables still installed in their normal places,

Re: Here's the solution

[TargetBoy]

I'll chime in because that is my experience as well. with both home and work PCs.

Once I get through the install of the windows updates, the systems performance pretty much hits a steady state, as long as I maintain it.

I do see others having this problem and I really don't know what the difference in usage is. I install and uninstall software fairly routinely in the course of my job, so that can't be the only factor.

Re: Here's the solution

[Anonymous Coward]

Libpng. Ring a bell?

Re:Here's the solution

[drsmithy]

You consider this article positive? Troll logic astounds me!

Slashdot is the Fox News of the open source world. If you aren't blaming Microsoft for everything from Ebola to your dog's farts, you're a paid shill.

Re: Here's the solution

[Anonymous Brave Guy]

If you want to install software that hasn't got into your distribution's package management system - you should compile it, make package and install package.

I've heard that argument before, but if we're assuming the behaviour of make install/make uninstall is sufficiently non-trivial to worry about the system getting messy, how are you supposed to make that package without becoming an expert on what each piece of software's make install would have done anyway? The closest I've seen to automating this process is tools like checkinstall, but since make install can do arbitrary things, no automation tool can be fully trusted if you haven't vetted the makefile behaviour first.

Re: Here's the solution

[fisted]

Yes.

Re: Here's the solution

[Windowser]

Whilst Linux and OS X have no registry to clog up things, running out of disk space is a lot more painful on *nix than it is on Windows.

If a Linux machine is left with zero space, it will still boot but you will not have any log of it since there is no space to write it.
If a Windows machine is left with zero space, it will fail to boot.
So failing to boot is less painfull ?
Your logic escape me

Re: Here's the solution

[Anonymous Coward]

You are such a fucking moron, users are not going to do that you idiot so it is a flawed system. If you can do that you can clean up the registry too.

Re: Here's the solution

[XPACT]

Operating systems like Unix, Linux, MacOS, Solaris, etc. don't have a registry, and don't have any significant "OS Decay".

O RLY? Try deleting portion of /var/sadm/install/contents file which might be a 30-50Megs of text. The contents file is the "registry" of Solaris.

Re: Here's the solution

[nine-times]

Apple took a different approach, which arguably works far better. Even if stuff is left behind, it just takes up a bit of disk space, and doesn't affect the system at all.

Depends on the app. Some have installers that can be pretty obtrusive as well, but Apple doesn't provide an "Add/Remove programs" thing to uninstall. Their older apps, you could just drag/drop wherever, which I generally preferred, but unfortunately some things really do need installers. Then there's their App Store apps, which are generally sandboxed in a way that, similar to the drag/drop installs, will only really leave behind preference files, which take up disk space but are unlikely to cause problems. On the other hand, some of those (e.g. Apple's own Xcode and Server apps) have installers that make changes to the system on first run, and those are pretty impossible to remove.

I'd like to see a system which monitors every change to the system during the install process and first-run, and then the uninstall automatically rolls all of those back. I'm sure you'd have to make some concessions, maybe allowing certain settings to be retained on an uninstall if the developer specifically excludes them, but it's always been downright silly how many things get left behind when you install/uninstall an application.

Re: Here's the solution

[jeremyp]

Woosh!

Re: Here's the solution

[behrooz0az]

Couldn't agree more.
and parent had the answer to his /etc/application/logs quoted. give it it's own damn partition if it's big, mount --bind it maybe?

Re: Here's the solution

[jeremyp]

OS X comes with a utility called Migration Assistant that can effectively duplicate an old machine onto new hardware. Technically, despite several iterations of new hardware and operating system upgrades, I am running the same installation as I was with 10.5 in 2007 and I have not experienced any operating system rot. My current four core 16Gb SSD machine is just as fast as my original 32 bit two core 4Gb MacBook pro.... .... oh wait...

Seriously though, I wonder how much operating system rot is really down to disk fragmentation. I don't think I've ever had a machine that suffers from it, but with Windows boxes I do defrag regularly.

Re: Here's the solution

[Anonymous Coward]

ROFLMAO. IME, the only thing more painful than maintaining a Windows system over the long term is maintaining a *nix system over the long term.

Let's consider Linux. First, you probably get to choose between a stable or a not stable version of your distro. Choose stable and you're OK as long as you don't need to run any software released in the last 3 years and you're OK with being forced to upgrade the whole OS after maybe 2 years anyway (which will quite possibly trash your entire machine to the point of not being able to boot, or at least breaking minor features like RAID arrays, assuming you actually managed to configure one of those properly in the first place after your distro's "user friendly" installer messed it up completely).

Or not.

Alternatively, choose unstable if you want to run more recent software but don't mind stuff breaking all the time instead of every couple of years on a schedule.

Either way, if you want anything that hasn't got into your distribution's package management system yet, you're almost invariably forced into compiling your own software and manually installing it with makefiles. Those might, if you're really lucky, also offer a make uninstall option that actually does cleanly uninstall. That might, if you're even luckier, still work six months later, as long as no-one inadvertently installed a new version of the manually compiled code over the top to "upgrade" it, or just ran make distclean without thinking leaving you with no idea what make uninstall should have done.

Or not

In any case, Linux is going to enforce absolutely no system hygiene at any point in this process.

Or.....whatt? -sniffs system- Seems ok.....

OS X is of course doing much better with a similar foundation, as anyone who has spoken the words "Apple" and "shellshock" in the same sentence over the past few days can testify. Or at least, they'll be able to testify, just as soon as they've finished wiping and reinstalling their botnetted systems, because the patch everyone else had within hours only arrived for Apple gear several days later and long after exploits were widely found in the wild.

You're absolutely right that we should be able to install many programs and uninstall them with no lingering effects. But the idea that the registry is the only thing preventing that on Windows

Is crazy, yes. The registry is not the only thing Windows does wrong/messily, agreed.

or that *nix systems do better

Is a lot more sane. You don't *have* to package source install trees into deb/rpm/whatever packages before dropping them into your root filesystem, but you do have the option (and are kinda nuts if you don't, especially on a serious system). Maybe there's an equivalent '.msi HOWTO' somewhere out there, I've never looked for it - but isn't this entire thread of conversation about how the MSI system leaves crap lying around after uninstallations that it shouldn't, so even if you do follow best practises for installing (originally) non-packaged software then later uninstall it you're left with junk?

is crazy. The only reason *nix systems don't break more often is that the only people running them are geeks and professionals, and those kinds of people are less likely to install random junk and more willing to dive in and fix internals when stuff goes wrong.

And the reason that people who are less likely to do dumbass shit and fuck about with systems best left unfucked gravitate away from 'maintaining a Windows system over the long term' is....?

Re: Here's the solution

[Anonymous Coward]

If the problem was really in the OS, then windows server which shares many of the same underpinnings as Windows desktop(s), would suffer the same fate. Since servers like domain controllers and exchange servers run for years without that issue, the problem seems to be from the crAPP that gets installed, as the parent explained, as well as the article. Bad headline to suggest the bad apps are M$'s problem

So the answer is to buy a Windows PC and lock it in a backroom and access it via services? Maybe use a Linux or OSX client? Fact is, even a highly controlled corporate desktop/laptop will slow down. It just happens.

Re: Here's the solution

[Anonymous Coward]

Why don't you just go back to system restore point #1? This will wipe the registry and unload every DLL ever installed. Things may remain on the disk, but windows will be restored to zero, while leaving documents intact. Simply reinstall the apps you want to use and you're back in business.

Re: Here's the solution

[Jane Q. Public]

The software you are using is garbage. Applications don't HAVE to store ANYTHING in the registry -- that's the vendor of your crappy games doing that. Also, each vendor supplies an uninstaller -- obviously yours aren't doing the job. I suggest you have a talk with the thick-headed developers who write your games. Or, just join the rest of us in 2014, get an SSD and don't worry about it.

Utter nonsense. Microsoft won't certify your product UNLESS you store certain things in the registry.

Certainly you can write your programs a different way (as I do), but don't expect to get your software Microsoft certified.

Re: Here's the solution

[Anonymous Coward]

Or they followed Microsoft's methodology for software development. You can complain all you want about incompetent developers/companies, but most just follow the Microsoft standard, which is to not use configuration files but use the registry. Far too often Microsoft treats programmers as hobbiests and doesn't take what they do seriously. It you don't actively keep up a MSFT certification, you'll never hear about You buy VisualStudio Enterprise and you're left up to you own. It's suggested that you buy MSDN. The companies I work for don't generally give developers access to it, they use it for the "free" software.

I blame MSFT for the state of Windows software. They define the culture that encourages bad software.

Re: Here's the solution

[Khyber]

I still have a 98SE machine that runs extremely quickly.

Here's the trick - LEARN THE FUCKING OS.

I'm not slowing down. I've taken timed measurements of boot-up times and maintained them over the course of several years.

I've been at a steady 13 second boot from power-on for the past 3 of those years on 7. 8 years on XP at 17 seconds, 13 years on 2K at 8 seconds. 98SE (on a 450MHz P3 laptop with 640MB RAM) at 6 seconds for about 15 years (one reinstall because of a botched driver, not the fault of the OS.)

I also had WinME running as smoothly as 98, but the mobo died there and that machine is hence retired.

And that's on top of changing hardware every other year where feasible. All machines still operational with the exception of the one mentioned right above.

Sounds like you and those you know don't know the OS, how it works, or how to utilize it.

Re: Here's the solution

[Chelloveck]

So let me get this straight... Your argument is that people are so content with Windows 8 that Microsoft has to start a poison well campaign against it to convince people to upgrade? Really?

Re: Here's the solution

[Anonymous Coward]

Win7 added the known folder constant FOLDERID_UserProgramFiles and by default this is %LOCALAPPDATA%\Programs which is what per-user applications are supposed to use...

Re: Here's the solution

[MobyDisk]

I can confirm this, having worked for a company that wrote backup software. While Microsoft does have some rules in place, they are not good enough. We had to make all kinds of exclusion rules to prevent backups of temp files, cache files, etc.

Re: Here's the solution

[mythosaz]

http://technet.microsoft.com/e...

Determine the Actual Size of the WinSxS Folder
Applies To: Windows 8.1, Windows Server 2012 R2

Dism.exe /Online /Cleanup-Image /AnalyzeComponentStore

That'll output the actual size of the sxs folder.

Re:Here's the solution

[Anonymous Coward]

If this is supposed to be positive press, it fails miserably!

- finally address a problem that has plagued pretty much every Windows OS since at least 95

- I avoided Windows 8 like the plague

On top of that, the question is not answered, nor is any reason giving to make us think windows 10 even attempts to fix degradation

Who's paying you to troll slahdot?

Re: Here's the solution

[MrVictor]

Since this is Windows we are talking about, I'll just go ahead and assume you were taking about %DEITY% and %CHORUS_OF_DEITIES%

Re: Here's the solution

[Anonymous Coward]

apt-get install logrotate

Re: Here's the solution

[aix tom]

The software that "stores everything in the registry" is not garbage per se. It's just software that was written by developers that listened to Microsoft between, say, 1995 and 2010. Because they were told "nonono, *.ini files are bad, store everything in this great registry thing we invented".

From my personal gut-feeling that "drive to use the registry for almost everything" peaked around 2005-2006, before the trend was reversed. At least judging from the work related "enterprise stuff" where I still (have to) have some contact to Windows machines.

Re: Here's the solution

[DarthVain]

I manage a bunch of legacy apps that all use INI files, and it is one of the things I like. Your DB move locations? Update the local INI file to reflect the new location. It takes notepad, a cut and a past, or about 5 seconds. Recently had to update a new .NET application for the same purpose... oh that will be actual development and support costs, etc... Poor design didn't help either.

Also makes testing easy, being able to switch between dev/test/prod DB instances simply by using a ";".

As an added bonus, because it doesn't use registry you also get around the privileges BS, when in a corporate environment can be a pain as usually no one has them. Of course security probably likes it.

Re: Here's the solution

[Anonymous Brave Guy]

Well, I was talking about Cygwin at that point. That's my excuse, and I'm sticking to it. ;-)

Re: Here's the solution

[WindowPane]

This is exactly part of the issue of the slowdown over time. The system spends more and more time searching for the correct files in the SxS jungle to use with each system call. Yes the registry can get muddled up but bloat caused by MS architecture failures seems to be the culprit in my opinion. I don't have metrics or references just 20 years experience as an admin. The SxS mess was created to fix the DLL hell mess.

An often overlooked killer is the environment path. Say for instance you install a program that prepends their path to the environment path and guess where the OS starts it's search every time it looks for an object on disk, in that folder. The entry %SystemRoot%\system32 should be the first in the path.

Re: Here's the solution

[Anonymous Coward]

WAKE UP, SHEEPLE!

Re: Here's the solution

[Anonymous Coward]

My dog is now farting blood. I will never use a Microsoft product again.

Re: Here's the solution

[Anonymous Coward]

I just looked through my registry, and find no sign of these 4,400 entries you mention.

Not saying iTunes hasn't dumped a lot of gratuitous crap into my registry, because it has. But this "entry for every file" thing? Not... in evidence.

Well, it might depend on the version of iTunes, but I'd bet that you just don't know where to look. Few people do, even many experienced Windows admins and devs, because almost none of them actually write installers. Of those who do, most just go through some clicky InstallShield Express wizard built in to Visual Studio.

Look under the key where MSI stores all the obfuscated GUIDs which represent installation component IDs. If you have a system with even a modicum of software installed on it, this area is just chock full of sub-keys. It gets better: on Windows XP systems with a lot of complicated software installed (e.g. multiple versions of MS Office and development tools), the number of keys can grow beyond what the built-in registry editor utility can display in the tree-view in the left-hand pane. So, you might need a third-party registry viewer to be able to see all the MSI component keys on XP; I don't recall whether later OS registry editor utilities had the same limitation.

- T

Re: Here's the solution

[Anonymous Coward]

I hear you. Just had a dev say "Come on its 2014 just give every VM 8 GB of RAM". And we were talking about increasing RAM for about 20 VMs that were currently at 2GB by 1GB ... Most devs have no fucking clue about systems administration. Just glad I'm dev lead and our architect and devops do have a clue.

Re: Here's the solution

[NJRoadfan]

I seem to recall reading about it on slashdot, but the closest reference I could find is here: http://stackoverflow.com/quest...

ClickOnce .NET applications may be what I was thinking about.

Re: Here's the solution

[markus.neifer]

This is my experience too. If you stop installing / uninstalling, chances are good you avoid system slowdown. Use virtualization to try things out. For example with Vmware or VirtualBox.

Re: Here's the solution

[Festeron]

Fortunately Microsoft has addressed this problem (as of April this year, so relatively quick in Microsoft time) so that the WinSxS folder can be cleaned up.

How did Microsoft address the WinSxS problem last April?

Re: Here's the solution

[lsatenstein]

If the problem was really in the OS, then windows server which shares many of the same underpinnings as Windows desktop(s), would suffer the same fate. Since servers like domain controllers and exchange servers run for years without that issue, the problem seems to be from the crAPP that gets installed, as the parent explained, as well as the article. Bad headline to suggest the bad apps are M$'s problem

The problem is one of recovery files that accumulate. MS never deletes a recovery file. If MS kept three or four generations only, then one could delete these obsolete disk fragmenters and performance would be restored. And a good defragger would help too.

Re:Here's the solution

[tedgyz]

Are you kidding? The only thing /. hates more than Microsoft is Apple. So I guess if you are Apple fanboy you might see it that way. Oh... Never mind.

Re: Here's the solution

[Maxwell]

So all windows software is garbage...including the makers of this suite called "Office". Would you be so kind as to call Microsoft for me? I am busy calling every other developer on the world

Re: Here's the solution

[toddestan]

I've never figured out that argument. Who, having experienced the crappiness that is iTunes on Windows, would want to buy a computer where the entire OS is created by the same company that made iTunes for Windows? I mean, it would be like running an OS created by the same people that wrote Java!

Re: Here's the solution

[budgenator]

If your narcissistic enough, you'll fingure that people will switch from Windows to OSX or IOS just to experience the full Itunes experience, if the version of wiindows is Win8/8.1 it not very much narcissism either.

Re: Here's the solution

[Anonymous Coward]

Agree with Ms. Public, a server is a very different environment than a desktop, app, environment, saloomy is comparing two totally different worlds. The MS OS has always been a bad design for desktop environment. But I have used Windows 8 now for two years and it has not slowed down and I have not done any clean up other than malware removal occasionally so perhaps MS did get the OS problem corrected with Windows 8.

Re: Here's the solution

[missneht]

Agree with Ms. Public, a server is a very different environment than a desktop, app, environment, saloomy is comparing two totally different worlds. The MS OS has always been a bad design for desktop environment. But I have used Windows 8 now for two years and it has not slowed down and I have not done any clean up other than malware removal occasionally so perhaps MS did get the OS problem corrected with Windows 8.

Re: Here's the solution

[tom229]

Arch, good guess. You're obviously using something debian based, probably Ubuntu. I haven't used Ubuntu in some years, but I don't remember them actively supporting the use of a ports-like system. It would certainly be possible in Ubuntu, but without community involvement, rather pointless.

Give Arch a try in a VM. I've yet to encounter a piece of software that hasn't had the source converted to a make package by someone in the AUR. Using the AUR is as simple as downloading the tarball, extracting, running makepkg, resolving dependencies, then pacman -S .

Oh, and don't use the base arch installer unless you have a lot of time on your hands. Go with something like ArchBang.

The bigger Problem is their "updates"

[Opportunist]

Sadly the way updates work with MS they become the far bigger problem. You can easily see this by installing a "clean" system, examine its timing (please don't even think about using system internal benchmarks...), then patch it and notice just how much speed you suddenly miss.

That's a problem you probably won't solve quickly...

Re:The bigger Problem is their "updates"

[exomondo]

You can easily see this by installing a "clean" system, examine its timing (please don't even think about using system internal benchmarks...), then patch it and notice just how much speed you suddenly miss.

Not that I can be bothered actually doing that but since you're saying that I'm guessing you've done it and had significant results, what were they and for which version?

Re:The bigger Problem is their "updates"

[Anonymous Coward]

The easiest test is on Windows 2000 with a clean install. Then apply the Services Pack 4. It's like hitting a wall.

Re:The bigger Problem is their "updates"

[Zynder]

Don't know about him, but I do believe we had an article on here back in the spring about the updates being a huge issue for Win XP. Dunno about 7 or anything else. Yeah, yeah, I know, Win XP is old and we shouldn't care but you go ahead and tell my 2007 model Thinkpad that. It still flies on XP, better than my newish AMD E2 on Win 7.

Re:The bigger Problem is their "updates"

[GigaplexNZ]

People don't tell you XP is old because of performance reasons. It's a security nightmare.

Re:The bigger Problem is their "updates"

[simplypeachy]

That was a one-off bug that slowly accumulated over a very long time for a rather obscure reason. Please provide compelling evidence. Or at least bullshit that looks, from a distance, like something evidence-shaped.

Re:The bigger Problem is their "updates"

[exomondo]

Don't know about him, but I do believe we had an article on here back in the spring about the updates being a huge issue for Win XP.

Yes a bug in the windows update algorithm for previously applied updates, that would occur predominantly when running windows update if you had IE 6 or 7 installed. Given that XP support has been dropped there is no reason to run windows update on it so it isn't an issue anymore.

Yeah, yeah, I know, Win XP is old and we shouldn't care but you go ahead and tell my 2007 model Thinkpad that. It still flies on XP

Well then this OS decay issue probably doesn't affect you.

Re:The bigger Problem is their "updates"

[Khyber]

Hah, I have a system running Win2K with SP4.

It's a 1.8 GHz P4 with 768MB PC-3200 DDR, on an IDE HDD.

It boots faster than my dual core 64-bit laptop with 2GB RAM with Windows 7. By about 50% less time.

2K with everything updated is roughly 2GB, total size.

Re:The bigger Problem is their "updates"

[Khyber]

What security nightmare? Still rolling sans firewall, DMZ, and the only thing I have to worry about is a DDoS hitting my multiplexed cable modems.

Re:The bigger Problem is their "updates"

[Anonymous Coward]

If the patch includes an update to the .Net Framework (any version, any of its libraries, or even any dependencies) or anything else that gets stuffed into the GAC, then you have to wait about 12-24 hours before running the post-update test. The reason is that there will be 4 concurrent compilation and linking processes running during time when the system idle timer goes above a certain threshold (based on number of virtual cores, amount of memory, and possibly even that stupid Windows Experience Index thing). These recompile and re-link the entire .Net Framework (whatever version just updated). Until they are completed with that task, everything runs a bit slower due to monitoring overhead (to see if system idle is idle enough), in-progress tasks (they don't abort until they've finished a unit of work), and lack of cached libraries (they're rebuilding the cache, so anything the compilation pool hasn't gotten to yet has to be done on-demand). For a large .NetFX update, that can take up to 24 hours, uninterrupted, on a slow machine.

Personally, I have a Win7 install that dates back to about 2011, just before SP1 came out (I think). It runs as fast as it ever did, with the exception of the first couple of minutes after startup. SQL Server 2012 Developer Edition is a damned resource hog.

Re:The bigger Problem is their "updates"

[kesuki]

I have a dell laptop with a 800 Mhz processor and it runs lubuntu faster than lubuntu on a lenovo quadcore 1.2ghz AMD apu...same os. one is what most would call obsolete and yes i installed with the same optical disc. the APU boots up faster but takes a full 30 seconds to respond after clicking the 'log out' button it is no lag if i sudo shutdown -h now from terminal.

knoppix warns that using the live media and eventually the system begins to slow, and that still applies to a write protected flash memory card. no i don't mean just leaving the computer booted up all the time the parts inside the computer slow down as they get older. there are a ton of things that cause problems that good design can deal with most issues people get with windows but that doesn't stop bitrot or normal failure or slowdown of the chip and other components...

the only thing that deals with component slowdown is buying new components. there is not one OS that doesn't have the same problems without requiring replacement of parts known to fail or slow down things.

Re:The bigger Problem is their "updates"

[DocSavage64109]

Boot time isn't a good benchmark. I had a 90Mhz pentium that would boot into windows 3.1 (from DOS) in something like 0.1 seconds.

Re:The bigger Problem is their "updates"

[Zynder]

I typically only use 2 devices to do anything that one would really be concerned with regarding security. That XP machine is only for web surfing, arduino programming, and a little bit of World of Warcraft. My banking and purchasing are done on my Win7 desktop which is fully patched, firewalled, AV'd, Firefoxxed, Noscripted, Ad Blocked, Peer Guardianed, and when I want it VPN'd.

Re:The bigger Problem is their "updates"

[Zynder]

I wasn't arguing so you can get your own evidence. I was just pointing out that there are issues with accumulated updates whether they are a "bug" or not. And speaking of, what classifies that as a bug anyway? How shitty does a system have to run before it is just a bug and not something systemic?

Re:The bigger Problem is their "updates"

[Zynder]

Never said it was. You asked "what were they and for which version" and I replied. It isn't even a slow process as many seem to claim. Take a fresh XP install and test drive it. Fairly quick. Then turn on Windows Update and let it download the 200+ updates and then drive it again. You'll ask where the other half of the engine went!

Given that XP support has been dropped there is no reason to run windows update on it so it isn't an issue anymore.

Are you a software/OS guy like most Slashdotters (I'm not, I'm hardware)? I ask because if you are, I can't believe you are actually advocating that I ignore every security patch that Windows put out- regardless if they will make no new ones. That's like telling me to not worry about getting a polio vaccine because it hasn't been a problem in ages. I'll get my shots (real or virtual) just in case. But FYI, to not be supported one or 2 updates pop up on windows update every month or two. I, unfortunately, can't remember what they are right now. Probably the Malicious Software tool that I never use (but install anyway).

Re:The bigger Problem is their "updates"

[Anonymous Coward]

There was an issue with the updates on Windows XP, but the problem was with the update process taking too many resources, not about installed updates causing a general slowdown. Here is one of the articles that came on Slashdot about that topic.

Re:The bigger Problem is their "updates"

[exomondo]

Never said it was. You asked "what were they and for which version" and I replied.

Well you're telling me your Thinkpad "flies on XP" so it can't be that bad.

Are you a software/OS guy like most Slashdotters (I'm not, I'm hardware)? I ask because if you are, I can't believe you are actually advocating that I ignore every security patch that Windows put out- regardless if they will make no new ones.

What? I said you don't need to run Windows Update anymore, the bug that was reported in the spring about updates causing a huge issue was in Windows Update. Since there are no more updates once you're up to date you no longer need to run Windows Update because there will be no more updates.

Re:The bigger Problem is their "updates"

[Luckyo]

When it can be properly reproduced by other people, and not just one person shitting up their PC with malware and then whining about "slow downs", you can start calling it a bug.

A good example was XP update overflow bug, that caused update process to become cripplingly slow right before the end of support.

Another is a known issue with XP SP2, which is more of a feature than a bug - it added several complex features to the OS, which upped system requirements.

Re:The bigger Problem is their "updates"

[Luckyo]

That's not boot of the system, but boot from OS into GUI.

Re:The bigger Problem is their "updates"

[symbolset]

This is a good place to point out that what to a Windows user is a successful OS life cycle isn't even close to what we would consider a good uptime.

Re:The bigger Problem is their "updates"

[CptJeanLuc]

Sadly the way updates work with MS they become the far bigger problem. You can easily see this by installing a "clean" system, examine its timing (please don't even think about using system internal benchmarks...), then patch it and notice just how much speed you suddenly miss.

Compared to osx and linux distro updates, Windows (at least Win 7) is a true dinosaur. Imagine how many man-hours are wasted worldwide while waiting for Windows to update, with a reboot required pretty much every single time. Even if you don't consider the time spent applying a patch during shutdown, there is often the additional waiting during boot, and more often than not it seems Windows want an additional reboot during startup. Which sucks hard if you have default dual-boot into Linux, because you fire up the PC, choose Windows, go grab a coffee, and when you come back ... behold, there is the Linux login. Because Windows of course decided to do some additional rebooting.

Yes, osx some times goes offline for a while when applying a large system patch, but this happens only every few moons, whereas with Windows you know you are in for a system update ride if you haven't touched that particular install in a couple weeks.

Re:The bigger Problem is their "updates"

[Anonymous Coward]

I post this nearly every time a discussion like this comes up, but here it is again. Windows 8 properly configured (install Classic Shell, 7TT, tiny window borders, turn off unnecessary effects) is going to be just about as fast and much more usable than XP.

Re:The bigger Problem is their "updates"

[rtb61]

It's called the system registry, that buggy plague ridden POS idea that has contaminated windows for decades, remains a continuous problem. Originally done so M$ could pry into what software people are running in one location and prevent undesired software from M$'s point of view from running, an idea that had to be abandoned for obvious reasons but that POS registry crap was still left behind. A brand new clean install more than anything else tidies up the registry and that speeds up boots, shut downs and application launches. The system registry is shite, get rid of it, finally for fukity fuck fucking sake.

Re:The bigger Problem is their "updates"

[simplypeachy]

I think the length of time it took to be officially recognised by Microsoft puts weight behind your question. I'd noticed it was slow for ages - even for freshly-installed machines. I didn't think too much of it, as it was always on old hardware.

And you're right, you weren't arguing. I apologise for flying off the handle at you when I should have been aiming elsewhere, your comment did not deserve that.

Re:The bigger Problem is their "updates"

[CaptSlaq]

It's called the system registry, that buggy plague ridden POS idea that has contaminated windows for decades, remains a continuous problem. Originally done so M$ could pry into what software people are running in one location and prevent undesired software from M$'s point of view from running, an idea that had to be abandoned for obvious reasons but that POS registry crap was still left behind. A brand new clean install more than anything else tidies up the registry and that speeds up boots, shut downs and application launches. The system registry is shite, get rid of it, finally for fukity fuck fucking sake.

Good luck with that. Registry keys are pretty much a staple of most software that's written for Win32. The idea actually isn't that bad: It's a place to stuff data about and settings for software in a standardized way. Thing is that my understanding of it is that the entire thing resides in RAM, so the fatter it is, the less that's available to the system. Someone with more/better information on this would be interesting to hear from.

Re:The bigger Problem is their "updates"

[operagost]

It does still push out the malicious software removal tool.

Re:The bigger Problem is their "updates"

[Khyber]

"no i don't mean just leaving the computer booted up all the time the parts inside the computer slow down as they get older."

If that is happening, you've got hardware engineering problems, not OS problems.

Just speaking from my meager 6 years of semiconductor fab and design experience. Shit doesn't degrade gracefully in hardware. It starts throwing errors left and right then dies outright. There is no steady slowdown that increases and then suddenly reverts when you reinstall the OS.

I've got 33MHz computers from 286 days, still just as fast as they ever were (one still runs a multi-node BBS!)

Re:The bigger Problem is their "updates"

[Khyber]

Given 3.1 is just a shell and not an actual OS, not a surprise. 3.1 still utilized all of the DOS OS, which wasn't very large at all, even at DOS 6.22/7.

Re:The bigger Problem is their "updates"

[Anonymous Coward]

The funny thing is I haven't seen any of this OS decay since XP. I have a Vista PC that I bought in 2007, a Windows 7 PC that I bought in 2011 and my current Windows 8 PC that I bought at the beginning of this year. Each is running from the original install and none of them show any problems with slowdown.

Re:The bigger Problem is their "updates"

[Anonymous Coward]

What is the phone number for the BBS? I would like to call it. Probably still have a copy of Terminate somewhere around too.

Re:The bigger Problem is their "updates"

[DocSavage64109]

That may be so. How long do you think it took DOS to boot up then?

Re:The bigger Problem is their "updates"

[Reziac]

I don't know about if it's kept in RAM (tho it's not that big compared to typical RAM on concurrent machines) but the easy way to keep it tidy is to apply this free tool (which I've been using for 15? years now, and have never seen it screw up):

http://personal.inet.fi/busine...

And for ghu's sakes, defrag. I don't care what Windows says. Defrag. Regularly.

You wouldn't run your car forever without changing the oil, would you? Computers need maintenance too.

As to the wipe and reinstall thing... I consider that the resort of ignorance, like buying a new car instead of changing the oil or brake pads in the old one. I have very old Windows setups (from 3.1 onward), some with over a decade in everyday use, that are still as slick as the day it was installed. How, you ask? I defrag and apply EasyCleaner regularly, and occasionally kill tempfiles. That's it.

Re:The bigger Problem is their "updates"

[toddestan]

Windows XP definitely got slower as the patches and service packs piled up. The original release of Windows XP ran acceptably on a P3 with 256MB of ram, which was a pretty typical computer when XP debuted. By the end, it was a total dog on a P4 with 1GB of ram, which would have have been a high-end machine back in 2001.

Really?

[Anonymous Coward]

This hasn't been a problem in my experience since around Vista, maybe even XP.. I really don't remember.

The only times I see anything remotely resembling this is when, over time, family and friends accumulate so many add-ins, toolbars, and other on-startup junk that there's a noticeable effect. That has nothing to do with OS 'decay': it's the user installing - wittingly or unwittingly - too many resource hungry programs. The solution is simple enough: just uninstall whatever is not actually needed.

Re: Really?

[Anonymous Coward]

This is Dicedot now. You expect stories on a higher level than "Have you tried turning it off and on again"?

Re:Really?

[RandomAdam]

As a counterpoint to this; I had a reasonable machine for work. Win7 Pro, then IT got hold of it and connected it to the new domain etc; now it is much slower. Booting, shutting down, launching programs...everything is slower then the day before.

Re:Really?

[Noah Haders]

That has nothing to do with OS 'decay': it's the user installing - wittingly or unwittingly - too many resource hungry programs. The solution is simple enough: just uninstall whatever is not actually needed.

no, its not quite that simple.

remember defragging hard drives? remember windows registry files? remember the "add or remove programs" control panel? all of these things represent design decisions that cause computers to be cruftier over time.

Re:Really?

[Anonymous Coward]

Real solution, buy a Mac.

Re: Really?

[Anonymous Coward]

inorite

Re:Really?

[RabidReindeer]

Add anti-virus. Have "fast search" caches begin to populate themselves. Other "services" begin to chime in. System maintenance services run. It's a wonder Windows ever actually gets to a usable mode at all.

Re:Really?

[MightyMartian]

I had some HP printer drivers that I couldn't get rid of on a Windows 7 machine, no matter what I did (well, I didn't boot into recovery console and delete the files that way, but that's dangerous territory), so yes, there are ill-behaved applications that can still leave their rotting remnants around the system.

Re:Really?

[Darinbob]

IT is probably a bigger cause of slowdown than Microsoft. Corporate approved spyware stuck on the PCs and Macs, validating that all software running is on the approved list, periodically running audits, etc.

Re:Really?

[Darinbob]

Another problem is even if the application has an uninstall file, if it gets corrupted or lost then it's useless and the standard Microsoft uninstallation will refuse to run. This is not hard to do actually, just installing an application on top of itself will often cause severe confusion if you try to uninstall later.

Re:Really?

[Blaskowicz]

Is that just that your user directory is stored on the network?

Re:Really?

[Frobnicator]

As a counterpoint to this; I had a reasonable machine for work. Win7 Pro, then IT got hold of it and connected it to the new domain etc; now it is much slower. Booting, shutting down, launching programs...everything is slower then the day before.

Well known problem. Once attached to a domain, Windows attempts to do all kinds of stupid things. One of the most common problems is the open/save file dialog. The OS attempts to display it, then blocks until it contacts the domain servers to look up the user's actual name. Then there are similar delays that happen as it goes out and probes each drive, which is a problem if they are mapped network drives as the display waits until everything is built before the UI appears.

On a machine that is disconnected from the domain, perhaps a laptop away from the office, it gets even worse. Internally there is a 45 second delay on each of the network probes, and between Windows 2000 through Windows 7 they all fired sequentially. So if you had your own friendly name plus three mapped drives, that's three minutes of waiting for network connections to time out. It is somewhat faster under Windows 8, but in bad cases can still take ages.

For these specific issues they will not fix the root problems of the shell blocking until after data is loaded or probing the domain for security settings as it would break many shell plugins. It can be made partially better by disabling some of the features; they include disabling certain group policies on shell extensions, turning off certain domain security and SCAPI settings, and disabling drive mappings whenever possible. When disconnected, removing all VPN lookups and disabling proxy detections can also help. Even with those improvements, attaching a machine to a domain introduces an immediate performance penalty on everything shell-related.

Another similar set of problems is apps that try to probe the MRU file list when files are on the network. Many parts of the OS try to cache things based on prior use, and once you're wired in to the corporate network these probes (which stupidly are often blocking tasks) can take seconds to run while on the network, or minutes to run when they time out when off the corporate network.

Re:Really?

[sound vision]

Windows changes a lot of its behavior as soon as you change that radio button from "Workgroup" to "Domain". Before you even get to the log in prompt, Windows has to connect to a domain controller to download Group Policy settings and the like. Having to run login credentials through a DC and load config from it as well will add at least a slight amount of lag - possibly a lot, depending on how responsive the network and the DC are. Then, there's the matter of which Group Policy settings the admins have chosen, many of those can slow down the computer on their own.

If IT "had their way" with it, they might have also loaded it up with antivirus or some other bloatware.

Re:Really?

[Bert64]

The real problem is that the install and/or uninstall is handled by the program itself, and not by a centralised package manager... So every installer is different, there is no consistency, not always a list of what exactly got installed or where, and no guarantee that the uninstall will actually remove everything.
You can also encounter really stupid problems, like the one you describe where you cant uninstall because the uninstall program is damaged. Sometimes you can reinstall over the top and then uninstall, but again due to inconsistency that doesnt always work.

Re:Really?

[matthekc83]

Yeah because now it talks to the network on startup, checks rules call "policies" your sysadmin might set, and talks to the network on shutdown. It makes administration much easier and your network safer but it does slow down your machine...

Re:Really?

[cyborg_monkey]

Bullshit. Plain and simple.

Re:Really?

[Anonymous Coward]

No wonder people want to use their mobile devices in the workplace instead of their PCs. Better user experiences for the management, information workers and data drones is all it takes.

Entropy

[Anonymous Coward]

Doesn't entropy always win?

Re:Entropy

[Anonymous Coward]

Nah, that's authority, no matter how much your fight.

unlikely

[Anonymous Coward]

Microsoft would have to redesign the whole OS to have its apps sandboxes for something like this to work

Re: unlikely

[savuporo]

Or you could actually keep is core along with settings on a dedicated partition, and have a meaningful way of diffing the changes to a stable initial snapshot to identify the reasons for things getting borked.
Trivial with mostly text based /etc, not so with registry, MSI binary databases, metabase and a fuckton of other often proprietary binary formats

Re:unlikely

[mlts]

The ironic thing is that third party companies have been able to repackage Windows programs so only one file is needed to run it. Not an installer... just an executable that provides a virtual environment for the program, and redirects all file and Registry changes to a specific directory in the user's homedir. A couple examples: VMWare ThinApp or Evalaze.

Yes, it takes a bit to create a clean system (VMs are perfect for this with snapshots), pop a "before" run, install the software, then click that it is done. The result is a single file that takes every single change the installer did, and puts it in a sandbox/partition.

If third party companies can do this, why can't MS extend their virtual redirects (which are used with some legacy applications to redirect stuff that would be stored in Program Files to the user's homedir) to include everything the program does? Container functionality is a core part of some other operating systems (RedHat 7), so why not Windows? That way, uninstallation of a program is just tossing the file it is in.

Sandboxes are not new either. I use sandboxie to ensure that what is in my web browser stays in my web browser and doesn't get out. This isn't a 100% solution since an undocumented MS API call would allow a program to "leak" out, but it is usable.

Re:unlikely

[mythosaz]

n.b. I'm currently working doing application virtualization for Windows apps.

The concept of virtualizing apps is fairly straightforward. When you launch a virtual app, you create a virtual file and registry system overlaid on top of the physical system. Any file owned by the app exists in the VFS. Any system file needed by the app exists on the real file system "below" the VFS, and requests for that file fall through. "Good" virtualization can create, start and stop services, etc. This works reasonably well for anything that sits in a folder, has a few registry values, has a self-contained service, but can go south fairly quickly.

The largest practical concern is dependencies. Did you write a .NET app? What version? Do you expect your client to have it? Are you going to put a complete .NET (or VB or whatever) library in your thin app? Are you going to capture your app's install changes on a 32 or 64-bit machine? Separate packages? Does your virtualized app need shell integration? Do you need to right-click on an app and have "Unzip" available in the context menu - and if so, are you depending on a service to present that to the machine -- meaning you can't work as a solid download like ThinApp, but only as a published application through something like App-V.

Does the installer behave PERFECTLY with regard to Window's file structures? Did the programmer make sure to use correct CSIDL locations, or does your virtualization suite protect you from programmers who didn't? Did the installer write your username to one of it's own values? Was the guy who packaged the app good? When he packaged an app that required a specific Java was he smart enough to set the kill bit for other versions of the SSV Helper in IE? Do two of your virtualized programs need to talk to each other or reach into each other's sandboxes? What strategy are you going to take to virtualize 10 different IE plugins or Office Add-Ons?

It can get tricky.

Microsoft's product (App-V) is pretty good, even though they took it from Softgrid. :)

"I avoided Windows 8"... try learning Windows 8

[Anonymous Coward]

http://winsupersite.com/windows-8/windows-8-tip-reset-or-refresh-your-pc

Re:"I avoided Windows 8"... try learning Windows 8

[mlts]

The funny thing is that on machines (ab)used often, backing it up with wbadmin, then nuking it and restoring it from the saved image actually seems to speed things up to the "freshly installed" speed.

Not sure why it helps, be it NTFS issues, actual registry files or other core OS files fragmented, or some other factor, but it does. A backup/format/restore seems to do just as good as a reinstall. It could even be a rootkit that hid itself from the backup utility would not be present when the system was restored, although it doesn't hurt to mount a drive on another box and scan the disk image for nasties just in case.

Application sandboxing

[whereiswaldo]

Like on a modern mobile device, sandbox your apps so they don't clutter the whole system and when they're erased, they're completely gone.

Re: Application sandboxing

[savuporo]

Except that "modern mobile devices" get messed up and bogged down exactly the same way - even if the apps are supposed to be sandboxed.
There is one million os wide settings , or system apps and services that can get screwed up and their internally stored data will start causing issues.
Is the battery drain on your android the same as it was after factory reset ? Didn't think so.

Re:Application sandboxing

[Anonymous Coward]

This is what I do.

Sandboxie is a bloody dream of a program for this.
It has some flaws, some programs still don't install under it. (like League of Legends, thank FUCK for that, up yours friends)

But for making applications separate, easy to clean up, refresh, backup and even make portable, it is glorious.
One thing I tend to do with any program is once I have installed it, I will copy and paste the whole folder in to a _backups folder.
Any time I screw something up, trivial to restore to base. (very useful for modding games for example)

As an added side effect, it also breaks most trialware software.

Re:Application sandboxing

[ayesnymous]

I like to use my work PC for software evaluation purposes. Then after I determine I like a piece of software, I'll install it on my personal computer.

Re: Application sandboxing

[abhi_beckert]

Except that "modern mobile devices" get messed up and bogged down exactly the same way - even if the apps are supposed to be sandboxed.
There is one million os wide settings , or system apps and services that can get screwed up and their internally stored data will start causing issues.
Is the battery drain on your android the same as it was after factory reset ? Didn't think so.

Android doesn't sandbox apps.

iOS does, and it doesn't suffer from this problem. All software is given a directory that they can read from/write to. There are a few places outside that which can be read, but virtually nothing has write access (except for a few cases where a system app will expose access to it's data via inter-app communication. Calendar for example has this).

When you uninstall the app, that directory is deleted. There is no trace at all that the app ever existed.

Re:Application sandboxing

[Zynder]

My Android devices need the app caches flushed out every couple months or the entire phone gets terribly slow. The sandbox has a hole in it or it doesn't work like you think it does. I'm a hardware guy so I have no idea what its issue is.

Re: Application sandboxing

[AuMatar]

Android does sandbox apps. The default internal directory for each app can only be read/written by itself. Prior to version 4.2, the SD card was public and could be read/written by anyone. 4.2 and later, only parts of the SD cared are publicly readable and only parts are publicly writable.

In both cases before and after 4.2 uninstalling will remove the private directory. It will also remove any private directory on the SD card, so long as the app used the default location. Some apps don't, purposely, so their data will persist if reinstalled.

Re: Application sandboxing

[jareth-0205]

Android doesn't sandbox apps.

Er.... Yes it does. Absolutely it does, it's right there in the docs.

http://source.android.com/devi...

Each app gets a separate Linux user, so it's data is separate and inaccessible to other apps.

Re:Application sandboxing

[CastrTroy]

Isn't this the way Metro Apps work? Seems to me like they were already headed down the right path with Windows 8.1 then. You really can't do anything much about old programs wanting to write to arbitrary parts of the disk, because you'll find a lot of applications that just plain won't work. I guess you could trick the application into thinking it's writing to a certain part of the disk when in reality it's just writing to a subdirectory in it's own private folder, but that would create even more problems, when the user decided to save a file, and couldn't find it later because it saved the file inside some virtual folder that only existed for that one application.

Personally I think it's OK if programs have arbitrary file access because it allow apps such as I have on my Surface 2 (RT) to access network drives just as easily as they would access any other file. On Android or iOS, an application has to be specifically coded to access network drives but not so on Windows (or Windows RT).

I think one thing that could be added would be for the OS to keep track of all registry keys edited by an application and be able to remove them after an application is uninstalled. You could possibly do the same for files, but then there would be risk of the user losing data they had created with that application.

Re:Application sandboxing

[blippo]

Is that it? I have needed to reinstall my phones, consistently after a few month although I haven't installed any new apps, etc.
It usually goes to a point where it takes 5-6 seconds for the phone to respond when answering a call - after the last reset,
the performance deteriorated rather soon again. Very annoying.

Re: Application sandboxing

[mlts]

Android doesn't jail() apps (where the application cannot see outside the space it sits in), but it does sandbox apps. Apps get their own UID, and by default, they cannot get into other apps spaces. /system is usually root owned and the whole volume is read-only, for example.

Recent versions of Android use SELinux, so if an app does get access it shouldn't have, it still is stuck in the role it was assigned. For example, some app getting root will still be constrained even with UID 0, so it couldn't remount /system read-write, for example.

Android 4.3 adds onto that by adding SELinux rules onto the external SD card, limiting its use. If you have root, you can use a utility like NextApp SD fix to change SELinux rules back to how they were previously, or SELinuxModeChanger to entirely disable SELinux on your device. Disclaimer: SELinux is a good thing overall, and killing it does weaken security.

iOS's security model is weakened by a jailbreak, while Android's is unaffected if the user has root (assuming the user didn't use the su app to give a rogue app root [1].)

Of course, Android's model has its issues... the all or nothing aspect [2] (where one can choose what stuff an app has access to in iOS), for example.

[1]: Newer apps have a special permission on install which shows the user that it might want root, and the su binary will warn or not allow access to any apps that don't declare that permission in their manifest.

[2]: Cyanogen's privacy features help, as well as XPrivacy. XPrivacy gives extremely fine grained control to what an app can use or cannot use. However, I'd not consider this part of Android proper, though it should be.

Re:Application sandboxing

[Anonymous Coward]

MS tried sandboxing with UAC. They didn't realise that they could have sandboxed individual GUI elements and instead decided to sandbox a single dialog.

"I don't they it means what they think it means."

Re:Application sandboxing

[Zynder]

I have reflashed my phone ROMs several times, but that is because I poke around in places I shouldn't and soft brick them. If clearing all the app caches doesn't help, and you're certain that you don't have a ton of useless background processes running you don't need, check to ensure that you haven't got your system drive almost full. Evidently SSDs slow down the fuller they get. There are several Google results that have quick lists of things that will help if you search for "android sluggish" or similar. The cache flushing solved all of the problems I've had as far as sluggishness and it seems (in my amateur opinion) that the Facebook app is the one hogging all of the cache, or at least the one to fill it up first. If you're a fan of the PC cleaning software CCleaner, they now have an Android version available that will quick flush all the cache for you. The Play store also has a lot of specialized apps for that. Since my phone is rooted, I use Titanium Backup to clean mine since they threw that in there as an extra function but I have used CCleaner as well. Can't comment on any other cache app.

Re: Application sandboxing

[stoploss]

XPrivacy gives extremely fine grained control to what an app can use or cannot use. However, I'd not consider this part of Android proper, though it should be.

Just another plug for XPrivacy. It's excellent. I use it in combination with AFWall to whitelist which apps can get through the Android iptables firewall. XPrivacy will handle network access control, but I'm a belt and suspenders kind of guy.

Re:Application sandboxing

[Shikaku]

Try running ccleaner on your android phone. Yep its on android too but running it after it got really slow helped a lot.

Re: Application sandboxing

[symbolset]

And in the transition we lost the ability to move apps to SD, which sucked on limited devices.

Re: Application sandboxing

[paulatz]

You are right to some extent: There is a tradeoff. A strict sandboxing will prevent many useful features; a lax sandboxing will not be completely effective

Yet, even a lax sandboxing can be extremely useful. In an Android phone it is relatively easy to keep track of which apps are using a lot of battery, and you can uninstall them from the same screen, this is possible thanks to sandboxing keeping track of where every system call is coming from. If you decide to give up and restart from scratch, it only take 5 minutes to erase all user data, and you have a reborn phone; eventually add 15 minutes to copy your pictures back, if you really want to. Compare this to the afternoon of cursing it takes to reinstall windows and all the programs, redo all the updates, restore the backup. find out that some stuff was not backed up because it as stored in hidden directories scattered around.

Re:Application sandboxing

[AmiMoJo]

You really can't do anything much about old programs wanting to write to arbitrary parts of the disk, because you'll find a lot of applications that just plain won't work. I guess you could trick the application into thinking it's writing to a certain part of the disk when in reality it's just writing to a subdirectory in it's own private folder, but that would create even more problems, when the user decided to save a file, and couldn't find it later because it saved the file inside some virtual folder that only existed for that one application.

Actually that's exactly what Windows has been doing since Vista. Both the filesystem and registry became virtualized, so apps couldn't shit all over them any more. Apps that attempt to write into virtualized directories are transparently redirected to safe locations.

To be fair it did break some stuff, but not much. The main issue was that it was slow, but they fixed that in Windows 7.

Re: Application sandboxing

[StikyPad]

iOS's security model is weakened by a jailbreak

Not true.

https://news.ycombinator.com/i...

Re: Application sandboxing

[ThatsNotPudding]

Each app gets a separate Linux user, so it's data is separate and inaccessible to other apps.

Ignoring the endless screens of permission requests with each update, including your mothers maiden name and bra size. But yeah, sandboxes are great.

Re:Application sandboxing

[Anonymous Coward]

I remember doing this on the computers at school. lol. Lots of risk with the amount of things and where they came from. I think I may have (inadvertently, but carelessly) caused some issues.

Re: Application sandboxing

[MobyDisk]

Then why are protestors in Hong Kong only infected by the government malware if they are using Jailbroken iPhones? I don't really know much about how an iPHone jailbreak works, but the consensus seems to be that it somehow compromises security. I'm unclear how though.
http://apple.slashdot.org/stor...

Re: Application sandboxing

[StikyPad]

Don't confuse widespread misinformation with consensus of knowledgeable people.

If they enabled SSH and didn't change the default password, then that's a possibility, but a JB doesn't install SSHd by default, nor does it increase the attack surface. The post I linked to explains all of that, though, so it's easier to just reference that than reiterate it all point-for-point here.

Re: Application sandboxing

[savuporo]

Re-imaging any version of windows on the install partition is possible within minutes, if you bothered to prepare for that. System Restore, WAIK, ADK etc, if this was actually important.
And its a ton easier on any unixy-box. And guess what, all this is even easier for a homogeneous hardware pool like a particular cell phone model, or a particular OEM PC model, with a preconfigured image that matches your hardware exactly - for a random home PC thats more work.

Re: Application sandboxing

[paulatz]

And its a ton easier on any unixy-box. And guess what, all this is even easier for a homogeneous hardware pool like a particular cell phone model, or a particular OEM PC model, with a preconfigured image that matches your hardware exactly - for a random home PC thats more work.

I don't agree on this. On a Linux box, if you used separate partition, it's as easy as save down a list of installed rpms (or deb), reinstall os, reinstall list of rpm. On OEM windows installation you normally only have a recovery partition that can only do automatic repair (that never works) or destroy everything and restore to the factory state.

I'm not talking about restoring the factory state, I'm talking about restoring your PC to a working state, with all your software and data as before but not fucked up

Re:Application sandboxing

[Anonymous Coward]

Ah, a tried and true method.

Antecdotes != Evidence

[Galaga88]

1. Is there any actual proof that OS decay is still a thing? I'm running Windows 8.1 that was upgraded from a Windows 7 install that was put on years ago, and I've seen zero performance issues.
2. Shouldn't the person asking this question have actually used Windows 8 before asking if Windows 10 will "finally" fix a problem that may or may not even exist?

Re:Antecdotes != Evidence

[Anonymous Coward]

> Is there any actual proof that OS decay is still a thing?

It's not. I stopped having to wipe my system every year the day I switched to Win 2k from Win 98SE. That was ~14 years ago. Since then, it's been one OS install until it's time for a major hardware upgrade.

Re:Antecdotes != Evidence

[ianbnet]

1. Is there any actual proof that OS decay is still a thing? I'm running Windows 8.1 that was upgraded from a Windows 7 install that was put on years ago, and I've seen zero performance issues.
2. Shouldn't the person asking this question have actually used Windows 8 before asking if Windows 10 will "finally" fix a problem that may or may not even exist?

Amen brother. I've also seen zero decay on my gaming rig, which has been through the Windows 7 --> 8 --> 8.1 etc without ever a refresh...

Re:Antecdotes != Evidence

[MrHanky]

My parents both use Windows 7, and both computers have slowed to beyond frustration. Perhaps the reason why yours still works is that you learnt how to use Windows. Most people never do.

Re:Antecdotes != Evidence

[Anonymous Coward]

Yeah... I use MSCONFIG to make sure NOTHING loads on startup, disable services I don't need and don't allow apps that run as a service, and my Windows box runs as fast as the day I bought it. Ditto for regular drive cleaning and defragging. If you are vigilant you CAN keep your system lean and mean.

Or just use Linux, I've never spent a single second doing any performance maintenance on any of my Linux boxes and they never seem to bog. Of course, until I can play BF4 on Linux and on ULTRA my one and only gaming rig will be running windows. So forever, probably.

Re:Antecdotes != Evidence

[exomondo]

I definitely remember it from the old Windows 9x days but not since the move to NT.

Re:Antecdotes != Evidence

[Anonymous Coward]

> 'm running Windows 8.1 that was upgraded from a Windows 7 install that was put on years ago, and I've seen zero performance issues.
Use it once in a while then.

Re:Antecdotes != Evidence

[Anonymous Coward]

> Amen brother. I've also seen zero decay on my gaming rig, which has been through the Windows 7 --> 8 --> 8.1 etc without ever a refresh...

Good for you. By the way what the fuck is a gaming rig? Do you mean you have a computer and bought some shitty stuff with blue LEDs to plug into it or did you install a computer in a truck?

It's still a fucking computer.

Re:Antecdotes != Evidence

[PopeRatzo]

Likewise. I've got a Windows 7 gaming rig that has seen LOTS of installs and uninstalls, driver updates and Windows updates and have seen zero performance reduction.

Maybe this was a thing in Windows 95, but I'm not sure it's a thing now as long as you're not getting infected with malware.

Re:Antecdotes != Evidence

[Hadlock]

SSDs help improve performance pretty dramatically, and also nullifies fragmented disk issues (which mostly disappeared when Microsoft started scheduling regular defrags of the drive in Windows version * )

Re:Antecdotes != Evidence

[Dutch Gun]

I used to have to do a clean install of Windows every few years to keep things performing well, but I don't recall doing that since the switch to NT-based systems (starting with 2000 for me personally). For users that keep installing malware/adware/spyware on their systems, it seems entirely likely that they'd have to do a clean re-install to get rid of all the cruft every once in a while. Some of that stuff is pretty hard to remove, and can really cause issues with system stability and performance.

When people talk about "OS decay", they're probably dealing with systems that have either a huge amount of software churn, a lot of crapware, or very often both. It's not so much about "learning how to use Windows 7" so much as not installing free, sketchy utilities that contain system-hogging spyware. Or perhaps it's better termed "learning not to abuse your operating system". People do the same sort of nonsense with their phones - install dozens of apps that all want do stay resident for whatever reason, and then they wonder where the battery life went. Same deal - if you give people the freedom to customize their device, some people will inevitably make bad choices.

I don't know if this applies to you parents or not, but I've certainly seen plenty of cringe-inducing systems for people to know just enough to be dangerous. My parent don't know enough to really do anything of consequence on their computer other than check e-mail, surf the net, and play solitaire, so their system (Windows XP) has stayed nice and tidy for the last seven or eight years (I think) they've had that machine.

Re:Antecdotes != Evidence

[Anonymous Coward]

It existed as of Windows 7.

Time at boot, for both Windows 7 and Ubuntu at the time was similar to a desktop, and for the hard drive to stop thrashing at about 90 seconds.

A year later, with both fully patched, Ubuntu was still 90 seconds. Windows 7 was about 600 seconds. Usage was pretty mixed between the two. Programming or serious work had Ubuntu booted, games had Windows booted.

Windows had the outer tracks on the disk, so should have been faster as well. I do have the times from a run. Oh, and no AV or anything constantly scanning to slow Windows down.

I have not had a clean install of both to test for Windows 8. I do have the timed numbers. Actually, some of the tests were observed by someone who was formerly one of Microsoft's project managers on Windows.

Re:Antecdotes != Evidence

[RyuuzakiTetsuya]

Get a spare disk and do a clean Windows 8.1 install on it and bench mark it compared to the ages old install of Windows.

This problem hits OSX too, and always makes me lose my Smug Mac User grin. It's even worse in OSX because the filesystem, HFS+, is absolutely bonkers about cruft that builds up. HFS+ is worse than Panic at the Disco. HFS+ is almost worse than scabies.

Re:Antecdotes != Evidence

[rahvin112]

How would you even know if you are affected. By your own admission you've not run anything other that what you have installed. You have NOTHING to compare it against. It could be 10 times slower than it should be an you wouldn't know because you haven't done a clean install. This is common knowledge to those that do routine format/reinstall. In the enthusiast category of user it's not uncommon to have this be a standard twice a year operation. If you want to verify it, do a format reinstall and compare the speed. You will find your computer feels new because it's so much faster and responsive.

Re:Antecdotes != Evidence

[Anonymous Coward]

You have a ton of crap in your registry I bet. The registry was the worst idea ever.

Re:Antecdotes != Evidence

[simplypeachy]

Why don't you diagnose the fault and actually help them, instead of siding on the side of bullshit? If the slow-down is that bad it should be pretty obvious to someone with some technical skill.

Re:Antecdotes != Evidence

[simplypeachy]

"You are welcome on my lawn."

Always wondered about people with fences around their property. Seems unnecessarily antisocial to me.

Re:Antecdotes != Evidence

[angel'o'sphere]

My ten year old Macs are as fast as they where when I bought them.
If you have speed issues on a Mac you likely caught one of the rare viruses or trojans!
That whole 'degrading of an OS' thing does not occur on unix based systems!

Re:Antecdotes != Evidence

[Russ1642]

I've only heard about it from friends and family. When they explain that their computer has 'slowed down these days' I can usually trace it to a single piece of shit software they've installed. Computers don't just get slower and slower. Something specific has to cause the slowdown and you can actually fix it.

Re:Antecdotes != Evidence

[TapeCutter]

I had a win98 install that lasted almost a decade but it took a fair bit of effort to hold back the entropy (I resisted upgrading because of one win98 game...). NT was definitely a landmark in stability as was XP. I've been on Win7 at home and work for a couple of years now. I've been doing C/C++ development regularly for the last 25ys as a job, a bad pointer would often bring Win98 to it's knees, not so much with NT/XP and I've never managed to crash Win7 with bad code. In fact Win7 has crashed on me exactly twice, once when the SSD died, and another time when the video card started smoking.

From the POV of system stability you could line up today's popular O/S's, throw a dart blindfolded, and be still be sure to hit a decent general purpose O/S. This doesn't mean they are flawlessly designed, however most of the bitching I see from geeks is just the geek not understanding how things work before attempting to "fix" the "problem", and most of the bitching I hear from non-geeks is about the non-geek's ignorance of how malware got onto their machine. .

Re: Antecdotes != Evidence

[jd]

Agreed about anecdotes. However, I can say that I have to reboot my Windows 7 PC weekly because of serious degradation in performance. I have installed a fair bit of software (the PATH can no longer be extended) but there's only about three games (Freeciv, Kerbal Space Program, Elite: Dangerous) and no apps, toolbars or junk. The rest of the software on there? MariaDB, Ingres, GRASS, QGIS (OSGEO is basically Cygwin, so I've now three incompatible Cygwin distros on Windows), HOL 4, Active Python, Active Perl, Erlang, Rust, Blender, PoVRay, BMRT - the sort of stuff you'd expect to find on any PC, nothing fancy.

And Netscape. Which is a horrible resource hog and is honestly not usable in its current form. I have abandoned all efforts to get Chrome usable. I'll probably deinstall both and switch to Amaya. Which barely does anything, but it does it tolerably.

Re:Antecdotes != Evidence

[Anonymous Coward]

Helps to turn off any and all folder animations etc. too.

Re:Antecdotes != Evidence

[exomondo]

The only crashes I've had with Windows 7 or 8 were from bad graphics card drivers causing a BSOD.

Re:Antecdotes != Evidence

[QuantumLeaper]

How old of an OS are you running? Since my friend who LOVES Macs, hated it when he upgraded the OS, the last time, it went from usable to a snail. He still used the Mac but downgraded the OS to the old version, he now uses a Win8.1 PC to play his games on, it runs a lot faster. And it only a few months newer than the Mac. I believe it the same CPU and similar video card. I had to fix his PC a few months back, his video card died, that's the only reason I know they both have the same CPU, and similar video cards, I told he should upgrade the video card, but he was a cheap skate.

Re:Antecdotes != Evidence

[Anonymous Coward]

The issue used to be caused by updates, etc that caused the hard drive to seek more, rather than everything being close to each other in the original install.

The "problem" appeared to dramatically improve around windows 7/8 simply because SSDs (where seek is much less of a problem) started trickling down to more users.

Re: Antecdotes != Evidence

[linuxrocks123]

Re Netscape:

What the hell? Dude, Netscape's been dead for 6 years. The browser is completely unmaintained and everything past 7.2 was a crappy respin of Firefox anyway. All that's left of the brand is a crappy web portal: http://netscape.aol.com/

I liked it, too. I used Netscape 7.2 long past its sell-by date. But why would you possibly be using Netscape in 2014?

Re:Antecdotes != Evidence

[loom_weaver]

Most of my changes are all the automatic security updates and OS patches for Microsoft products (Office, VS, .NET, etc.)

I could avoid OS decay if I turned them off but should I?

Re:Antecdotes != Evidence

[gargleblast]

Invite yourself around, sit down at their machines, right click on the task bar and select 'Start Task Manager.' Look over particularly the Processes tab and its CPU and memory columns, also the Performance tab and its Resource Monitor button. Take some notes, and think a while. You might earn yourself a free dinner.

Re:Antecdotes != Evidence

[lgw]

Depending on where you live, anti-social can be a strong survival trait.

Re:Antecdotes != Evidence

[Barlo_Mung_42]

My main dev box at work started as a Vista system. Upgraded through Win7 and now it's been upgraded to Win8.1. It only feels slow compared to my much newer system at home. It doesn't take any longer to load files or reboot than it ever did. In fact it's only gotten faster as I've upgraded.
What I don't do is install a bunch of crap on it that's going to run in the background. At home it's much the same only I run most of the systems as a standard (non-admin) user. Want to stop pretty much all drive by installs in their tracks? Run as a standard user.

Re:Antecdotes != Evidence

[LordLimecat]

The problem is people install applications which offer to install a startup widget.

Thats not a technical problem that can be solved with technical means unless you mean to remove the end user's ability to set up startup applications.

Re:Antecdotes != Evidence

[LordLimecat]

That whole 'degrading of an OS' thing does not occur on unix based systems!

Thats the most handwavy OS-elitist thing I've ever heard.

There is nothing about "unix-based systems" as opposed to NT-based or Linux-based that makes it not "get slower over time". Getting slower over time could be caused by one of several factors:
  * Fragmentation (on non-SSDs).
  * Buggy drivers
  * Viruses
  * Unneeded applications running at startup

The first one is an aspect of filesystems, not OSes (OSX doesnt use a "unix-based" filesystem, it uses HFS+ which AFAIK is proprietary Apple). It could be argued that NTFS doesnt do as good of a job as preventing fragmentation as other filesystems, but it really hasnt been an issue in about 7 years. Im not sure about HFS+, but my understanding is that the FS is quite old at this point and I wouldnt be surprised if it had its share of issues.

The next two can affect any OS; buggy drivers tend to affect microkernels and hybrid kernels like NT more than monolithic kernels, but "Unixy" OSes have had their share of buggy driver problems (I've seen FreeBSD 8.3 boot hangs due to dell USB keyboards, for example). OSX itself has had a number of buggy drivers recently (or so it seems-- with the hardware issues in iOS 8.x). Viruses absolutely affect OSX, and Unix offers no protection here.

That just leaves startup apps, and as I recall that was a major feature added to Mountain Lion. And certainly as a userland feature, startup applications have zero to do with the kernel on the backend.

The whole "its unix and therefore magical" thing makes really good marketing material, but isnt based in anything resembling reality.

Re:Antecdotes != Evidence

[Anonymous Coward]

Certainly adware, malware crap. Ignorant users click yes on everything then complain. Running Windows 8/8.1 on 6 machines with not a single issue. All humming with no performance degradation

Re:Antecdotes != Evidence

[thegarbz]

I'm inclined to believe it's the type of software being churned that would make a difference. Games are a classic example of things typically done right. A folder on the HDD with most of the stuff. A folder in the user profile to store the saves and settings, and a folder in the registry for the DRM shit. And most of the invasive stuff to the system (DRM, Battle.NET, Origin etc) are shared between games and thus not churned.

Now compare that with say changing antivirus software vendors, and filling up the computer with spyware and crapware from yet another worthless program downloaded from CNET and you may have a very different take on the whole idea.

Re:Antecdotes != Evidence

[Anonymous Coward]

Amen brother. I've also seen zero decay on my gaming rig, which has been through the Windows 7 --> 8 --> 8.1 etc without ever a refresh...

You do realise that means you, by definition, have no point of reference. "My system is not slower than my old Windows 7 install" is not the same as "My system is not slower a clean installation".

Re:Antecdotes != Evidence

[ArsonSmith]

Non social or Unsocial. Antisocial is picking people off with a rifle from your front door as they walk/drive by.

Re:Antecdotes != Evidence

[ArsonSmith]

I think that's exactly what this article is talking about. That single piece of shit software being some version of Windows that was either pre-installed or that they installed.

Re:Antecdotes != Evidence

[Anonymous Coward]

Hmm.. well... Maybe there is an chance...

More and more games get ported to SteamOS/Linux..

Latest are Borderlands 2 and Baldur's Gate Enhanced Edition (Baldur's gate is comming soon).
Others are Metro: Last Light, The Witcher 2, X-Com Enemy Unknown and sequel, Legends of Grimrock, ShadowRun and more..

So - maybe not probable, but possible.. I guess so..

Re:Antecdotes != Evidence

[Anonymous Coward]

Whoops - I ment to say "Baldur's gate 2 is comming soon" sorry...

Re:Antecdotes != Evidence

[Anonymous Coward]

It is funny how you dismiss anecdotes and ask for evidence, and then present your own anecdote as an evidence...

Re:Antecdotes != Evidence

[Anonymous Coward]

Unfortunately, it's hard to avoid crap-ware. I helped a lady recently that complained her computer had gotten slow. Turned out it was bundled crap that came with a printer. This is a big problem Windows have, and there's no easy fix except preventing people from installing stuff.

Ideally, Windows apps should be sandboxed. If they want out of their sandbox, users should be allowed to both deny and give rights without the app knowing, so that you can't write an app that nags you until you give in.

Re:Antecdotes != Evidence

[Anonymous Coward]

I have used all windows systems since 2.0. As in the windows did not overlap. On NT kernel based systems you can screw it up with bad extensions (drivers, window hooks, anti virus, etc.) like install something and delete files. But you DO NOT have to reinstall ever. You must maintain it. Often this involves clearing the registry of the offending pointer to bogus files, deleted paths, etc. There are apps to do it, I do it manually since I am a Windows weenie. Reinstall is just the nuclear bomb version of resetting bad config. If you know how to use WinDbg and ProcessMonitor you can debug the problems very quickly and fix them, but it requires significant skill.

The best thing to do is not install crappy software... avoid the problem in the first place.

Re:Antecdotes != Evidence

[paziek]

Maybe you just didn't notice that as you update on regular basis. I had Windows 7 that didn't see any update for like 2-3 years and then I decided to do it. I had no idea about this decay issue until I told my coworker about how much my system slowed down after huge upgrade.

Re:Antecdotes != Evidence

[Anonymous Coward]

That's what you think! Try wiping and installing fresh, it'll be like a new machine.

Re:Antecdotes != Evidence

[guacamole]

IMO, the biggest issue is not the OS, but the third party apps. Think of the dozens of apps that the users installed over the life time of a PC. A lot of them will run scripts and services on start up and shut down. Many processes will hang in memory: automatic software updaters, license managers, useless tray apps, and daemons that may "call home" the company. Also remember about software that's either borderline or outright spyware or malware, such as the infamous Search Protect garbage by conduit, that's highly likely to be installed at some point on your PC if you download lots of freeware apps and you're not careful.

Re:Antecdotes != Evidence

[Barsteward]

" For users that keep installing malware/adware/spyware on their systems, it seems entirely likely that they'd have to do a clean re-install to get rid of all the cruft every once in a while." - you can't blame the ordinary non-tech users for that. if they use duff browsers then those things are installed silently plus if they don't read an install screen correctly they won't see "installl xxx toolbar" is ticked as a default.

Re:Antecdotes != Evidence

[Anonymous Coward]

Windows does have two problems that at contribute to the system getting slower.

One is the registry. While a forgotten file in a forgotten directory doesn't take time, the registry is this huge blob that needs to be loaded. And everything forgets to erase things from the registry, or leaves them there deliberately in the hope that you will install the program again.

The other is that everybody gets to write their own installer. While Microsoft did finally create MSI, far from everyone is using it, and those custom installers do leave things all over the system. On a Linux system, you either stick to the system package manager, and nothing gets lost, or you install into a single directory under /opt or /usr/local. When that directory is deleted, everything in it is gone. The exception being installing from source, unless you do that via the package manager.

And the combination of the two, even though a program is installed through MSI, it can still write all over the registry, and the system won't know about it.

The solution would be that the software needs to inform the installer which folders and registry keys it needs, and have the system enforce that it can only access those. I.e. a form of sandboxing. Linux doesn't have this either, it is more of an honor system, where everybody follows the rules.

Re:Antecdotes != Evidence

[Anonymous Coward]

Very well said. I've not experienced 'OS decay' since the 9x days (and that was more related to disc fragmentation than anything else) as I tend to keep an extremely tight ship, software-wise. The people whose machines I see who complain about slowdown generally have helper apps for a couple of printers, all the update utilities that come with their machine, some awful tsr that does something related to media playbace, itunes, and a whole host of other stuff that I wouldn't touch with a 20-foot bargepole.

Sometimes the problem is in layer 8 rather than the tech.

Re:Antecdotes != Evidence

[Bert64]

It's not that unix is magical, but there are several very important differences that make unix systems far less susceptible to these problems than windows...

1, The biggest difference is probably the use of package management on unix vs arbitrary binary installers on windows... with a package manager, every install, update and uninstall is controlled by the same process which keeps track of what got installed and is able to cleanly remove it again, with windows an "installer" is just a binary program that you are trusting to write files all over the place but you have no real idea what its doing or if its working correctly. With the package manager its very easy to identify what package installed any given file etc. If you go outside of the package manager on unix and try to overwrite system packages by hand you can have serious problems too.

2, Transparency - Unix systems are much simpler and better understood, the boot process is usually just a series of scripts for instance, the filesystem is laid out in a mostly logical hierarchy and most configuration is stored in individual human readable text files, its much easier to understand exactly whats going on and much more difficult for poorly written programs to hide performance crippling cruft in unexpected places.

3, Lack of third party drivers - on most unix systems, drivers typically ship with the OS, get updated when the OS does and get tested together... Windows systems typically have a random collection of disparate drivers which sometimes don't play well together or with updates to other parts of the system. The other problems mentioned above also apply to drivers as well as userland.

Re:Antecdotes != Evidence

[bemymonkey]

Agreed. I'm still using my Win7 SP1 installs, and the last time they were reinstalled was when the MSDN SP1-integrated Win7 DVDs came out... still running like champs.

Re:Antecdotes != Evidence

[paulatz]

Of course, until I can play BF4 on Linux and on ULTRA my one and only gaming rig will be running windows. So forever, probably.

It will most likely take around 7 years and a child, then you won't be able to run BF4 on Linux nor on Windows because of lack of time.

Re:Antecdotes != Evidence

[Bert64]

Many applications install the startup widget without asking, and often without informing the user at all. It could be solved through the use of a proper package manager and standard package format so that the package manager rather than an arbitrary installer program actually controls the installation and always gives the user the choice.

Re:Antecdotes != Evidence

[Dutch Gun]

I agree. It's definitely hard to avoid, and it's a huge problem on Windows. Even installing software from "reputable" companies can get you into trouble. A lot of companies make shady deals to promote various products (like the Ask toolbar installing with Java). I installed a new HP printer, and was prompted to install a shit-load of various 3rd party applications, NONE of which were necessary to get the device to print, but does who-knows-what to your system. There was no option I could discern for installing the more "necessary" bits separately (like functionality for controlling higher-level printer functions, such as cleaning the heads), so I just ended up using the default (and completely functional) basic driver available in Windows. I actually have to pull out my printer heads and clean them manually from time to time, though, which is sort of sad.

While it's true users do install a lot of sketchy stuff, it's hard to blame people for everything that can go wrong. In fairness, even when legitimate companies are doing this, it's really hard for the average user to know they need to be wary until they get bitten a few times. Over the years, I've had to become hypersensitive about installing software just to keep my system crapware free. It's sad when you can't even trust the discs you get with new hardware purchases.

Re:Antecdotes != Evidence

[Anonymous Coward]

It was a thing in XP.

Not since then, unless you install a LOT of dodgy apps that shit all over.

Re:Antecdotes != Evidence

[Anonymous Coward]

I'm not a Windows user, but their file systems still fragment?

Re:Antecdotes != Evidence

[simplypeachy]

I got a leaflet through the door this week from our local neighbourhood watch. It said that the police's attention was on sorting out the dog poo and rowdy behaviour from people leaving one of the pubs. I live in such a lame area and I'm genuinely grateful for it.

Re:Antecdotes != Evidence

[SimonTheSoundMan]

I got very paranoid about my Mac slowing down. I tested all sorts before doing a full wipe.

Before wipe, booted in 38 seconds, tested some apps load time, I can't remember off my head but it let's say 10 seconds. Wipe, machine took 35 seconds to boot, apps took the same time to open.

Meanwhile, two Windows machines I have with Windows 7. These machines do not get used much. Before wipe took 3 minutes to login screen and a good 2 minutes after login to be usable (don't get this in OS X), wipe the machines took 50 seconds to boot and were usable after login. I've been to businesses where their quad core i7 machines get switch on in the morning as soon as staff arrive at their desk, they then go make a coffee, eat breakfast and come back and the machine would just about be booted. Meanwhile I'm there with an iMac from 5 years ago that has never been wiped and it boots in 30 seconds.

Now on a retina MBP. Takes 11 seconds from boot to login, most of that time is EFI doing POST etc.

Re:Antecdotes != Evidence

[AmiMoJo]

Macs get it just as badly, as does every OS that allows programs to install things that hook into the OS or run in the background. For example, if you install an app that lets the Finder show thumbnails of some image format it previously could not display you have just added extra code that needs to be loaded, and extra RAM use to store it. Your computer got a bit slower. Similarly if you install an app to show some extra data on your desktop, now it is running when you boot your computer (longer boot time) and use resources (RAM, CPU time).

My friend installed Norton on his Mac years ago, it was crippled instantly. He removed it eventually and everything was fine again. There is nothing magical about Mac OS.

Re:Antecdotes != Evidence

[LordLimecat]

with windows an "installer" is just a binary program that you are trusting to write files all over the place but you have no real idea what its doing or if its working correctly.

This is not correct in the least.

Windows Installer programs (.msi), which are the backend for many programs, IS structured and is not "just a binary". Its not quite the same as a package manager like apt-get, but shares more in common with that than with "just a binary" able to "write files all over the place".

When the Installer service isnt used, there are a number of other options like InstallShield, which again are structured, not "just binaries". It is true that you can download a binary that doesnt use any installer and does just randomly write, but thats true of ANY non-walled garden OS. Notably, for the longest time (and probably still now) the best way to get good nVidia performance on Linux was to download a large binary blob driver from nVidia. It did not use a package manager, it simply executed via a shell script.

If you go outside of the package manager on unix and try to overwrite system packages by hand you can have serious problems too.

But it is sometimes necessary, and it isnt a feature of "unix". Windows has plenty of package managers, they just arent popular for a number of reasons that have nothing to do with its heritage as an NT-based OS and everything to do with the culture around the respective OSes and the common license model for software on those OSes. Most software people use on Windows is NOT OSS, and thus would be utterly unsuitable for a repository.

2, Transparency - Unix systems are much simpler and better understood, the boot process is usually just a series of scripts for instance,

Thats also not true. The "better understood" could perhaps apply to the FOSS *nixes, but not OSX. Series of scripts is just plain wrong; every OS follows a pretty standard progression of "boot record --> bootloader --> OS / drivers". Scripts dont kick in until after that; prior is a lot of "binary stuff". The filesystem structure you mention is also there on Windows, it just looks different; programs in Program Files, Window bits in the Windows folder, user data in the Users folder (and programdata, when a program is badly behaved and tries to stick userdata in program files). Im aware of the Unix heirarchy, but youre being deceptive if you claim that programs dont violate it all the time. Binaries and config stored in /usr/local, /usr/bin, /opt, /var (!), logs stored outside of /var, and so on-- I've seen it all.

3, Lack of third party drivers - on most unix systems, drivers typically ship with the OS, get updated when the OS does and get tested together

This is vaguely and sort of true. The kernel types are different; you might as well complain that a pencil isnt as well bristled as a paintbrush. And certainly there ARE driver problems on Linux, even when they're built in-- and as theyre built in theyre often far more of a PITA when they do break (like the old e1000 adapter bug that bricked the NICs).

Windows systems typically have a random collection of disparate drivers which sometimes don't play well together or with updates to other parts of the system.

This is very rare. If you got your system from an OEM, they provide the drivers (packaged from the hardware vendors) and generally there are no issues-- it works out of the box. If you build your own, you just grab the motherboard and video drivers and you're set. Actual conflicts are not common at all; the closest I've seen is buggy webcam drivers from logitech causing bluescreen, and that not for several years now. Since then I've heard of iOS bugs breaking wifi / mobile data, Linux bugs bricking e1000 NICs, FreeBSD USB driver bugs causing hang on boot (8.

Re:Antecdotes != Evidence

[PopeRatzo]

Non social or Unsocial. Antisocial is picking people off with a rifle from your front door as they walk/drive by.

In the US, we call that "home security".

Re:Antecdotes != Evidence

[SuiteSisterMary]

(I resisted upgrading because of one win98 game...).

Dammit, man, which game?!

Re:Antecdotes != Evidence

[Anonymous Coward]

Well if we're going to talk anecdotes, I have a Vista computer (yes, I know) that isn't all that old. It doesn't run much software and upon starting up (or waking up) the disk thrashes violently to the point of slowing down my browser (Chrome, FireFox, IE, it doesn't matter). I have yet to work on a WIndows box that didn't experience the decay mentioned in the original article. I have Linux and OSX boxes that are older, more used and don't slow down.

Truthfully, Vista isn't that bad of a version of Windows except for the decay, but that's a Windows thing not a Vista thing. To me it's a fact of life that one has to deal with because I know, regardless of the target OS i program to, my company will make me work on a Windows box because of Office.

Re:Antecdotes != Evidence

[Anonymous Coward]

My parents both use Windows 7, and both computers have slowed to beyond frustration. Perhaps the reason why yours still works is that you learnt how to use Windows. Most people never do.

Which supports the idea that "OS decay" is really mismanagement on the user's part, not some intrinsic "feature" of the OS.

Re:Antecdotes != Evidence

[angel'o'sphere]

The "magic" is: you simply remove what you have installed. And thats it. On windows that is more or less impossible. I mean: who came to the idea that you need a "deinstaller"?

Your examples make sense, ofc additional functionality costs, however on windows random crap costs suddenly speed ... your boot time example e.g. is wrong.

Such an desktop app would be started after boot, in the user context, and the Mac would be useable and responsive already even with a back ground app starting to display something on the desktop.

BTW: who made this stupid "subject" :D Anecdotes are Evidence just like any other evidence is ... anyway, another topic.

Re:Antecdotes != Evidence

[operagost]

Do you let people use your computer and dig through all your files?

Re:Antecdotes != Evidence

[AmiMoJo]

The "magic" is: you simply remove what you have installed. And thats it. On windows that is more or less impossible.

No, it's easy. Since Vista all apps have been using a virtualized registry and filesystem. Each app is segregated into it's own little area, even if it thinks it can write files and registry keys all over the place.

Re:Antecdotes != Evidence

[RyuuzakiTetsuya]

Going from boot to login is one metric, but overall usability is another.

http://arstechnica.com/staff/2...

Hfs+ is a fucking dog. John Siracusa isn't some Johnny come lately to the Mac world. He's been one of the faithful since well into the classic days.

Re:Antecdotes != Evidence

[angel'o'sphere]

Ah, interesting. Never noticed that,
But why is it still slowing down over time, and Windows 7 and 8 seem to have the same problem?

Re:Antecdotes != Evidence

[LordLimecat]

One is the registry. While a forgotten file in a forgotten directory doesn't take time, the registry is this huge blob that needs to be loaded.

The entire registry usually does not exceed 30MB, and the entire thing is not loaded at once. It is loaded piecemeal when used. You might as well argue that the fact that there are 80 zillion plist files on OSX causes it to get slower over time (maybe it does, but only if OSX were really badly engineered).

The other is that everybody gets to write their own installer.

Most dont, there are a handful of popular ones including the built in one. And lets face it, if you're running someone elses code on your box they could technically write their own installer no matter what OS you're on. Conventions may vary due to other factors, but the "Unixness" of an OS has no impact on your ability or lack thereof to write an installer.

And the combination of the two, even though a program is installed through MSI, it can still write all over the registry, and the system won't know about it.

MSI files describe what will be modified and when; thats their job, and how they can perform atomic installations. Any binary you run on any OS is generally going to be able to write anywhere, in the absence of a Mandatory Access Control system.

The solution would be that the software needs to inform the installer which folders and registry keys it needs,

It does.

Re:Antecdotes != Evidence

[angel'o'sphere]

Fragmentation is a Windows myth.

Installed files belonging to the OS, never get moved or get changed. They don't "fragment". They are always on the exact same hard drive sectors.

User files are a different story. But modern OSes since NT times reallocate the whole file anew when it is written again. That means if you open an document and change it and write it again, it shows up elsewhere on the disk as a consecutive list of blocks.

The whole "its unix and therefore magical" thing makes really good marketing material, but isnt based in anything resembling reality.
No one said anything about magic. Point is, a 20 year old unix system is a fast as it was when it got installed. A 20 year old windows system won't even boot due to y2k bugs. And a 10 year old windows system is: for what ever reason unusable now.

No idea why you want to blame that on "bad drivers" etc. when that is obviously an inherent problem of the platform.

Re:Antecdotes != Evidence

[angel'o'sphere]

Thats also not true. The "better understood" could perhaps apply to the FOSS *nixes, but not OSX. Series of scripts is just plain wrong; every OS follows a pretty standard progression of "boot record --> bootloader --> OS / drivers". Scripts dont kick in until after that; prior is a lot of "binary stuff".
That is complete bollocks.
After the kernel is up, there is no distinction between "scripts" and "binaries" at all. After all a script is just text loaded and executed by a binary, wow, that was difficult to grasp.

because I work with both types of systems (Linux, FreeBSD, OSX, Windows XP/7/8) and I've seen these problems crop up on all of them.
If you can manage to get a Linux/OS X or other *nix system to behave like windows, then you don't qualify for this discussion and you certainly should consider a different career. Hint: stop using a *nix system as root. That might help.

Re:Antecdotes != Evidence

[servant]

Not evidence, but my wife runs an employer provided and 'professionally maintained' laptop. She never had a BSOD on her XP laptop, this Win7 machine has had 2 BSOD incidents in the last 2 days.

Her use is strictly business and no 'roaming' on the internet for non-business. Her home machine is a fun only box (personally owned PC). Still XP. Never had a BSOD that was not explained by 3rd party utility software (patches from the vendor fixed).

yes, still anecdotal not sufficient for 'hard proof' evidence.

Re:Antecdotes != Evidence

[angel'o'sphere]

I have an old Mac that runs 10.6.8.
My new one runs 10.9. And my Mac mini 10.8.
Never tried to upgrade the old one, for various reasons.
But my oldest (meanwhile sold) G5 17" usually liked the upgrades.
I upgraded my ex GFs Powerbook from 10.5 to 10.7 and it got a little bit faster. No idea about your friend, perhaps a bad computer versus new OS combination?

Re:Antecdotes != Evidence

[LordLimecat]

The kernel doesnt come up until after that "binary stuff", aka the boot record and MBR.

After the kernel is up, there is no distinction between "scripts" and "binaries" at all.

Being a pedant myself, you are correct that scripts are technically binary, but in the context a binary is referring to compiled bytecode, while a script is human-readable shell-interpreted code. There is a massive difference in that
1) binaries generally cannot be easily reverse engineered or altered
2) binaries are much, much faster (being compiled)
3) scripts rely on a lot of pre-existing infrastructure-- the kernel must be up, the userland must be up, the command interpreter must be up. Binaries can run straight off the disk (ie, the bootloader).

If you can manage to get a Linux/OS X or other *nix system to behave like windows,

Go grab an older version of pfSense 2.0 (would probably work with versions of FreeBSD 8.3), do a fresh install and plug a USB keyboard in. TA-DA, driver bug causing boot hangs. I think it occurred with certain combinations of keyboards and motherboards, but it was a bonafide documented driver bug.

Go grab a copy of Ubuntu 12.10 and install on a machine with certain Intel NICs. TA-DA, another driver bug bricking your NIC. Driver had to be blacklisted in a post-release update.

Go find and install SoftEther VPN or S5 proxy. Watch where they stick their binaries, log files, and user config-- and then tell me how it conforms to the magical Unix standard.

Hint: stop using a *nix system as root.

When I figure out how to run kernel-mode drivers without root privileges or install packages without root, I'll let you know.

Re:Antecdotes != Evidence

[LordLimecat]

MBR in first line should be bootloader. Whoops

Re: Antecdotes != Evidence

[jd]

Firefox is Netscrape.

Re: Antecdotes != Evidence

[linuxrocks123]

Uh, no it's not, dude. Netscape was the original browser to use the Mozilla rendering engine, now called Gecko, and Firefox uses Gecko, but that doesn't make Firefox Netscape. There are other non-Netscape browsers based on Gecko, like SeaMonkey and K-Meleon. Calling Firefox Netscape is about as accurate as calling Chrome, Safari, or Opera Konqueror.

If you want a Netscape-like browser, by the way, you should look into SeaMonkey. I keep toying with the idea of switching to it but never get around to it because I've got over 100 tabs open.

Re:Antecdotes != Evidence

[eharvill]

I'd be curious to see the statistics of perceived OS Decay between installs on mechanical drives vs SSDs. I would always reload my system every 12-18 months on XP and then Win7. I don't think I reloaded Win7 again after I got my first SSD for my boot drive. I also did a clean install of Win 8.1 on an SSD, but it's only been 2 months so far.

Re:Antecdotes != Evidence

[Dutch Gun]

Fortunately, all the latest versions of the major browsers no longer make it so easy to install third party crap silently. If people are using duff browsers, as you put it, then there's really nothing to be done for them. It's not like anyone charges money for a modern browser, and even the latest versions of Chrome and Firefox are still available for Windows XP service pack 2.

I agree that it's hard to blame users for the sleazy tricks some companies use to install this crap, which is why I think there's an argument to be made that most people are actually better off in a well-tended walled garden (sad to say). Installations that do things like that should never get let past the gates. MS has recently realized this, which is why they started cleaning their app store a bit. Unless people can trust what's in the store more than what they get out in the wild, there's zero point to using an "official" app store.

Moreover, I think the media needs to do more in calling out companies that engage this sort of behavior. The negative PR from doing so needs to outweigh the financial advantages of getting in bed with sleazy companies like Ask.com and their horrid toolbar. At least, that would be the ideal, but it doesn't seem to be happening. These sorts of tactics are really hurting the PC platform by making it unbelievably treacherous for ordinary users to manage their own computers. It's no wonder people are switching in droves to smartphones or tablets as their "primary" computing platforms. At least it's not quite as easy (although not impossible, of course) for a single rogue app to compromise your entire system.

Re:Antecdotes != Evidence

[toddestan]

Welcome to Windows prefetch. Basically, Windows will preload your programs and data that it thinks you might use into ram after you boot. The idea being that if the program is already cached when you try to launch it, then it will launch faster. The actual result seems to be lots of disk thrashing after boot, and the more ram you have the more Windows will thrash as Windows will prefetch until it's full. At least they toned it down a bit with Windows 7, as one of Vista's faults was that it was far too aggressive about prefetching.

Re:Antecdotes != Evidence

[missneht]

You seem to misunderstand Galag88, he is not running Windows 7, he is running Windows 8, which was upgraded to 8 from previously 7 on his system. I have two separate computers one with Windows 7 and the other bought new two years ago with Windows 8, the 7 computer is slow, slow and slower, the 8 computer is still running very fast. So it does look like Windows 8 did get rid of the problem that existed in previous versions. If version 10 has the same OS ability that 8 has, then 10 should work that well also.

Re:Antecdotes != Evidence

[MrHanky]

Unlikely. I've got a Windows 7 install myself, running in dual boot for gaming. I believe it's from 2010, and since then moved first to an SSD and then from AMD to Intel (which takes a bit of time and magic, but is possible). It still runs just fine, but I almost only install software I need (i.e. games).

Come to think of it, I actually did end up with some garbage that took over my IE start page and search engine at one time. The problem with Windows is that it encourages bad usage, and that crapware is so much more readily available than good free software tools.

Looks you never used OS/2 :)

[martiniturbide]

OS/2 installation never decay as Windows. It is all on the config.sys, you simple delete all the stuff you don't like and it is done. http://www.os2world.com./

bennet haselton under a psuedonym?

[rogoshen1]

j/k this post was only 1 paragraph, instead of 3 libraries of congress. The style and narrative is the same though. =/

Re:bennet haselton under a psuedonym?

[Anonymous Coward]

Or worse. He's grown a small buy loyal fanbase that seek to replicate his style.

No, I have not had that experience since XP

[Anonymous Coward]

This is not a painstakingly maintained system, but I am not in the habit of installing toolbars and updater services. If an application installs any of those, they are deactivated or uninstalled. People install "registry cleaners" and "optimizers" and download "codecs", "players" and "downloaders". That's what ruins their systems. Once malware or anti-malware has gotten in, there's no way to restore a system to its former speed and security.

Re:No, I have not had that experience since XP

[apraetor]

Agreed. I've noticed the aforementioned tendency for PCs "in the wild" to slow down, but not once has it been due to the length of time since the last clean install. The people with this issue fall into 2 overall categories: people who know nothing about system maintenance, and people who know just enough to be dangerous. Group 1 typically end up with tons of bloatware, malware, and random spyware toolbars installed. Group 2 is aware of the problems plaguing Group 1 and think that "registry cleaners" are the solution, turning a perceived slowness into real software problems which bog things down.

Re:No, I have not had that experience since XP

[simplypeachy]

But PC Tools' software is convincing and alarmist, surely it wouldn't lie to me!

Re:No, I have not had that experience since XP

[apraetor]

To be fair, though, it would be nice to have the ability to do something similar to what OS X has supported for ages: copying the app folder and user home folders onto a new machine and having essentially the entire user-space experience transplanted, as well as being able to simply write over the existing System folder with one (from a backup image or from the installer image) and have a working computer. I know a lot of the reasons that's viable is because of the limited hardware support required, but given the nature of WOW64 and the potent self-healing Windows 8.1 has (I've used PowerShell to run repairs on systems with basically the whole Windows folder thrashed and had it fix itself) I'm hopeful it is in the future.

Re:No, I have not had that experience since XP

[Anonymous Coward]

Windows can do self-contained programs perfectly well, just use .jar files. (That's pretty much what .app bundles are on OS X.)

Re:No, I have not had that experience since XP

[apraetor]

That's not accurate. A .jar file is run within the system's JVM (for security you shouldn't even install that junk), so it's sandboxed. Yes, you can do self-contained apps in Windows, but it requires the program's author to at least have allowed for it to be possible; you can't take any arbitrary program and just move it to a new computer and have it work. Any app that's been converted for use from a USB drive is "portable", but it's still not convenient-by-default. OS X .app bundles are packages -- the equivalent to a program's subfolder within \Program Files\ on a Windows disk. Unlike Windows, though, they keep all their dependencies outside of System-level frameworks inside the package.

It's the Windows Installer's fault

[Anonymous Coward]

Installed applications in Windows should be entirely self-contained. They should have their own directory, their own temp files, and their own registry hive. When the application is removed, all of this should vanish as well.

Re:It's the Windows Installer's fault

[Darinbob]

I'd like to see package dependencies too. Microsoft applications are every much as convoluted as Linux apps when it comes to the files they depend upon. DLLs, shared directories, etc. But when uninstalling the applications they don't always uninstall the shared stuff cleanly. Ie, an app wants vbrun300.dll or such, so you visit the relevant Microsoft site and get it, but then you uninstall the original application but the dll is left behind; and there is no uninstaller for these libraries, they don't appear in the control panel.

I used to have a utility that would monitor all system changes during installation so that it could clean up later when uninstalling. Almost every time there would be some junk left over even after a successful uninstall. There would even be junk left over if you installed and immediately uninstalled without ever using the application.

Re:It's the Windows Installer's fault

[Joe_Dragon]

and you can say good buy to.

your email app from being able to see any office files.

no more flash , java , quicktime and more on the web.

adobe apps can't work with each other

no more visual pinball working with pinmame (at least both are open source and can have both join to one app)

IDE / codeing apps may be come hard to do.

No more NV or ATI driver apps

games can't have mods or map editors.

need I go on?

Re:It's the Windows Installer's fault

[Anonymous Coward]

All that started because disk space was scarce, applications wanted to offload common tasks, licensing third party libraries was getting ridiculous, and it all seems like a good idea... until DLL Hell was noticed. Please correct me or add your own experiences.
Today, we have massive 'winsxs' directories providing division between the scores (?) of vbrun300.dll variants. When 'winsxs' breaks, you'd better hope you notice it as the problem before putting in endless time trying to fix it; that's a reload waiting to happen.
You know what is clean? Having every application provide for itself. The most stable programs I have are nearly or entirely self contained. Portable versions of applications make me weep with joy. What makes me howl in pain are applications that permanently alter the system when they are installed.

Re:It's the Windows Installer's fault

[jcupitt65]

The big problem with the Windows model is that everyone can directly modify the registry and one badly-written installer can really mess it up.

OS X's solution is for programs to simply declare (for example) what associations they'd like (there's a small XML file called info.plist in the app bundle), and then for the file manager to update the associations for them as programs are installed and uninstalled by being dragged around.

Because the list of associations is being managed by one (hopefully) sane program, the chances of some random installer causing havoc are removed.

Not possible

[AqD]

I personally never experienced that for daily use. Installing/uninstalling applications and updates do since there are always some left-over garbage, but that's simply not addressable unless Windows kills all non-standard installers and forces them to play by Microsoft's rule (sadly even their own left garbage, but it's the first step to make them manageable), as it is on various Linux distributions.

With SSD, since it gets slower with more writes, a reborn system wouldn't be faster. It'd be pointless and you should really just dump it after 5 years.

Re:Not possible

[jader3rd]

but that's simply not addressable unless Windows kills all non-standard installers and forces them to play by Microsoft's rule

But Microsoft can't do that. Microsoft lets users do whatever they want to with their computers.

Of course not

[StripedCow]

You may find this interesting:
http://en.wikipedia.org/wiki/P...

Re:Of course not

[Barlo_Mung_42]

Windows has started going the other way. Win8.1 runs faster on my main dev box than Vista did and one of the main goals of Win10 is to bring along old systems and get as many people as possible to move on up. No doubt this has hurt the hardware folks though as there's less and less reason to buy a new system.

Mostly a layer 8 issue

[Anonymous Coward]

Funny. I never had to do a wipe since XP. Despite running my machines mostly 24/7. Maybe just learn how not to mess up your OS...

Read-only Windows directory

[Anonymous Coward]

I have this dream of running windows in a fully locked down fashion.
No writes allowed to the Windows directory. Any attempted writes result in immediate termination/quarantine of said app.
A long time ago I created a 20gb windows partition. Silly me. Now I'm cleaning out garbage out of the c: every few weeks just to keep the thing running.

Vague click-bait

[JoeyRox]

Saying the OS decays and slows down without any supporting data is about as informative as saying society is in decay. Btw I've never had this problem on any of my XP or Win7 systems.

Re:Vague click-bait

[Anonymous Coward]

Had an XP system slow down on both start up and shut down. Reinstalled the OS and it was much, much better. Haven't seen this with W7

Re:Vague click-bait

[Jumunquo]

Same experience. WinXP has serious fragmentation issues, and the included Disk Defragmenter is not enough to fix it. I have had success with third-party defragmenters that do more - in particular Norton Speed Disk, which is unfortunately part of the horribly bloated Norton Utilities. Even disabling everything, the Norton Control Center still starts up, so I had to remove the startup entry from the registry manually. However, it's worth the trouble because it really does return the system to its original snappiness.

Haven't seen it on Win7, although I am running on a well-performing SSD, so that mask some of the problems.

Re:Vague click-bait

[Anonymous Coward]

You just get used to it. I've heard this from many, many, many, many people; most of them trained Windows pros with the certificates and everything. 100% of them have realized they were wrong simply by benchmarking. You should too, and join reality with the rest of us.

Excuse Me?

[Jahoda]

If your disk develops bad sectors, the OS most certainly does not "have to work around them". Any modern drive will self correct its own bad sectors upon identifying them. If a disk is developing so many bad sectors that this is a constant problem, then the disk is about to fail, and you should expect performance to be degraded. This has nothing to do with Windows.

Re:Excuse Me?

[aardvarkjoe]

That, combined with the "device drive congestion" phrase (WTF is that supposed to mean?) makes it obvious that the submitter doesn't really understand anything about the topic.

And given that samzenpus gave it the green light, he obviously doesn't understand it either. Which I suppose is unsurprising.

Re:Excuse Me?

[simplypeachy]

I heard you have to reverse the direction of the SATA cable to clean that crap out. Same goes for slow broadband - either reverse the Ethernet cable or turn the WiFi router upside down.

Re:Excuse Me?

[pklinken]

This. Another trick is to put the cables in the oven for an hour or so at 70c, it resets the ionisation patterns and really makes everything feel snappier.

Re:Excuse Me?

[Anonymous Coward]

You can't reverse an ethernet cable; It's bi-directional. For slow broadband, you shake the cable to loosen up the stuck bits.

Re:Excuse Me?

[omnichad]

"device drive" must be "device driver", but still doesn't explain how you'd have any more than one driver per device actually loaded.

Re:Excuse Me?

[Anonymous Coward]

If a disk doesn't remap its bad sectors (from a bug, or because it's out of them) then NTFS has a $BadClus file to keep track of and avoid them.

Meh...

[Anonymous Coward]

Well, running Windows 8.1 Pro at the moment. Works as fast as I installed it back then on my Intel 520 SSD. Boots for the same amount of time and general tasks happen within a click. I don't know about you but my system is flawless. Maybe you should stop installing antivirus software and just use your PC as you would your smartphone ... and that means use it with a little concern and skepticism of the software you install.

OS Decay is largekly a myth.

[metrix007]

Since..let's say Vista, OS decay doesn't happen.

• No need to defrag anymore.
• Leftover registry entries don't slow down anything (the registry is an optimized database. Having a few extra records does not affect query time).
• Having extra disk space used up by unused applications likewise does not affect performance

You know what does affect OS performance? Malware. Silly PC users not installing updates and getting infected, or just infecting themselves because a dancing pig said it was a good idea.

Microsoft doesn't have to fix OS decay. They need to continue to educate their users about security and then make a version of notepad that can understand unix newlines.

Re:OS Decay is largekly a myth.

[martiniturbide]

Hi.

It is not a myth. The OS decays. All the stuff you install and de-install, all the crappy drivers that you install on the OS makes the registry gets bigger, it makes Windows to under-perform and slows it down. No matter if you un-install all the extra things you don't use anymore, the OS will never get back to the original state. That is why a full reinstall will make it better.

But there is one thing that I noticed. For the first time ever on Windows 8.1 the "refresh" functionality of the OS worked very good. It took a long time but it recovered to a refreshed state.

Re:OS Decay is largekly a myth.

[Anonymous Coward]

No need to defrag anymore.

The OS does this for you these days. With 8, most of the other clean-up is done at the side as well, which explains the submission as well as the mentioning of the driver problems, as the submitter hasn't have the opportunity to have the relaxing experience with class drivers.

make a version of notepad that can understand unix newlines.

Just install Emacs or Vim, depending of your religion.

Re:OS Decay is largekly a myth.

[El_Muerte_TDS]

WinSxS

Re:OS Decay is largekly a myth.

[metrix007]

Hi,

You're incorrect.

As I stated, the registry as an optimized database. A few extra records do not affect query time.

I will be happy to met money that my 3 year old install of Windows 7 will not have any speed decrease over a new install on the same hardware.

If the OS slows down, it is because there is something causing it that you can remove. It isn't due to "decay".

Re:OS Decay is largekly a myth.

[exomondo]

What of it? It's the side-by-side assembly folder to manage multiple versions of the same DLL on an application-specific basis, the UNIX way (which IMHO is better) is to have the version in the filename. I'm not quite sure how you equate this to "OS decay" though, sure it could eat up more disk space if you install and uninstall a lot of applications with poor uninstallers that don't remove their files properly but that won't necessarily slow the OS down.

Re:OS Decay is largekly a myth.

[Darinbob]

But refresh, as advertised, only works for modern "metro" apps. It does not do a clean refresh for the desktop which is what everyone uses. Or more specifically, for metro it will reinstall the apps for you but for desktop you will have to do this yourself.

Or at least this is what it said when I read the description from Microsoft. Thus I have never actually tried it because I don't want to blow my system away and reinstall everything.

Re:OS Decay is largekly a myth.

[Anonymous Coward]

Exactly.

When Windows XP was around after about 6-7 years from its release I figured out what was causing this "decay" symptom: anti-virus.

I observed that after uninstalling anti-virus software Windows XP was suddenly magically as fast as it was on Day #1. Likewise, installing anti-virus again made the advertisements about them not affecting system performance seem true. NOT! Given enough time, you will observe this "decay" in performance, and when you normally hear people say something like "it's about the time again" (to reinstall Windows) after 1-2 years of running.

Replace the harddrive with an SSD and problem is incredibly mitigated (though not technically solved -- it really is a problem with anti-virus software).

Re:OS Decay is largekly a myth.

[Anonymous Coward]

Hi,

You're incorrect.

As I stated, the registry as an optimized database. A few extra records do not affect query time.

I will be happy to met money that my 3 year old install of Windows 7 will not have any speed decrease over a new install on the same hardware.

If the OS slows down, it is because there is something causing it that you can remove. It isn't due to "decay".

^this, most people blame the OS, it is nearly always due to software that has left services running, startup files etc etc. While a lot of cruft gets left behind in the registry that particular cruft has no performance impact. The performance impact is the malware, the toolbars, java, tray items etc etc that people allow to be installed, OS decay itself hasn't been an issue since defragging was added to the OS itself. Someone blaming OS decay is usually a sure sign of someone that doesn't understand how much shit they have installed and really needs to hand their computer over to someone to clean it up.

Re:OS Decay is largekly a myth.

[Bite The Pillow]

I know it is a database, and slightly optimized. "A few records" would not affect query time, especially if they were not in the query path.

What about a lot of records? And how about a lot of records that are in the query path?

Re:OS Decay is largekly a myth.

[Anonymous Coward]

I will be happy to met money

I met money once. It was awkward, because I couldn't think of anything to say.
Afterwards, I thought to myself, "Damn, that was money."

Re:OS Decay is largekly a myth.

[benjymouse]

I know it is a database, and slightly optimized. "A few records" would not affect query time, especially if they were not in the query path.

What about a lot of records? And how about a lot of records that are in the query path?

It's a database. IIRC it uses B* trees. Search time is proportional to the logarithm of total number of records. Even "a lot" of records may not cause the height of the tree to increase. You generally need to *square* the number of records to double to search time.

At the same time, the registry hives are really, really robust. Windows keeps to redundant copies and even protects writes through the kernel transaction manager as well as the journal of the file system. Corruption is virtually impossible until the hard drive decays to a state where even the redundancy cannot make up for it anymore. Unlike text files, both metadata *and* data are guaranteed to either succeed in an atomic transaction.

(compare to the Unix way, where config files can be corrupted if the system/power fails during a write: File system do not guarantees *data* consistency for regular files, only *metadata* consistency, i.e. the fs guarantees that its internal structures will not cause it to go haywire on your files afterwards)

I suspect that this is actually the reason why there's a myth about corruption of the registry: With all of the redundancy, the registry is often the last component to fail when a drive succumbs. At the same time, Windows will refuse to start *if* the registry is corrupted. At that point the drive is in such a bad state, that even restoring/repairing the registry corruption will not save the drive.

Re:OS Decay is largekly a myth.

[Anonymous Coward]

It's a database. IIRC it uses B* trees. Search time is proportional to the logarithm of total number of records. Even "a lot" of records may not cause the height of the tree to increase. You generally need to *square* the number of records to double to search time.

SQL Server is based on B* tree and, just like Windows, can behave really badly when a poorly designed schema is being used. The problem with the Registry is that it's just a simple key-value store and Microsoft has layered upon it several layers of API and redirecting references for most things. Want to look up what program and parameters to use to open a particular file? Sure, lookup HKCR\.extension, then look up the related "file class", get the Command/Open key, oh shit that's got a GUID reference instead of a file path, look up HKCR\CLSID\GUID, oh bugger it's just a filename, go searching through HKLM\Software\Microsoft\Windows\CurrentVersion\Installer\... to find the full path for the program, etc., etc.. And if users have gone through the install/uninstall cycle a few times there are often dead-end paths causing the need to restart the search at various points along the way. It's really fucking inefficient.

Re:OS Decay is largekly a myth.

[Anonymous Coward]

Microsoft doesn't have to fix OS decay. They need to continue to educate their users about security and then make a version of notepad that can understand unix newlines.

I used to use Wordpad for that until they replaced it with the full-blown version of Word. Now I just use jEdit and am much happier for it.

Re:OS Decay is largekly a myth.

[SuiteSisterMary]

Again, it's all relative. Do you really think it takes your modern (less than 10 years old?) multicore computer THAT much time to parse though even several million records? In a binary database format? Enough time that you'd even notice?

Re:OS Decay is largekly a myth.

[MobyDisk]

There's more to it than that.

Since the registry is so often modified, it tends to become fragmented. Microsoft addresses that with background defragmentation, but that induces more I/O cycles. Even on an SSD that means more FAT entries so there is some cost. Windows also makes periodic backups of the registry, which takes time and causes more fragmentation. And while a query for a single specific registry key won't be TOO much slower on a larger more fragmented registry, lots of operations involve enumerating which IS slower. So the more entries under HKCR\CLSID, the longer it takes to instantiate a COM object. Some of that is dealt with by caching, but now there is more data to cache so less memory is available. Similarly, the more hardware devise you have used (displays, USB keys, ...), the longer it takes to search for a driver since they are all in the registry under HKLM. If you run Process Monitor you will be amazed at how an idle Windows PC is constantly enumerating registry keys.

Re:OS Decay is largekly a myth.

[metrix007]

I always run process monitor.

Look, lets make the claim even simpler.

Several years worth of crust in the registry from uninstalled or leftover applications will at most, have a negligible performance impact on querying or enumerating thee registry.

There.

Re:OS Decay is largekly a myth.

[MobyDisk]

I wish I had time to test this. In theory, I agree with you, but in practice, it doesn't seem to hold-up. I would love to setup a PC, run some benchmarks, install everything in the universe, rerun the benchmarks, leave it up for a year, rerun the benchmarks,uninstall, rerun the benchmarks, .... I would use a VM, but then one could attribute the performance lost to the host operating system.

Re:OS Decay is largekly a myth.

[Anonymous Coward]

I would be happy to take your money, as I have done to every single person who has made the claim you have. Learn to benchmark your system.

Re:OS Decay is largekly a myth.

[zb84]

It's not up to data or metadata to ensure consistency and avoid corruption by themselves. If that data's backend storage is a database, it's up to database. If it's a filesystem, it's up to that filesystem. Hence, on UNIX, stuff you're talking about depends on filesystem implementation. /etc is less "safer" than Windows Registry if it lies on ext2, and far more "safer" than it if it lies on properly configured ZFS. Other than that, your post stands correctly.

Re:OS Decay is largekly a myth.

[toddestan]

Actually, I wouldn't call registry corruption a myth, as it was a very real problem back in the Windows 95 and Windows 98 days. I'm sure that's one of the reasons why the registry is robust as it is now is from the lessons Microsoft learned the hard way from earlier versions of Windows.

XP

[Charliemopps]

I'm sure many of you have had the experience of taking a five-year-old PC, wiping it clean, putting the exact same OS on as it had before, and the PC is reborn, running several times faster than it did before the wipe.

Are you running XP? Ever since Win7 I've not had this problem since I switched off XP. Learn to uninstall things and learn how to manage startup.

All you're fixing by reinstalling the OS are all the configuration mistakes you made over the years. Stop making mistakes, or learn how to correct them, and you don't need to reinstall.

You could also set a restore point just after instal and revert...
Or reinstall the OS into the same directory it currently resides, having the same basic effect.

Your problem seems to be your lack of expertise in windows, not with the OS itself.

Also, I'm not a MS fanboy... hate em... but what you're talking about is not a windows problem. It's a user problem. I guess they should make it easier to deal with, but the fact of the matter its far easier to fix this in windows than Linux. And far easier to make linux unbootable via the same mistakes than windows.

Re:XP

[Noah Haders]

it's funny to think that in 30 years all of the Linux, Windows, and OSX will be like Solaris, and everything will be iOS/Android.

Re:XP

[Anonymous Coward]

20 years ago I left a comfortable SUN workstation to work on Linux.

I think 10 years from now, it will still be here.

Android is likely to get rebranded during that time, as Android becomes synonymous with unmanageable POS with a store infested with malware, the name will disappear and come back as something else.

Re:XP

[Noah Haders]

Tizen? CNux?

Get a SSD

[valinor89]

My 8 month old Win 8.1 install is as snappy as the first day and has only decayed because my SSD has a bug. (SAMSUNG 840 EVO) I thought my install was getting slow but after a SSD refresh it is almost as snappy as the first day. Has it decayed? Most likely, but when the difference is 1 or 2 seconds at startup you hardly notice.

We're already there

[Cowclops]

I used to reinstall XP every year or two to get it back to a fresh copy, but I ran the same install of 7 from 2009 until... well, the present though that computer is my second one instead of my main one now. Including installing games and at one point switching from an nvidia to an ATI video card. Runs as awesome as ever. This problem has basically been solved... SSDs and huge amounts of ram help.

Re:We're already there

[Anonymous Coward]

I used to reinstall XP every year or two to get it back to a fresh copy, but I ran the same install of 7 from 2009 until... well, the present though that computer is my second one instead of my main one now. Including installing games and at one point switching from an nvidia to an ATI video card. Runs as awesome as ever. This problem has basically been solved... SSDs and huge amounts of ram are the remedy.

FTFY. As mentioned before, you just don't notice it because you likely have SSD and 4-8 gigs of ram. Try running windows 7 on 2 gigs of ram and any non-SSD drive. It's measurably sluggish after a few weeks, and unusable after about a month.

Microsoft has solved this problem ages ago.

[140Mandak262Jamuna]

None of the machines are expected to be run for more than three years. You are supposed to dutifully buy a brand new machine with the latest OS every three years. If you had done that and had given unto Microsoft what was Microsoft's, then you would not have this problem at all.

actually...

[buddyglass]

The last Windows OS I ran was XP, prior to migrating to Mac. I didn't run anti-virus, never had any malware problems, and never had it slow down. I ran CCleaner maybe once a month and/or after installing / uninstalling an app or installing OS updates.

Not doing another backup.....

[djsmiley]

You deserve everything that MS decides to inflict upon you...

LOL. You expect MS to fix the problem ...

[UnknownSoldier]

... let alone understand it ? /sarcasm Naive ...

The continual bloat of _registry_ is the cause of the problem. That is not going away anytime soon.

Hmm, so why don't Unix machines have this problem ... gee, maybe because they don't use a single bloated binary config file.

ReactOS to the rescue!

[Anonymous Coward]

Windows is too stable for you? You want OS decay? Try ReactOS instead! It's an alpha quality Windows XP clone, and it doesn't even support NTFS. That's right, your FAT32 will be toast after the first blue screen. Reinstall every day, it's fun!!

Seriously, get some fucking perspective.

PIG Services like MSSQL Express

[Anonymous Coward]

This is what slows a PC down. Remove MSSQL Express now and you are back in black. Back. In. Black.

Install your favorite distro of Linux

[Anonymous Coward]

I work here in the state of Guerrero, Mexico in the dept. of scientific investigation of UAGro universities and was given a large touch screen Dell. Unfortunately it had windows 8 as the OS. I was having so much trouble getting anything done at all that I just went ahead and wiped it clean and installed Linux. I'm glad I did I haven't had any problems that a quick search for a fix didn't take care of. By the way if you are a windows user as I was then you might like Elementery OS. I also like other distros but I have found it to be so much more intuitively designed that I find that I can get things done more quickly. ...Why did I wait so long to make the switch?

Malware

[Anonymous Coward]

Probably the accumulation of malware stealing CPU cycles. It's hard to avoid with the current state of the art on any commercial desktop OS, except by being very careful and selective when browsing, opening email attachments, etc. And sometimes even that's not enough.

Windows XP was different...

[Lumpy]

XP was fast as hell until you patched it up to SP3. Microsoft borked the hell out of that OS. Windows 7 I have not had the gradual slowdown problem at all.

Re:Windows XP was different...

[Anonymous Coward]

Lumpy (12016) wrote:

XP was fast as hell until you patched it up to SP3. Microsoft borked the hell out of that OS.

Hint: XP SP3 was almost entirely about improving security:

New features in Service Pack 3

        NX APIs for application developers to enable Data Execution Prevention for their code, independent of system-wide compatibility enforcement settings[73]
        Turns black hole router detection on by default[74]
        Support for SHA-2 signatures in X.509 certificates[74]
        Network Access Protection client
        Group Policy support for IEEE 802.1X authentication for wired network adapters.[75]
        Credential Security Support Provider[76]
        Descriptive Security options in Group Policy/Local Security Policy user interface
        An updated version of the Microsoft Enhanced Cryptographic Provider Module (RSAENH) that is FIPS 140-2 certified (SHA-256, SHA-384 and SHA-512 algorithms)[74]
        Installing without requiring a product key during setup for retail and OEM versions

If the choice is fast vs. secure and you choose fast, don't complain when you get owned.

Re:Windows XP was different...

[Circlotron]

XP was fast as hell until you patched it up to SP3.

I'm with you. Nowadays I run XP SP2 as it comes off the CD with zero updates and no network or internet access in VirtualBox on Mint 17 for the occasions I must. (COM port behaviour with Wine can be iffy.) SSD and 16 Gigs of ram and it boogies. SP2 to SP3 made the wheels fall off.

Re:Windows XP was different...

[Anonymous Coward]

Same experience for me.

I'm fighting a loosing battle on Windows 8, on almost identical hardware :\

(And I'm very aggressive - I got almost nothing installed)

Difference is that I'm using the Windows 8 for dev. Maybe its the dev tools that bogs it down?

I'm not blaming the opensource tools, but I got visual studio. VS seems to hook into all kind of random stuff :O

Re:Windows XP was different...

[Anonymous Coward]

Microsoft is an anagram for borked. The clowns can't figure out how to steer a company. If they were a crew on a ship the coarse changes would look like a drunken sailor. They also seem to be breaking more stuff as they go. So now we jump from 9 to 10. Guess what, in Windows numerology that puts the next release square on a windows ME release again. I've been one step away from Linux and it looks like that final push has come.

Re:Windows XP was different...

[Anonymous Coward]

So providing free updates that slow the computer down to force users to upgrade and pay them for new OS is a failure?

It's their business model and it beats having to offer improvements and attract people into upgrading. Their problem is that Windows 7 was not slow and buggy enough. Business users will be able to stick with it for years more and save on upgrade and training costs. Many are still sticking with XP even though Microsoft has blocked available security updates for software so hacking will drive sales.

Why avoid 8.1?

[Anonymous Coward]

Honestly, I don't know what the complaints about Windows are going on about. I'm typing this on a Win 8 tablet, sitting in the sun, running real programs (including dev environment) that an iPad or Droid couldn't hope to handle. It took me an hour to get used to the new start page, and anyone who takes longer is really just not trying. Sure, if you are a Win hater, you've got good reasons and stick to them. If you like Windows, 8 is awesome. As Irving Berlin might say, anything your iPad can do, my Venue can do better (at three times the price).

Re:Why avoid 8.1?

[Jumunquo]

Well, you're on a tablet - that's why! Microsoft made the bone-headed decision to force mouse-and-keyboard uses to use the same interface, and not only it is awkward, but it is just plain inefficient. Also, most of the people who would take the time to try Windows 8.1 or buy it are not Win haters. They are the bonafide Windows users. That's why so desktop users are sticking with Win7, my gaming rig included. Or you could disable most of that Metro crap with Classic Shell.

Re:Why avoid 8.1?

[exomondo]

There are simple workarounds to the issue though. There's the boot-to-desktop option, there's the option of Classic Shell if you really need a start menu and then there's complete shell replacements like LiteStep.

Now with boot-to-desktop you don't have a start menu, for those indoctrinated in Windows usage that may be an issue so you probably want either Classic Shell or to adapt your workflow but for most of us that often or mostly work on platforms other than Windows (which don't have a start menu) this isn't a big deal. Application launching mechanisms like those on OS X map to Windows without the start menu: Dock (Taskbar), Finder (Explorer), Spotlight (Search), Desktop Icons (Desktop Icons), Launchpad (Start Screen) ... so the lack of a start menu is easy for those of us that use multiple platforms.

Re:LOL. You expect MS to fix the problem ...

[Lumpy]

No but the Dumb-ass developers and Distro makers throw config files all over hell in random places.

If your config files are not in /etc or ~/.appname then the developer is a complete moron. /opt/dev/random/stinky/appname/config.cfg is NOT acceptable.

if they figure it out

[iceperson]

maybe they can teach Google the trick. My wife's Note 2 was taking 15 or so seconds to open the messaging or camera apps and finally today I just did a hard reset and started from scratch.

It doesn't matter.

[NoMaster]

It doesn't matter whether Win10 finally addresses "OS decay" or not.

All that matters is that there's a succession of stories so they've got you talking about it - from website to website, day after day, one story after another.

Oh, and that you attack Mac users for being "sheep"...

Sounds like a 90's problem

[Anonymous Coward]

I have to say that I don't believe I experienced "OS decay" in any of my recent boxes. Recent includes one box that is some 7 years old and is running Windows 7 from pretty much the time it was released (ran Vista beforehand, and yes, that one was painful, but consistently so).

Windows 7 hardly degrades over time, at least in my experience. Sure, the more apps/devices you install the more time it may take to boot and the more memory these startup apps will consume, but I think that mysterious 'OS decay' is not all that mysterious anymore, and can be easily avoided by not installing crapware and an occasional (yearly?) visit to msconfig's Startup pane, to get rid of stuff that you may have once thought is useful but really isn't.

Another thing that pretty much killed OS Decay is SSD. SSD trims bootup times so significantly, and improves responsiveness of apps - that the so called 'OS Decay' - more apps running on startup and having to swap in/out of memory - has a much lesser effect even if you do suffer from it.

Last - I believe much of the perceived 'OS Decay' has a lot more to do with 'App Decay', or perhaps it's more accurate to call it the opposite - 'App Bloat'. Generally speaking, there's a lot less emphasis for memory optimization with every passing year - just take a look at how much RAM your browser windows are consuming. They're typically ultra fast - but they're trading RAM for speed. If you have a 2GB or 4GB system, OS reinstall or not - you're going to have a slow box as you start opening more than a handful of tabs, combined with a few Chrome extensions. At 8GB you could do OK if you're not too demanding. But admittedly, I only did away with memory related issues completely as I got a 16GB box. Combine a system that's constantly approaching out-of-memory with a magnetic disk, and you have a recipe for a performance disaster.

An XP problem?

[Animats]

That seems to have been a problem only with Windows XP. I didn't have it with Windows NT 3.1, NT 4, Windows 2000, or Windows 7. (I skipped XP).

Ubuntu Linux seems to have an incredible number of background processes that aren't really necessary.

Not OS decay, service/app bloat

[gatfirls]

Pretty much every app wants to have some update service or helper or whatever. Most of the time "cleaning" a PC that's dragging is a simple as running msconfig and unchecking most of the startup app and services.

I usually get a "b-bbut the last guy/tech/business said I have to reformat and reinstall! :( "

Re:Not OS decay, service/app bloat

[Anonymous Coward]

Figures. The difference between Windows and Linux support is that more people "know" how to fix Windows. They're usually the ones who complain about how hard it is to fix.

Well, I never ...

[CaptainDork]

... had to do a backup, disk wipe, and reinstall.

Ever.

Since Windows 3.1.

Not ever.

I know how to clean the damn thongs without performing radical surgery, and so does every professional I work with.

Re:Well, I never ...

[Anonymous Coward]

It's almost as though a mass-marketed consumer OS that does nearly everything for everyone will experience a wider variety of use cases and experiences.

Re:Well, I never ...

[simplypeachy]

Bingo. If the system is slow, you diagnose it. Once you diagnose it you fix the cause.

Either that or you just wildly accuse the registry, Microsoft or a conspiracy by RAM manufacturers. The Internet told me that registry cleaners and driver boosters work, so they must! Anyone blaming their damn thongs on anything but their washing powder is just...I'm sorry. I just couldn't help myself :-)

Re:Well, I never ...

[CaptainDork]

I hate when I do that. :)

Re:Well, I never ...

[Anonymous Coward]

I know how to clean the damn thongs without performing radical surgery, and so does every professional I work with.

/me awkwardly avoids eye contact
TMI?

Re:Well, I never ...

[Anonymous Coward]

Haha, I've heard that one MANY times, and 100% of the time, a benchmark proves this completely false. Try again though :)

Re:LOL. You expect MS to fix the problem ...

[ipstas]

at least you can simply rm -rf /opt/dev/random/stinky/appname/config.cfg and the problem is gone. Unfortunately you can not simply delete the folder in windows, more likely you will have later an issue with uninstaller saying it can not uninstall the application

Some credit, some criticism

[Anonymous Coward]

After installing the preview today and using it for a couple of hours, my conclusion is that we are still putting lipstick on a pig. The new Start Menu is a joke... it's somewhat more useful but the inclusion of XBOX, weather, health, news, etc on an Enterprise install is nonsense. The flat pastel UI is still crap. And the melding of Metro and classic windowed apps is bad... there is no easy link to Control Panel, only a link to the hobbled PC settings Metro app. The install is straight up Win 8... they barely changed anything from where I sit.

On the plus side, even in a VM it is relatively snappy, which I did find with Win 8.1. And the default install is only 10GB, which easily destroys previous versions. Our 8.1 test VM came in at 17GB.

If it's still going to take them a year to tweak this hunk, there is no question I am all Linux all the time from here on out. I disagree with the one UI for all platforms approach, and with the consumer apps for Enterprise users configuration. Yes I know I can fix that with Group Policy, but how about they ship an Enterprise edition in an Enterprise configuration right out of the box.

My only Windows will be a Win7 VM for the two apps that don't run on Linux (Visual Studio and MS SQL) and my one piece of incompatible hardware, a Dymo label printer.

OS decay

[koan]

Never had this problem, Apple has it though.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Randomly placed config files in Linux is a problem, but nowhere near as large as the problems caused by the registry on Windows. In Microsoft's defense, when the registry was originally thought up probably in the early to mid 1990s, it was thought to be a good idea. However, the test of time has not been kind to the Windows registry.

Microsoft would do better to follow in the footsteps of Apple when they switched from System 9 to OS X and redo EVERYTHING in the Windows operating system instead of keep adding more duct tape to fix each new problem with the ancient os.

LOL. You expect MS to fix the problem ...

[Anonymous Coward]

I think what the commenter was attempting to state is that Unix systems tend to maintain their configuration as files on the filesystem which result in:
O(1)
lookup time for configuration.

Whereas the Windows registry is a database which results in:
O(log n)
lookup time for configuration.
This is also coupled with the fact that entries in the registry continuously accumulate and are historically not scrubbed when a program is removed from the system.

Remember looking up program configuration information tends to happen at program start which results in slower start times.

How about get rid of the registry?

[Anonymous Coward]

That subsystem never shrinks, I remember in Windows 2000 perhaps, there was actually a gui setting to dictate the size of the registry. Anyone remember this?

No it probably won't, just use an image

[NemoinSpace]

Actually Windows 8 seems to do a better job than windows 7. After 2 years I haven't touched my windows 8 system in this respect. As an added bonus most of my data is on network drives, so the ease of laying down a base image is the least painful of all.
Linux suffers just as much from cruft buildup and most of my linux installs are on VM snapshots. Defrag a hard disk? Wow haven't even thought about that in years.

Re:LOL. You expect MS to fix the problem ...

[dltaylor]

systemd will fix that for you, with the added "feature" of parsing text

Decay is required...

[eyepeepackets]

for their business model to function...and they won't break it.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Relative to appname it is... In the shop where I work, /etc is exclusively for system provided packages configuration , ~/.appname for user specific configuration and /opt/appname/etc/configuration.cfg for application installed outside of those provided by the OS.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

at least you can simply rm -rf /opt/dev/random/stinky/appname/config.cfg and the problem is gone.
Unfortunately you can not simply delete the folder in windows, more likely you will have later an issue with uninstaller saying it can not uninstall the application

Why would you remove the application's configuration files before you remove the application? Are you meaning to re-install it? Then uninstall the application, or run the Repair option in Programs and Features. If that part is broken, blame the package maintainer. And if you still need to blow away the configuration files, do it after the application is uninstalled.

As for the registry, there's a great front end that fully resembles a file system view and dozens of third party tools that make parsing it a snap. Most of the time you want the software hive.

It's nobody's fault that you don't know how to use Windows, much like it's nobody's fault that I don't know how to use Linux.

ARE YOU LIKE STUPID????

[Anonymous Coward]

It is simple to Windows running at top speed.

1) fix the PAGEFILE. Go inot the settings and change ti to fixed size - 2x-3x size of ram - both of minimum and maximum size. Do not let WInodws manage it! It wipes it out and builds each time you boot the system. Since MS builds file from 0 sector out, and does not look to size it needs, it become fragmented over all the holes that have been left by other processing creating and deleting files.

2) Get a real DISK FRAG Tool. IOBIT has free one the is great. It will even defrag the PAGEFILE (if you fix it) and well was the registry, during boot. Further it will defrag on the fly when your use of your machine usage is low, to keep it in shape.

3) Clean your registry! IOBIT also has a great free tool Advance System Care. It will scan and cleanup most of your booboos. Normally the first time, I run it on machine that is "slow". It wil fix upwards 9000 errors the registry. Also will tracck down many junk files left in temp directories. One machine was helping on it cleared from one temp directory over 20000 opjects left behind from a word processor.

4) Dump the System Restore from time to time. This is just junk removal. It has its place and can be of great help. But if you have not installed any new in say the last month, disable it, to clear out the trash, and reable it.

Again it is simple HOUSEKEEPING!!!!!! And stop whining.

If you really want to run the FASTEST WINDOWS install??? PC with 8GB or 16GB of ram, Install Linux 64. Install a virtial machine software (say VMware) Create a windows VM. Do what is listed above. Fire it up. It will be faster that loading windows on the bare metal of the same machine.

Re:ARE YOU LIKE STUPID????

[ewhac]

1) fix the PAGEFILE. Go inot the settings and change ti to fixed size - 2x-3x size of ram - both of minimum and maximum size. Do not let WInodws manage it! [ ... ]

Better still, move PAGEFILE.SYS off of C: entirely, preferably on to its own spindle if you can. That way the swapper isn't having a fight with every other application in the system for accessing system files; and PAGEFILE.SYS itself won't become fragmented.

Consider moving %TEMP% and %TMP% off of C: as well.

4) Dump the System Restore from time to time. This is just junk removal. [ ... ]

Sadly, this appears to be an all-or-nothing affair -- on XP, you can either delete all restore points or none of them. It would be nice to delete those that are, say, more than a year old.

Re:ARE YOU LIKE STUPID????

[simplypeachy]

In the unlikely event that anyone is this deep, looking for meaningful advice, I offer it here. I could address parent's other foaming-at-the-mouth rants, but I'd just be repeating other explosive and rude comments I've already made on this article.

"4) Dump the System Restore from time to time. This is just junk removal. It has its place and can be of great help. But if you have not installed any new in say the last month, disable it, to clear out the trash, and reable it."

No. Do not "dump" it ever. Follow the Control Panel into "System and Security" and then "System". Click on the "System Protection" link from the left-hand pane. You will notice a setting for each fixed disk for System Restore. On the disk that's so hurting for free (i.e. wasted) space, click it then click "Configure".

Here you can choose how much disk space it consumes and you can then completely forget about System Restore until it saves your bacon. You will be glad that you let Windows decide when to delete Restore Points because I can assure you: it knows better than you do.

Re:ARE YOU LIKE STUPID????

[mr_mischief]

Better yet, don't use a spindle for this stuff at all. Get a cheap SSD. Yes, it's write-limited. If you're only using it for temporary files and the page file then there's no long-term data to lose and it's a hell of a lot faster.

If you're on a gaming rig with a whole lot of memory, the page file doesn't need to be 2 to 3 times the size of RAM, either. With 16 or 32 GiB of RAM and only running the OS, one game, a voice app or such, and a game library app (Steam, Origin, etc) then you only need a page file for stuff that's proactively paging. Set it to 4 GiB and forget it.

Also, 7 and 8 handle dynamic page file sizing better than XP so it's not as much a concern to have a fixed-size one on a spinning disk as it used to be. It may still help if you pick the optimal size, though.

NO!

[darkain]

NO!

http://en.wikipedia.org/wiki/B...

No shit

[Sycraft-fu]

I've found that there are really no issues with regards to running a new OS for long periods of time. There was a time when regular reinstalls were a part of my regimen but that is long past. I reinstall only when there's a specific need. That doesn't mean I just put up with a slow computer either, I demand very fast performance from my systems. A reinstall just isn't needed to maintain that.

Likewise, components have gotten much better, and upgrades more incremental, so I've found the need to buy new hardware that would necessitate a reinstall to be less.

My guess? This person either has crap on their system that causes issues, or this is just a "magic incantation" they've always done for supposedly better performance without understanding or examination.

Re:No shit

[Zynder]

Do you have some special set of things you do to maintain that performance or do you do the same things all the CNet type sites tell you- turn off superfetch, kill Windows Search, etc? I already do most of those but I'm always looking for one more clock cycle of performance!

Re:No shit

[simplypeachy]

You forgot to turn off System Restore as well! System Restore killed my SSD! Once it's dead we can start on the carcass of the DNS Client service. Pretty sure Microsoft is using that to spy on my web habits...

Re:No shit

[Zynder]

Oh it's off. There is a huge list of things I turn off and I didn't feel like listing them all, even if I could remember them. Normally I have to go look up one of those "69 Million Tips to Make Your Windows Fly" articles to refresh my memory.

Re:No shit

[Sycraft-fu]

Nope. I find none of that stuff helps. All I do is:

1) Run a good AV scanner (NOD32) as a last line of defense to make sure I don't get malware.

2) Uninstall shit when I don't need it.

3) Have normal maintenance shit like have trim run on the drives at regular intervals and such.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

... so why don't Unix machines have this problem ... gee, maybe because they don't use a single bloated binary config file.

Just give Poettering some time, he'll take care of this.

There is an easy workaround

[Dr. Spork]

When I install Windows, I work hard to set up everything exactly as I like on install day. Then I make a backup of the OS partition - which has only programs, no photos, videos, etc. - using Acronis TrueImage. Then I proceed as normal, and when something gets screwed up, I just restore from backup. This completely undoes any effects of winrot, and the system immediately feels like it was installed that day. What I usually do then is update my applications and settings, and immediately make a new backup. A full restore takes about 4 minutes, and a backup with max compression takes something like 12. I find this so convenient that I use no antivirus. When I start to suspect that I may have installed malware, I just restore from a backup, and four minutes later, my system is perfect. I've been doing this since Win2K days, and if this method weren't available to me, I wouldn't be using Windows.

Re:There is an easy workaround

[Anonymous Coward]

Why? there are several cheap programs (ui for disk filter driver) that "freeze" your disk partition except manually precise folders setup in configuration... you can leave documents, desktop and ProgramData "unfreezed" and everything else protected... after next reboot any changes done to protected areas are gone. Just unfreeze all to install and update. I use ShadowProtect as last barrier to virus, malware and rootkits, i know RollBackRx offer freeware somehow similar and microsoft used to have SteadyState

Best thing since ME

[ChadSmith4920]

Windows ME was the first Windows operating system with the now-common System Restore feature, allowing you to "roll back" your computer to a previous date and hopefully remove any recent errors.

But is Linux any better?

[Anonymous Coward]

Is Linux really any better, though? I've had a Fedora system that I've just been updating continuously for several years now. It was fast at first, but after pulling in so many updates for so long my system fucking crawls. It got much worse after systemd was installed. My slow system got even slower!

Re:But is Linux any better?

[countach74]

Hmm... I've never noticed any slow down. And I've done all sorts of terrible things to my system, package-management wise.

Re:But is Linux any better?

[jonwil]

I ran Gentoo for a while and my system actually got a little tiny bit faster over time as big heavy packages got rebuilt with newer GCC versions that optimized the package better than the last GCC version did.

Of course that didn't make up for the fact that it could take the best part of 3 days or more for the ancient Pentium 4 I was running it on to do a full emerge pass if certain larger packages all happened to be updated at once.

Re:But is Linux any better?

[Anonymous Coward]

Never noticed slowdown any issues on Debian. Maybe it's your package manager, or you should clean your /tmp folders onces in a while.

Re:But is Linux any better?

[Gavagai80]

Might be a Fedora issue or more likely something specific to your PC. Haven't seen any slowdown with Kubuntu.

Re:But is Linux any better?

[bingoUV]

Fedora is never supported for over 18 months. I've run some for 20 months, never faced this. Updates dry up, so effects after that wouldn't be possible.

Re:But is Linux any better?

[Anonymous Coward]

Bullshit, there are always continual updates to various parts of the system.

Re:But is Linux any better?

[Anonymous Coward]

Either you're lucky or you're oblivious. I think the second one is far more likely.

Re:But is Linux any better?

[Barsteward]

you'd best move to opensuse then, mine got faster after systemd was introduced. or perhaps you're just trolling systemd

Re:But is Linux any better?

[greenfruitsalad]

nonsense. on my thinkpad, i installed "debian potato" in 2001, but as i ran dist-upgrade every few years, i've had to:
* add more RAM (128MB hasn't been enough since woody)
* migrate the installation to >20GB hdd (etch-and-half just wouldn't fit with all my applications)
* with the stupid vesa driver, I'm forced to run windowmaker as my UI (everything better needs 3d and enlightenment is no longer available)
* the last release that properly supported apm (power management) was woody
* oss (sound) went completely to hell with etch (or maybe even sarge)

so now, even though i have a perfectly fine as-new laptop, i'm forced to buy a more powerful machine just because kdm/gdm alone would eat my entire ram (384MB) for breakfast.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Config files go either under

* /etc/app - system wide, distribution provided
* $HOME/.config/app - user specific, don't just dump to $HOME
* /opt/app - foreign applications, system wide

Then you have 3 places. Maybe 4 for legacy $HOME/.app stuff

Why is this not the same as "The Registry"? Because it's filesystem level and there is no stupid GUUIDs for keys. Imagine if some application installed their config files under unreadable with GUUID paths. Then the under does not know WTF those things ever suppose to be.

But that's a system that COM and COM+ invented. It works great until someone does not follow the rules, which turns out to be quite a few users and vendors.

Fix for Win 7

[Anonymous Coward]

Just make the following into a batch file and run as Admin:

@echo off
FOR /F "tokens=1,2*" %%V IN ('bcdedit') DO SET adminTest=%%V
IF (%adminTest%)==(Access) goto noAdmin
for /F "tokens=*" %%G in ('wevtutil.exe el') DO (call :do_clear "%%G")
echo.
echo Event Logs have been cleared! ^
goto theEnd :do_clear
echo clearing %1
wevtutil.exe cl %1
goto :eof :noAdmin
echo You must run this script as an Administrator!
echo ^ :theEnd
pause>NUL

Hu?

[WaffleMonster]

No idea what TFA is talking about.. Only "decay" I've noticed is caused by people getting suckered into installing malware.

Re:Hu?

[manu0601]

TFA does not tell us about the fixes to the problem either. Useless link, don't waste your time and skip it.

Re:Hu?

[Anonymous Coward]

A comment from someone who actually know something about computers, Good to see that there still is few left on slashdot!

Re:Hu?

[Anonymous Coward]

Or have corporate IT departments that auto-deploy it for you (thanks again, McAfee).

Re:Hu?

[Anonymous Coward]

Well, this world is full of morons who hose something up and then blame the technology for their problems. A majority of users do this, in fact.

Don't wait on the wiping...

[dnebin]

When I get my new PCs, the first thing I do is wipe them clean and reinstall windows.

I'm just careful when I grab the disk labeled "Ubuntu" because I know that version of windows doesn't come with all of the problems the OP identified.

Re:Don't wait on the wiping...

[Jumunquo]

That's just sad - why are you even buying new PCs with Windows preloaded and paying for something you'll wipe immediately? You're clearly not the clueless user so just buy the parts yourself and stick them together. Dell/HP basically use the cheapest, crappiest parts they can get, and that's why their standard warranty is 1 year when most quality computer parts have a 3-year warranty. That is why the Apple fanboys always rag on PCs. For just slightly more than Dell/HP, you can build a faster, quieter quality computer that will last 10+ years and with parts you can individually upgrade. SSDs are cheap now too, and they give the huge speed boost. Dell/HP gives you like a pathetic 60GB hard drive and change you an arm and a leg to upgrade it.

Re:Don't wait on the wiping...

[dnebin]

Sorry, but I think you might have missed the sarcasm that I was aiming for...

But I will add that I no longer build my own PCs. Not because I can't, but mostly because I had a hard time settling for the low-end, cheap parts (if I'm going to build a rig, it's going to be the hottest rig around, at least for the next week or so). Buying OTS ended up saving me a lot of time and money. I usually go w/ HP's, bump up the memory and otherwise go w/ the system as-is. They may pick up cheaper parts, but for the most part I've had nothing but success getting linux running on them.

LOL. You expect MS to fix the problem ...

[DraugTheWhopper]

The continual bloat of _registry_ is the cause of the problem. That is not going away anytime soon.

I've got 277 items in my add/remove list, dating back to about 2.5 years ago, yet my system is almost as fast as it was back then. Why? I'm not naive. I have no PUPs, malware, or other unwanted programs, I run MBAM and MBAE instead of a "real" AV (much lighter real-time protection), and I'm smart about what I let run at boot. Registry bloat is not a problem, it's clueless users who cannot maintain their system. The only issue right now is that my memory usage is a bit high after a clean boot, but that's because I'm running Nvidia drivers, Geforce Experience, uTorrent, Steam, three different cloud storage programs, and a file indexer covering 500,000 files (Everything FTW!).

Win7 has been good for me too

[Radical Moderate]

I rebuilt my XP boxes every year or so. I've had Win7 on the same box for 3.5 years, no problems. It could be that today's hardware is so over-powered, not even MS can generate enough bloat to slow it down.

That said, it would be nice if Win10 forced apps to behave better and kept the registry clean. Not holding my breath.

Re:Win7 has been good for me too

[sandytaru]

That's what I'm suspecting. I'm on a year old power rig using Windows 8.1, and I haven't noticed any slowdown at all yet. Usually about a year out is when I start debating whether I should reinstall Windows. Haven't felt the need yet at all.

Re:LOL. You expect MS to fix the problem ...

[simplypeachy]

You should understand what you are talking about before going off on one, otherwise you just make yourself look silly. The registry is cached in RAM while the system boots. It's tiny. Have you ever even gone to see how big the registry hives are? Are you now telling me that accessing parts of such small files is "the cause" of a Windows system's performance degrading? Do you have any idea how long registry reads and writes take?

Just because bullshit is repeated ad nauseam doesn't make it true.

Re:LOL. You expect MS to fix the problem ...

[DocHoncho]

As for the registry, there's a great front end that fully resembles a file system view...

I've called regedit.exe a lot of things over the years, and great was never one of them.

windows 7 addressed this

[Billly Gates]

Decay is caused by poorly written apps writing to the registry each time run. Windows treats it like a database with copies and forks and you have a recipe for disaster.

UAC controls and warnings if an app didn't install right fixes this

Re:windows 7 addressed this

[toonces33]

On newer versions of Windows, a typical application will no longer have permissions to write to HKLM\Software - any per-run gunk should get stored in HKCU instead, and it is a lot less painful to wipe your profile and start from a clean HKCU hive than it is to reinstall all of Windows.

I have long suspected that the problem is simply bad registry fragmentation, but I haven't tested the theory. For what it is worth, there are tools out there that claim to defragment the registry. I suppose they might help.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

I've called regedit.exe a lot of things over the years, and great was never one of them.

It's certainly better than that piece of shit dconf-editor that Canonical uses to administer dconf, which is the current Windows registry favorite alternative in use with Linux applications right now. dconf-editor is extremely limited compared to regedit - fucking hell you can't even use dconf-editor with just the keyboard!

Sorry, pet hate. Linux has a registry - dconf (originally gconf) and no-one's written a respectable GUI to manage it yet.

Dealing with slowdowns

[Sigma 7]

Even if Windows slows down over time, there's easy ways to deal with it.

Since Windows XP, you have a program called "MSConfig" that allows you to remove any startup programs, especially ones that are pure redundancy or are otherwise not useful.

And with modern systems - Web browsers slow down the system more than anything junk that accumulates in the OS. I've had both Firefox and Chrome running at the same time, with the resulting commit charge around 8GB, sometimes approaching 12GB. Once I stopped using one of the two browsers, the constant thrashing stopped, and everything else is much more responsive. (Firefox is still freezing, but that's a memory leak issue.)

NTFS | Windows Registry?

[Anonymous Coward]

I think the big problem here is the way NTFS file system and Windows Registry works...

http://www.howtogeek.com/115229/htg-explains-why-linux-doesnt-need-defragmenting/

Betteridge's law...

[Wootery]

Betteridge's law of headlines:

Any headline which ends in a question mark can be answered by no.

I'll believe it when I see it. It's not just Windows that has this problem, after all. Android and Mac suffer from it, and even Linux isn't immune (or there'd be no Paco).

Re:Betteridge's law...

[guruevi]

I've never seen either a Mac or a Linux box suffer from 'decay'. The box will work progressively slower compared to newer, faster systems which you may get used to eg. at your job but I've never seen the box slow down any measurable amount (unless of course, the hard drive is getting close to full etc). Windows on the other hand simply slows down due to updates and antivirus getting heavier and heavier, Windows machines will also fill their own hard drives (I manage some boxes that have been installed when XP first came out, even with minimal user data, the machine has gotten cleaned up multiple times due to a full hard drive (mainly logs, updates and system restore data)

Re:Betteridge's law...

[TechyImmigrant]

>I've never seen either a Mac or a Linux box suffer from 'decay'.

So you weren't a big user of MacOS 7?

Re:Betteridge's law...

[NJRoadfan]

Rearranging the load order of System Extensions and Control Panels was fun. It took Apple until 7.5 to add that nifty Extensions Manager control panel to the OS instead of a 3rd party add-on.

Re:Betteridge's law...

[Anonymous Coward]

What if the same article was titled, "Will Windows 10 Still Suffer OS Decay?"

Re:LOL. You expect MS to fix the problem ...

[Gothmolly]

Don't worry, with systemd we'll catch up to Windows any day!

Piffle.

[Anonymous Coward]

"OS decay" doesn't exist... except for Microsoft releases...

Personally, I wouldn't be surprised if it isn't intentional.

How else can MS release a "newer and faster" system when no real improvement can be done?

The bigger Problem is their

[Anonymous Coward]

Either you get system decay or upgrade hell.

Even linux has the upgrade hell problem--and it's forced onto you since the latest -rc type apps are readily available. I tried flat installs on a DellPowerEdge, DL380, System328 and an X3650 blades.... AND nothing worked from opensuse 12.1 to latest, Ubuntu 13, CentOs... only Ubuntu 12.04LTS worked on just the DL380. Either issues with graphics, or RAID or CPU. Really it all comes down to graphics hardware... really, really, no one is stick to the VESA standard it appears.

Re: The bigger Problem is their

[Anonymous Coward]

Never had any problems installing slakckware, debia,n or ubuntu on any machine and make. Don't know what you're talking about.

Re: The bigger Problem is their

[Anonymous Coward]

Yeah installing works just fine. Then you install gpu drivers for 30 minutes on a good day. For 2 hours on a bad day. Then you try to get your wifi to work for another 2 or 3 hours. On a good day the printer might actually be compatible.

Then a few weeks later you run security updates and realise that new kernel update killed the gpu drivers and you're back in the world of the 90's trying to remember where you saved the installer and navigate to it via the ultimate crutch of linux aka CLI. At this point the driver install may actually work again, or as it often happens it won't and you are left thinking - "wouldn't it be easier to just wipe the whole effin' thing and start again?"

Re:The bigger Problem is their

[steak]

why are you installing X on servers? /*nix_elitist

Re: The bigger Problem is their

[Anonymous Coward]

That is pretty much my experience every single time I have tried to install any of various Linux distros over the past couple of decades. Eventually I just give up and reinstall Windows.

The OS isn't something I want to mess with, I only want it to support my hardware and run my software. Nothing does that better than Windows.

actually no

[Anonymous Coward]

I have a pc I bought in 2009 with Vista on it. I only installed Win 7 at one point. Don't remember if it was an upgrade or clean install but it never slowed down compared to before.

I know people who format and reinstall every other week. I have no idea what they are doing to their PC.

You just answered your own question.

[Anonymous Coward]

Microsoft does not come up with new versions of Windows that will breathe new life into five year old computers. Microsoft designed windows to be a PITA to install on an old PC so that you'll upgrade to a new one with the latest Windows pre-installed. Microsoft is about putting as much planned obsolescence into their software as the customer will allow. Microsoft won't embrace interoperability, standards, and robustness until they absolutely have no other choice.

Re:LOL. You expect MS to fix the problem ...

[caseih]

Unless this is new in the last month, no it certainly doesn't. Stop spreading this kind of misinformation.

Has your mom stopped blowing the milkman?

[Anonymous Coward]

How about it?

Windows 7 fixed that

[Anonymous Coward]

By automating defrags; something that people never did on their own. They didn't even know about it. And page files didn't grow and shrink as much, becoming fragmented as a result. Worst thing about the page files is the system would refuse to defragment them. So you'd get systems with 16,000 little fragments all over.

Reading all the comments 'defending' Windows ...

[angel'o'sphere]

Windows ... : http://baetzler.de/humor/haiku...

One of my favorites:

Yesterday it worked.
Today it is not working
Windows is like that.

Windows is not an operating system, it is a menace to society.

However that puts the Linux movement into a bad 'catch up' (with what?) situation. A shame it is not more prevalent meanwhile.

Markets are not driven by price or functionality ... regarding what we 'could' do with computers/software everything you perceive right now is 30 years behind what *I* did 20 years ago at the university.

The last 50 years of research have no impact at all at current day computing, except for silicon and SOCs etc.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Why do you think it is faster for the operating system to access myriad various bits of information as required from a multitude of idiomatic files scattered all over the place than it is to read a single simply organized hierarchical structure that easily fits into memory?
Windows has many design flaws, but settling on a standard centralized configuration structure isn't one of them. People who don't like it are just special snowflakes who don't like being told what to do.

Re:LOL. You expect MS to fix the problem ...

[kiphat]

1. Problem with application.
2. Decide it's the config file and it's easier to start from scratch
3. Rename config file to configfile.bak
4. Start application
5. Problem solved.

The problem I have with this...

[roc97007]

...is that for Microsoft to create an OS that doesn't slow and become wonky over time removes one of the primary reasons to upgrade to a new version of Windows. Already Microsoft is dealing with Old Windows That Won't Go Away (XP, and now Win7). It is in their best interest for the OS to degrade over time. I can't imagine this obvious cash cow going away. And if so, what replaces it? MSFT tried floating OS as subscription before, and it didn't fly. Unlike the x-box, some phones and their competitor's platforms, Microsoft sells OS's and applications, not hardware. So an OS you can buy once and use forever (or for the life of the hardware) just isn't part of their business model.

So.... what, then?

This is a serious question. I'm a user of MSFT products. Until certain apps get ported to Linux, I'm likely to continue to be a user of MSFT products. But the OS to me has never been the app. It's a program loader and resource manager in which I run the apps that I actually use. I have no interest in new versions of the OS, as long as it'll still run my programs. I was one of the people who didn't leave XP until forced. And I won't leave Win7 until forced. I don't look forward to OS upgrades, I want to get work done. It seems to me that this frame of mind directly contradicts Microsoft's business model of endless costly upgrades. How are endless non-costly upgrades going to work for them? (It certainly works for me, but I don't really believe it yet.)

Re:The problem I have with this...

[Overzeetop]

Say what? Why would I "fix" a problem by buying a new OS. If things get massively out of whack, you reinstall the original OS. What forces people out is dropping of driver support and bug maintenance and, quite honestly, you are going to expect that after a while.

I'll agree - upgrades should offer productivity benefits, and for the most part the past 3-4 iterations have been essentially cosmetic upgrades and shuffling of the deck chairs, making it harder - not easier - to maintain the OS (since controls are now buried in 2-3 different places, and the obvious ones are dumbed down for basic users). There are some advantages to 8 in maintenance, but they're overshadowed by the layering of control. Adaptive screen brightness on my laptop, for example, is controlled in 3 separate places, and it's an OR function for the "on" condition - all three must be turned off to eliminate what I can only describe as the most annoyingly 1980s version of a screen dimmer ("fading" in halting 5% steps based on average screen luminance instead of room brightness).

Re:The problem I have with this...

[Anonymous Coward]

Device drivers only getting Windows certification if they are designed to overuse the hardware to artificially reduce lifespan? (i.e. allow GPUs to run hotter than they should, juggle around data on a SSD every now and then to run the 'writes left' count down a bit more etc). Dead hardware = new computer = new windows license sale.

Re:The problem I have with this...

[roc97007]

> Say what? Why would I "fix" a problem by buying a new OS. If things get massively out of whack, you reinstall the original OS.

That works fine for gamers. My nephew, a rabid gamer reinstalls his OS probably once a week. It's his answer for any issue. But, he doesn't do anything with his PC except game and chat.

But I use a lot of different tools in my work, and reinstalling everything is problematic enough that I avoid it for as long as possible. It's not a matter of reinstalling Crysis, move your saved game folder off the thumb drive, done. If I'm going to go through that process, it has to be carefully planned, and I have to look at the support lifespan of the OS as part of the criteria.

Typically, if the OS gets massively out of whack, I find the problem and fix it. It's that gradual collection of random and hard to find cruft that's more of the issue.

Re:LOL. You expect MS to fix the problem ...

[AbRASiON]

Yep, it's a registry thing.
Until stuff is sandboxed properly this is going to continue to occur (although it's slowly gotten better over the years) it's still not ideal.

BSD

[Anonymous Coward]

Microsoft needs to license BSD...

Re:LOL. You expect MS to fix the problem ...

[Rich0]

... so why don't Unix machines have this problem ... gee, maybe because they don't use a single bloated binary config file.

Just give Poettering some time, he'll take care of this.

Actually, if anything he is pushing to improve things - the goal is to allow stateless systems - as in you can mount your distro as /usr on a ram drive and have everything work. If you use something like systemd then all the OS-provided stuff is in /usr and will be cleaned by the package manager, and the only stuff in /etc is stuff the admin puts there (and presumably can clean up themselves). Also, just about all the config file templates can be overrided on a per-line basis so that what you put in /etc is the minimum necessary to do the job.

Shut up liar

[sproketboy]

The performance of Windoze is designed to degrade over time. IT'S BY DESIGN!

MS desperation

[Anonymous Coward]

Wow, a slashvertisment telling us that next version of Windows will finally work. Really, we mean it this time, promise. What do we get with Windows 10? These are the titillating bits so far: the start menu will come back, and it won't suck donkey balls if you keep using it for more than a few months. Wow. I can hardly wait.

Re:LOL. You expect MS to fix the problem ...

[Jeremi]

Registry bloat is not a problem, it's clueless users who cannot maintain their system.

In other words, it's a problem. A solution that requires all users to have technical knowledge isn't a solution, it's a fantasy.

Take only what you need to survive

[Deathlizard]

It's not still a thing. I can make a case that it hasn't been a thing since Windows XP.

I've had systems with Windows XP on them that have been fine for years without any major issues. The secret is simple, and it comes from SpaceBalls Nonetheless.

Take only what you need to survive

Install what you need and that's it. If you start installing and uninstalling everything under the sun of course it's going to get slower because you're filling up the system with bootup garbage or adware you probably don't need, and it's probably not going to remove itself cleanly because you're relying on some Monkey in a Code Zoo that doesn't know how an MSI Installer works to make a clean uninstaller.

FrankenWindows

[Anonymous Coward]

Seriously? With every version since XP being buggier and more spaghetti-coded than the former, you really think Microsoft is capable of fixing this problem?

Their software just keeps filling up with cruft. They need to start over completely from scratch.

OS-provided structured sharing mechanisms

[tepples]

and you can say good buy to.

your email app from being able to see any office files.

When you drag from the file manager or Office to your e-mail program, the operating system would make a read-only view of the file in the e-mail program's space.

no more flash , java , quicktime and more on the web.

Three words: good fcuking riddance.

adobe apps can't work with each other

Having been published under the same private key (that of Adobe Systems) would let them run in the same sandbox, if the model you envision is anything like the model of Android.

no more visual pinball working with pinmame (at least both are open source and can have both join to one app)

They could join through more structured data sharing mechanisms, such as local sockets set up by firing an intent.

IDE / codeing apps may be come hard to do.

Console gamers and iOS fans would say "good riddance". But they could work the same way AIDE does on Android.

No more NV or ATI driver apps

Again, having been published under the same private key (that of NVIDIA or AMD) would let them run in the same sandbox.

games can't have mods or map editors.

Console gamers and iOS fans would say "good riddance", as I described in another article, because modding helps cheating. But a mod could be installable through the same share mechanism I mentioned above.

CCleaner, anyone?

[drkvogel4600]

Anyone who's stuck with Windows and is vaguely intelligent will surely know about CCleaner which knows where all that junk lives and gets rid of it, fixes registry problems, makes it easier to remove unneeded services etc... I've restored several crippled Windows machines back to full health just by running CCleaner with default settings. If Microsoft had any decency they would buy Piriform and make it a standard feature; Microsoft Unfuck or something like that.

Shake the bad bits out of the cable

[tepples]

The one kernel of truth in your joke is that sometimes you do have to reorient the Wi-Fi AP to improve signal coverage. And sometimes you have to unplug and reseat cables in order to get devices into an operable state. One phone rep recommended disconnecting the USB line and "shaking the bad bits out".

Re:Shake the bad bits out of the cable

[omnichad]

You have to get creative to try to get a consumer to check both ends of a USB cable to make sure it's actually plugged in.

IDE for tablets

[tepples]

and everything will be iOS/Android

Good luck running Visual Studio or Xcode on one of those. But I will grant that Android has AIDE, which might sub for Eclipse in a pinch.

Reduced functionality mode

[tepples]

Trial software can work around this by going into reduced functionality mode until it verifies payment. Much already does, such as the "lite" versions of apps on Android and iOS. So does Doom, which comes with only the first episode unless you register.

DRM in game servers

[tepples]

Your server isn't getting games installed on it, which put all kinds of settings in the registry, then removed later when the game is old and tired, leaving behind cruft (including DRM bullsit) in the registry.

You'd be surprised at how many games have the same digital restrictions management BS in the game's dedicated server app that they have on the client.

Mirekusoft Install Monitor

[Anonymous Coward]

Mirekusoft Install Monitor (www.mirekusoft.com) was created specifically to address that situation.

LOL. You expect MS to fix the problem ...

[Anonymous Coward]

If you had any clue about how databases work you'd know that there's absolutely no problem with having useless stuff in the registry.

Because guess what, Windows DOES NOT READ THE ENTIRE FILE !!!

It accesses the required fields, like a database, doesn't touch the rest, caches entries when needed, ...

But then, that "bloat" problem doesn't actually truly exist. We're not in Windows 95 or even XP times anymore, that was 12 years and more years ago.

Windows on 32 GB?

[tepples]

Or, just join the rest of us in 2014, get an SSD and don't worry about it.

Windows itself takes the majority of the 32 GB SSD that comes with a Windows 8.1 tablet such as a Transformer Book, Aspire Switch, or older Surface Pro. I imagine that most people don't want to have to manually shuffle data between such a small internal SSD and an external HDD.

Re:Windows on 32 GB?

[exomondo]

He said get an SSD, so if you're going to do that get one bigger than 32GB.

Portable apps

[giorgist]

Has Microsoft discovered portable apps ? When I delete it, they are GONE. In fact I can have multiple copies and launched different version. If Microsoft had an elegant sand boxing system such as in mobiles, you are home and hosed.

Just have a vm

[Anonymous Coward]

It might be better but crap software on top of crap software on top of spyware that sneaks on, which never gets fully cleaned up, all the slowness is error handling to keep the os from crashing or actually doing someone it has to jump through hoops to perform.

Want your computer to run fast? leave it barebone and stable. Install a vm u can toss anytime without worry and use that for your dirty work...

It's like your stereo...

[jpellino]

brand new, it sounds new, different, impressive, then six months down the line you ears get used to it and you think something happened to the gear. I expect OS slowdown is a lot like this.

Lol... if you have this problem

[Anonymous Coward]

then you suck at windows.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

You certainly don't understand it. I bet you're one of those idiots that actually believes "Registry Cleaners" provide some kind of measurable benefit. Newsflash: They don't.

What "OS Decay"?

[Torp]

It's not a generic term. Call it "Windows Decay" by its proper name. I don't know of any other operating systems that have this problem.

Will they fix explorer.exe GDI leaks?

[Anonymous Coward]

Will they fix the explorer.exe's GDI leak that's been around for ages?
Or the fact that sometimes I cannot paste between Word and Outlook and must use mspaint as a go between?
There are countless niggles they do not fix

Re:Windows 7 gaming rig

[DocSavage64109]

Yep, I just ran "C:\Users\Dzov>CMD /K WMIC OS GET InstallDate" and got an install date of some 5 years ago:" InstallDate 20090808155049.000000-300"

I don't even think that was on the same motherboard and cpu (second gen i5), but I'm not really sure. I don't know if it's the same speed or not, but it runs fine.

Re:device drive congestion

[DocSavage64109]

Maybe he only has 512MB of ram and antivirus running. That'll cause a fair bit of hd thrashing.

Sad... this was largely addressed in Windows 8

[Aryeh Goretsky]

Hello,

What NetworkWorld freelancer Andy Patrizio complains about, cruft or OS decay, in the RTFA was largely addressed by Microsoft in Windows 8.

Microsoft worked in depth with silicon developers (i.e., the folks who make chips/chipsets for various things that require drivers like motherboards, videocards, network adapters and so forth) as well as software developers that used drivers (anti-malware, encryption, backup and so forth) to ensure not just that installation and removal went smoothly, but that performance was within acceptable levels, which in particular had been a problem for some of the bloatier anti-malware programs often seen pre-loaded onto consumer-targeted PCs, not just during startup and shutdown, but also during common day-to-day activities.

Since Mr. Patrizio didn't bother to use Windows 8 for any length of time, though, he didn't find out about the performance improvements, which, I suppose, is why we are commenting on his rather sad polemic.

Regards

Aryeh Goretsky

Blah Blah, i know nothing of computers.

[Anonymous Coward]

Why does a complete computer illiterate get their rantings published on /.?

Experience proves its a problem, so why?

[Anonymous Coward]

Lots of people here are saying, just uninstall all the cruft that is causing this. But that doesn't seem to work.

I have absolutely experienced this as a computer repair man. I never saw this kind of issue on my MacOS 9, MacOSX jobs. Why did this happen in Windows 95,98,ME,XP,Vista,7. I would go though and uninstall everything, see only a little improvement. Then end up reinstalling the whole OS and reapplying all the updates to get a snappy system.

This IS a problem. Its not user induced because MacOS doesn't seem to experience this.

So barring dumb users....Why does this happen to only Windows?

Re:Reading all the comments 'defending' Windows ..

[TechyImmigrant]

>The last 50 years of research have no impact at all at current day computing, except for silicon and SOCs etc.

This.

Doing new stuff in silicon is like shooting fish in a barrel. You just go find some research paper that has a good idea, then go and design it in hardware. I little hard math is enough to ensure that no one else is doing it.

Software seems stuck in a research time warp. Plan 9 was how long ago? Yet we don't have network transparent multi processing and per process name spaces in popular OSs.

Solutions

[River of Souls]

This is a real problem, I have experienced it myself on my computers and at work where I work in IT on some 4,000+ PCs. The help desk probably re-images a dozen machines a week partly due to this problem. I know for me it is often because I am downloading various interesting software packages, and then I am too lazy to uninstall them. And a large percentage of software does not uninstall cleanly - not really a Microsoft problem there, not completely.

Personally I kind of like Windows 8.1, and I really only think it was a marketing flub to try an force a touch interface on people. Bad Microsoft, no biscuit. But I digress...

* On some machines we use a product called Citrix Provisioning Server (used to be Ardence), booting the machines off a network read-only drive, and we have other software that saves select important user settings and files. We refer to this as "stateless" and is the closest thing you get to being immune to this problem. Unless you really have skillz and screw up the master image this is based on. This has been the "Always runs like new" experience for us.

Other ways to achieve a similar effect:
Use virtualization
* Windows 8.1 includes a FULL version of Hyper-V, a type I hyper-visor that is fast (you could use others as well of course). Basically, install Windows twice, one being the host and put nothing on there but the guest. Then immediately make a snapshot of the guest. Use that VM for web surfing or any activity that will introduce cruft, etc., and you can always revert to the snapshot and be pristine one again (of course you will need to do updates again, re-install software, etc.). This also would let you use Win7 as the guest, if you like that OS better. XenClient Enterprise is another nice one but it costs money (no, I don't work for Citrix, but I am a Citrix admin). Oh, and although this is similar in effect to backing up with an image, it is much faster and you don't have to buy something like Acronis (although it is nice). I can't recommend things like VMware Workstation, VMware Fusion, Parallels, etc. because Type II hyper-visors like these cause a big performance hit for everyday use, especially if you have invested in a nice machine and want to take advantage of it. These have good utility for other things though.

Make your own thin client
* If you have access to Microsoft Enterprise licensing, you can use the ThinPC version of Windows, which is made to turn a PC into a thin client and includes a write filter. Such a machine will not retain anything on reboot. So you would need a way to save data / settings elsewhere. But, you can turn the write filter off to install things permanently and then turn it on again. Effectively making an "appliance" with the apps you need, but doesn't really get slow over time (at least no where near as much or as fast). Great way to test things.

Microsoft did include a "refresh your PC" built right into Windows 8.1, but I will admit I have not tried it myself yet.

It's called the Windows Store

[Anonymous Coward]

It's called the Windows Store. It's part of Windows 8 and newer and applications in it install and uninstall exactly the way you think they should. They built it and gave it to you and no one showed up.

Three steps folks...no decay.

[Nocode5000]

1) CCleaner - it just works, clean the reg too....seriously its the only reg cleaner I use. 2) Autoruns - Dragons here! so really only tweak things on the login tab unless you uber l33t 3) Defrag - optional on your 1TB drive with 970GB free -- Close contenders: add/remove progs, msconfig, PC optimser pro....yeah right

apps

[Anonymous Coward]

You dont add and remove apps, this isnt apple.

We call them Applications, Programs and executables.

just use the old ones...

[bumba2014]

We have a few web-servers, the new ones come automatically with Windows 2012 version, there are terrible, everything is put somewhere else. And a lot of stuff doesn't work, because of some new security measurements... It's no fun working with them...

system

[Tom]

As you add and remove apps, as Windows writes more and more temporary and junk files, over time, a system just slows down.

Yeah, it's a damn hard problem to solve. No surprise it's taken them 20 years to figure out that you could just put all of the files that belong to one application into a few folders exclusive to that application and then wipe them when the app is removed. Instead of, say, the absolute dumbest thing you can do, which is scattering them all over the place without keeping a record so you are absolutely guaranteed to never, ever, find them again.

Re:system

[mrprogrammerman]

If you want to keep track of what gets installed where check out Install Monitor (www.mirekusoft.com). It's designed to solve that exact problem.

Re:system

[Tom]

Which basically means that Microsoft is selling a car without mirrors, brakes and seatbelts, but you have plenty of options buying one from a 3rd party. Provided you remember to do it...

Re:system

[Anonymous Coward]

Instead of, say, the absolute dumbest thing you can do, which is scattering them all over the place without keeping a record so you are absolutely guaranteed to never, ever, find them again.

Properly written installers don't do that, even installers written before MSI was introduced, e.g. those written with InstallShield 5 back in the NT3.51 period. That said, there were lots of improperly written installers, especially from small shops and shareware authors.

- T

Solution to the problem of OS Decay?

[lippydude]

Ubuntu 3D Desktop

Ubuntu Demo

Distro Watch

Windows 8.1 is garbage!

[Anonymous Coward]

It doesn't even seem to allow playing solataire without being subjugated to being harrassed to send payment to M$ or some other obscene corporate entity. My poor mother is begging me to install Ubuntu on the laptop I bought her when the HD on her desktop died. She just wants to check her email, view Facebook, and play Solatire. After seeing that Ubuntu has morphed into a not very differntiated clone of Win8 I'm about at the end of my rope to find a useable non-money sucking system for her.

Do you mean

[Chrisq]

Do you mean that there are people who don't completely reinstall their computer with a different OS every few months, struggle to get all the devices working, then when its finally stable see some other distribution that looks worth a try?

Same q as "is this feature an MS talking point"

[CptJeanLuc]

If it's not part of the Win 10 sales pitch, then my 99% confidence answer is no. They are not going to invest in some feature that no customers will know about (because noone is talking about it).

Plus, as someone else pointed out, it may very well be part of the master plan, that system rot leads to users wanting to upgrade. And there definitely is some system rot going on - someone wrote in another post that this is no longer a problem since Win 7, but I have experienced this countless times with Win 7 over the years, with just standard installs without any fancy tinkering on my part.

Rubbish

[ledow]

No. It doesn't. It hasn't for years. I had a WinXP image that followed me for 8 years and never got reinstalled. It had any amount of stuff installed over it and was in constant daily use (took it to work, worked all day on it, brought it home, played games all evening on it).

It's not an inherent property of Windows that it "slows down" or any such nonsense. If you ask it to run 20 services on startup, it will be slower than if you ask it to run 10. It's a given. The trick is to make sure that NOTHING IS RUNNING unless it needs to be.

Computers DO NOT GET SLOWER WITH AGE. They are the same speed to within MILLIONTHS of a second. If you ask them to do more then, yes, they will seem slower. Don't ask them to do more - remove unwanted programs but most importantly do NOT let things run on startup or in the background unless they are vital. Hint: Almost nothing is vital. QuickTime does not need to be in your startup. Java does not need it's QuickStarter. Adobe stuff needs NOTHING running in the background. And so on.

Do that, and the computer does not slow down at all. I have an 8-year-old XP image to prove it until I stopped using it a couple of years ago (and not because it was slow - because I was managing Windows 7/8 networks).

If you manage your machine properly (and you're working in IT if you post here, I assume), it does not happen. And it's no more a burden than having to reinstall everything after a format. I have NEVER formatted a machine to clean it. I've gone back to known-good images on work machines, but those images have histories going back years too - but kept PRISTINE so they could re-image nicely.

If you format, as far as I'm concerned, that's a harder version of the "reboot will fix it" mantra. A total cop-out. I have brought machines back from the dead (five minutes to get to Windows logon, down to 45 seconds on the same hardware) by proper management of the machine and pruning only third-party services and junk on startup.

Stop making excuses and doing the "Microsoft-fix". Manage your machine properly and it's never an issue.

Wife Aggro

[MeesterCat]

The only time I ever experience this issue is when my wife has decided not to use her laptop and has instead decided to use my carefully tended to machine.

I clicked the 10000 extra smiley button

[Tyr07]

I have found work systems that run like the day they were installed, with software added and removed over years.

Proper software being installed and removed has no problems, and doesn't slow the system down.
However, people who are fucking absent minded, install any software they feel like because the smiley banner said so, then try to remove it, run into fucking problems.
Operating systems don't decay.
People put stupid shit on them.

It's like a garage, you kept sticking shit into it, removing a few things, putting some others back but until you bother to properly reorganize it or clean it out, it gets cluttered.

There are a few people whose garage is sacred ground, they make sure to only store what things are worthy of their garage, and it's always immaculate.

So I'm going to rephrase this topic

"Will Windows 10 finally babysit me so I can do everything I want while pretending to wear the big boy pants"

No, probably not. But fortunately with more app based windows sanctified only software, one day, you might have that.

FUCKING IDIOTS

[Anonymous Coward]

Microsoft spent a lot of time and money engineering system slow downs.

They had board meetings that weighted just how small a change, what setting of search indexing, could be engineered to break.

Proof: Last update to Windows XP was to fix a deliberate system slowdown that was added to Windows XP - that makes the system crawl. They had to remove it before the end of the service period so as not to draw attention (after, it would draw more attention) and to remove it so in ten years nobody would even think that for many many years people were running on sabotaged OSes.

Windows XP: Search is the feature!

Windows Vista: Windows XP's search is shit! Use this!

Windows 7 (is that right?): Windows Vista's search is shit! Use this!

Windows 8: LOOOOOOOOOL we're just fucking with you, in a plan that was hatched ten years ago, push out this, then skip 9, then 10 to make people feel they have to pay $100 FUCKING DOLLARS for an OS. What bullshit.

Windows 9: LOOK SEARCH WORKS! We took the start menu, made it fullscreen....theeeeen shrunk it back down again! buy this shit like a mindless fucking cunt that you are.

Evidence

[maestroX]

OS decay really impacted Windows XP performance, which can be reproduced by installing initial XP, SP1 and SP3 from CD.

• IE6 notoriously slowed down browsing with large cache (>500Mb).
• SP1 performance/stability issues with drivers, requiring more memory
• SP3 requiring more memory still

This is without crapware, anti virus using only Microsoft drivers.

Re:Evidence

[Anonymous Coward]

OK, that's your anecdote. Here's mine: This is being typed on a WinXP system which was upgraded from an OEM Win2000 installation - sans crapware, which was actually an option back then. A lot of modern software is slow on it, but that's because the software expects to be running on something modern. I have never had an OS re-installation. I have had problems with bad video drivers and one bad NIC driver, but the system is still very close to its original feel after nearly 15 years.

How can I be sure? I still have the Win2000 image from immediately before performing the upgrade to XP SP2. Recently, I had to restore that to use software that doesn't work on XP (at least, I couldn't get it to run properly) for a client. While running Win2000, I noticed that the overall performance wasn't appreciably better than with XP. I also tried some older dev software (C++Builder 5, Delphi 7, and so on) so I could compare those later with the restored current XP image; they run only slightly slower on the current XP image.

The only time this XP system got slow across the board was after installing a software licensing library required for a client's project. That would bring the system to a halt at unpredictable times for anywhere from a few seconds to a couple of minutes. After finally removing that stuff (client has moved on), the system is back to being as snappy as a 14-year old system can be.

Yes, I expect to replace it soon, and I will probably convert it to a VMWare image on the new system.

- T

Get an SSD...

[Anonymous Coward]

... and you could forget about OS decay.

No, no.

[Anonymous Coward]

You need to swab your RAM and empty the bitbucket.

While you're doing that, you might as well top off your monitor with liquid crystal. It's not related, but since you're doing maintenance anyway, you might as well.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

I used to work for a company making USB devices. Every time I plugged in new hardware (and sometimes after I reflashed its firmware) Windows correctly detected it, discovered the driver needed, and stored the association in the registry. That caused quite a bit of registry growth over time, naturally. But the registry is a database. O(N log N) doesn't cause slowdowns. And the practical result was that detecting the same hardware was a lot faster on the second try.

BTW, the registry isn't a single binary config file. It's already partitioned. But you don't need to care how exactly - the API is stable and unified so details do vary between Windows releases. As for bloat, well, having security on every level is not free. But text is much more bloated than binary to start with.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

You have no fucking idea what you're talking about.

Re:LOL. You expect MS to fix the problem ...

[AmiMoJo]

The registry is a database. It is designed to store vast numbers of keys in a hierarchy, and a default install of Windows will have tens of thousands of them. Even the most bloated apps will only add a fraction of a percent to that. Performance of the registry really isn't the issue.

Performance problems come from app that hook in to other things, particularly explorer. That is done via the registry, but the registry itself isn't the problem. For example, if you install Adobe Reader it sets up a DLL with a hook that makes Explorer load it in order to provide thumbnail previews of PDF files. This slows the machine down because now Explorer takes longer to load, uses more memory and executes extra code when a PDF file is displayed as a thumbnail.

No.

[StikyPad]

The answer is no. Really, your question could have ended with anything after "Will Windows N finally address," and the answer would have been the same. Windows 10 will not finally address anything.

Next question.

Re:LOL. You expect MS to fix the problem ...

[Spacelord]

You forgot .local, .gconf, .gnome, .gnome2, .kde

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Of course, you realize that the reason that Linux has not caught on as a desktop environment is that it requires too much Linux background knowledge to run/maintain a Linux system? If requiring system-specific knowledge is a mark against an O/S, Linux has a long ways to go to be a viable O/S.

No, but clearly the OS is.

[Overzeetop]

The OS should be doing all of this automatically.

And, while I'm at it, WTF is up with keeping a pagefile in 2014? I have 24GB of RAM and turn the pagefile off. I've never seen my commit above 13GB, even with multiple RAM heavy programs running. (FWIW, I've never encountered a PF problem in 5 years of running without one)

Really?

[Anonymous Coward]

One program: Revo Uninstaller.

We recently had issues with STK at work. A multi-thousand dollar (with our toolboxess) software program just stopped working. We tried an uninstall-reboot-reinstall. Still didn't fix the problem.

So we ran Revo Uninstaller on it. Revo creates a system restore point, does a quick scan for the program, runs the application uninstaller, then scours the disk and registry for "cruft" left behind. For STK, that amounted to over a thousand files and roughly a hundred thousand assorted registry entries left behind. Revo then prompted if we wanted to remove them.... of course we did. After that we reinstalled and worked as new.

A simple "uninstall" just won't cut it these days. Application developers are lazy and don't clean up after themselves.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

You have missed Dconf didn't You? Gnome developers decided that single bloated binary config file is exactly what was missing in linux desktops.

Hummmm

[Anonymous Coward]

ok, of course if you install and uninstall on a daily basis, your machine is going to be crap, Temporary files are not deleted correctly to my satisfaction even when you purge the temp folders of the user profile and the machine, there is always stuff in there that needs to be removed manually.

Defrag is still needed and windows 7 left the defrag on scehdule on a daily basis so your HD are always compact BUT, the decay will still be there because of your hard drive degradation whether you clean your machine or not, it's a bunch of factors that no OS can circumvent unless you have a SSD HD or maybe have an OS that moves your file all over when sectors gets corrupted.

I have Windows 7 and it works perfectly and havent reinstalled over 3 years and it runs 24/7

Sure, malware was a factor...

[Anonymous Coward]

...but the main reason Windows XP got slower over time was because Windows Update was designed to slow down exponentially as Microsoft released more updates. After they finally fixed that, a Windows XP install works just fine after 10 years of proper use.

Re:LOL. You expect MS to fix the problem ...

[Anonymous Coward]

Yes... no single bloated binary config file, instead, at least three of them: gconf, dconf, xconf

It isn't decay...

[tsqr]

It's waxy yellow dll buildup.

*shrugs*

[Mysticalfruit]

Our house went windows free years ago and hasn't looked back.

Viva la Linux.

Decay? What decay?

[Anonymous Coward]

Windows still stores EVERY SINGLE update in your windows folder so it can be removed/restored at any moment. This clutters up the HDD way too much. But the OS gettting slower has nothing to do with that. As some others point out it isn't due to the OS itself getting slower it's how you operate the OS. Maintaining startup, services and some other factors can lead to a perfectly running machine several years down the line.

Breaking the cycle?

[Anonymous Coward]

ME - bad
XP - good
Vista - bad
7 - good
8 - bad

So MS have finally decided to break out of that good-bad-good-bad cycle by skipping version 9?

Re:Breaking the cycle?

[Daniel Klugh]

"XP" was the sequel to "2000". "ME" is regular MS-Windows, "XP" is WinNT V5.1.

Every OS sucks

[ilsaloving]

Just to save everyone a whole lot of time and energy, and basically summarize every argument that is going to appear for this article:

https://www.youtube.com/watch?...

Put OS on a separate drive and image backups

[kimgkimg]

This is why I put my OS on a separate drive and take an image of that drive when I have everything set up the way I want it set up. This allows me to freshly start from that point at a future date without having to re-install everything from scratch.

Ever used Sandboxie?

[mrjimorg]

There are situations where your being held hostage - if you want xyz then you have to download and run this app. If you need xyz then your forced to comprimise the integrity of your system. This shouldn't be the case. Sandboxie has a system in which you can run those apps in a box, then choose what you want to bring over from that box (xyz) without bringing over everything (bad virus). I used this in order to get Minecraft mods for my son from evil websites like iLivid.
If sandboxie-like functionality was built into the core OS then it's a simple matter to fix issues on grandma's computer when she installs EVILVIRUS in order to get a recipe. If the gui is clean and clear then grandma can fix it herself. It also presents a clean system to the sandboxed app so the virus can't steal your private info.
If Microsoft had any sense "if(0)" then they would use this technology along with the containers idea from docker as their future installation mechanism. This also make differential backups much easier. It win win win all over the place.

I have to format Linux boxes more often

[Anonymous Coward]

Really, it is a matter of how proficient you are with the OS. I often break my Linux box that I use at work when trying to get something more abnormal working, I know that if I had the skills and knowledge I could fix my Linux box as easily as I fix my Windows box at home, but that is just not the case. The internals of a GNU/Linux system is a lot harder to understand than Windows and varies a lot by distribution.

For example I broke my Linux laptop hotspot some time ago and I can not get it working again, I looked around the net and it seems I need to downgrade some bizarre network package I never heard about, I will just format the box and get it working again.

Now Linux is a lot harder to break in the first place, but when it does I usually just format it (which is a lot less painful than formatting a Windows boxes because all applications keep their stuff in my home folder). The problem is that in Linux you run into the abnormal situations far more often than in Windows. On the other hand in Windows you run into Spyware and Crapware far more easily.

It's because there isn't clean system/user demarc

[Anonymous Coward]

Whether you see this depends on how observant you are, how much stuff you install/deinstall, how badly those programs overlap their shared objects (typically dynamic link libraries, which is what MS calls run-time shareables on Windows), and how overpowered your system is.

It's 100% a fact, though. The problem is that historically windows was architected as a personal computer operating system - designed from the ground up for a single-tasking singer user with no external connections. As Windows has matured into a fully networkable system that allows sharing parts of the system with other users (and implemented true multitasking and service daemons,so that some processes are not actually user processes) many of the bad design decisions have been re-engineered out of existence, but several of them are still hanging around causing problems. The worst of these is that Microsoft decided early on to let user applications install code into the C:/windows directory. This created the infamous ".dll hell" as well as system decay.

In a system like MacOS or Darwin or VMS or a well-architected Unix implementation, all the files belonging to a user application live in a known place, and other applications that might depend on the first (say, add-ons or plug-ins for a spreadsheet app) all look in that place to find anything they might need. In Windows, part of the application live in their own poorly defined places (obscure registry locations and oddly named folders) and parts are loaded into the system folders. In later versions of windows, these places have gotten better defined (for example, app folders are mostly, usually in C:/Program Files instead of just some randomly named place) but there's still the problem of .dll files that are expected to be shared by applications.

Example: You load a program to balance your checkbook, and it puts a file named "BlagC.DLL" in your C:\windows\system folder. Then you load the ButtWiper3000 program to monitor toilet bowl cleanliness, and it wants to also load a DLL with the same name, because both these programs were built from the Blagovich C compiler, which always generates this DLL. The version that came with ButtWiper3000 is newer, so it overwrites the old one. So if you uninstalled ButtWiper3000, it can't uninstall completely, or you'd be unable to balance your checkbook. So none of these programs will completely uninstall, for fear of breaking something else, and cruft accumulates. This is just one of the many bad effects of having no strongly enforced application code isolation ("sandboxing" if you prefer).

If you install a hundred crappy apps, and then uninstall 95 of them, your system will slow down. Will you notice? Not necessarily! You might be ideologically committed to not noticing, or the slowdown might be in your network processing which was already bottlenecked by a crappy Internet connection, or the slowdown might be in your core processing, and you were already bottlenecked by your crappy disk drives, or (most likely) the slowdown will happen so gradually over such a long period of time that you could not possibly tell without reinstalling, or (second most likely) you will buy a new system before you finish slowing this one down enough to notice.

The situation has gotten better with every release of windows since Vista, and will probably continue to get better. It's already possible to install a few commerical applications without special privileges - this indicates that they are properly isolated from the rest of the system.

Re:LOL. You expect MS to fix the problem ...

[Lumpy]

/opt/app is not acceptable either. /etc

  $HOME/.config (I'd rather it be called etc to be consistent) is fine but /opt is just bullshit.

two places, no more. having to go hunting for crap is one of the biggest problems with linux.
and it has always been there. Gnome X11 and a lot of others crap in $HOME/ And it's all because the developers are lazy and need to be beaten with a sack of doorknobs.

Can't back up applications?!?!

[hoggoth]

> another backup, disk wipe, and reinstall.

The biggest problem for me is there is no way to backup and restore your installed applications! The 6 month shuffle goes more like: backup data, disk wipe, reinstall OS, reinstall every single application you use finding all of the serial numbers and resetting all of your preferences, restore data.

WTF?!

What I have to do

[Bumbledum]

OK I get no slow down over time, BUT and there's a BIG BUT, I have multiple apps for maintenance installed. I have a LOT of apps on my pc for my music and video and photo processing equpiment and studio. I also run games at times. Once a week though I look for any new crap ware that's been snuck on and remove it, then i do a check for any windows updates that are optional and decide whether or not to install. I then run driver and program updates with THREE driver update check programs (yes they each catch things the others miss) and three programs that check for software updates in addition to the self updating programs ( and again things are picked up). I then run registry cleaners, and an error check program that cleans out temp files, active X and .com errors, etc. I follow that with a full virus scan and then a full disk scan. Reboot and a chkdsk on reboot. I also run diskeeper and keep my files consolidated. Yes, it takes a couple hours a week, but I haven't had any slowdown in two and a half years. Pain in the ass, but to me worth it for productivity. Tuck

7 year old XP build

[whipnet]

I have an original XP build laptop that I have used every single day since 2007 and it ran like a champ until MS recently dropped XP from support. Strange how now it's started running like crap. Coincidence? I think not.

Sigh.. another computer novice writes an article

[Anonymous Coward]

Okay... There is no such thing as "OS Decay". The reason a windows based PC slows down over time is because of a crappy file that keeps the settings for almost EVERYTHING in or on the OS called the "registry". It's a crappy invention from MS that should have NEVER been invented! That along with other drivers and 3rd party files that get loaded at start-up slow down your PC. The solution? Re-load the OS. That has been the answer since the Windows 95 days, and as long as MS uses a registry, it will continue to be the solution for years to come, my young one.

The real solution is to come up with an OS from the ground up that needs no more than 1 or 2 files to boot up, and modules, or "extensions" that add functionality and that are user installable at the click and drag of a mouse. OS is slow? drag the extension to the "disabled" folder and restart. Oh wait... This sounds familiar... That's right, it was already invented and it was called Mac OS 1.x-9.2.2. But Steve Jobs killed it off in favor of an OS that REALLY suffers "Decay" called BSD UNIX with a GUI over top called Mac OS-X.

Sarcastic, but 100% truth :-/