More news on Windows 8. This time around, Gabe Aul, a director of program management in Windows, blogged about the changes Microsoft has made to Windows 8’s boot process. The results are impressive – a boot time not much slower than waking from sleep on current Windows 7 and Mac OS X machines. This is, of course, a vital component of getting Windows NT ready for tablets.
Windows 7 can already boot pretty fast as it is – I never timed it, but on my SSD-equipped workstation, booting Windows 7 is pretty fast. Of course, the BIOS takes a lot longer to load, and for some reason, ever since I installed the SSD, I get another detecting screen thingie which adds another few seconds to the process. However, the loading of Windows 7 (or Ubuntu 11.04 for that matter, even though Ubuntu is on a regular hard drive) is pretty damn fast.
But, this process can still be made faster. You might be wondering at this point – why bother? Doesn’t everybody sleep and hibernate by now? Well, as data gathered by Microsoft suggests – not really. Most people still prefer full and clean reboots and shutdown/boot cycles. They found out that 57% of desktop PC users and 45% of laptop users shutdown their machines instead of sleeping or hibernating them. Common reasons cited by users is that they don’t want their computers sleeping since it still draws power, draining the battery and upping energy usage. On top of that, many people prefer the idea of a ‘fresh start’.
So, Microsoft still set about to improve boot times in Windows 8, and they did so in a fairly clever way. Most likely due to the componentisation of the lower levels of Windows NT, Microsoft can now actually hibernate only the kernel. In Windows 7 and earlier, when you initiate a shutdown, all user sessions are closed, and in the kernel session, all services and drivers are closed so you have a complete shutdown.
In Windows 8, things are done differently. All user-related stuff is still properly shutdown like before, but everything related to the kernel goes into hibernation. “Now here’s the key difference for Windows 8: as in Windows 7, we close the user sessions, but instead of closing the kernel session, we hibernate it,” Aul explains, “Compared to a full hibernate, which includes a lot of memory pages in use by apps, session 0 hibernation data is much smaller, which takes substantially less time to write to disk.”
It also takes substantially less time to boot. “It’s faster because resuming the hibernated system session is comparatively less work than doing a full system initialization, but it’s also faster because we added a new multi-phase resume capability, which is able to use all of the cores in a multi-core system in parallel, to split the work of reading from the hiberfile and decompressing the contents,” Aul adds, “For those of you who prefer hibernating, this also results in faster resumes from hibernate as well.”
Microsoft also uploaded a video of the boot process of a laptop running Windows 8, but they do not detail the specifications of this machine. It most likely has an SSD, and they note it has UEFI as well. My new media centre computer has UEFI as well, and it sure rocks by a heck of a lot faster than BIOS.
While this all looks fine and dandy, the real tests will come outside of controlled environments. The diagram above does list the time between winlogon and desktop-ready, but this period will be different for everyone, depending on how much crap you have installed.
For whatever it’s worth, it’s an interesting way to improve boot times.
Windows 8 is turning out to be a very interesting OS after all. I do quite enjoy using Windows 7 and I also use Linux at work exclusively. Windows 8 has got my interest piqued!
Meh.. just add WinFS and everything else that sounded good that was left out last time a-round…
M$ should just look pretty, do not much and live off the shares it has in Apple, like Stevie baby….
POST DELETED
Edited 2011-09-09 17:25 UTC
All their doing here is making shutdown the equivalent of a partial hibernate.
No doubt they will now have …
– Sleep
– Hibernate
– Shut Down
– Shut Down and Reset
So what will happen here is that most installs will still do a reset and require a ‘full’ restart and the only time you really get this ‘fast’ boot is when you shut down your computer for the night and restart it in the morning.
It also looks like the address space of services is restored so problematic memory gobbling services won’t get reset in a regular restart.
Edited 2011-09-09 18:12 UTC
Exactly. And this gets at the primary issue with Windows 7 (and Windows Vista before it): regardless of how fast the OS is “in use”, when you have to wait between 10 and 60 minutes for it to update itself every few days, when all you want to do is turn it on (or off) and start (stop) working, all that theoretical speed is negated from the end-user perspective. This is one of the reasons people love the iPad (and to a lesser extent, the Mac)–it’s generally hassle-free, while updates are fast, easy, not forced down your throat while you’re trying to work, and most important, few and far-between. Turning on/off a computer, starting the task you want to start, whatever it may be, should not be an ordeal, it should be instant, like an appliance, like a TV. As long as Microsoft fails to understand this, Windows will continue to lose market share to all manner of competitors, iOS and Android included.
Oh, and don’t get me started on driver installation. Why is it that on Mac and Linux, you can plug in any old mouse, any old keyboard, any old thumb drive, any old memory card reader, (almost) any old printer or camera, and it instantly “just works” while on Windows it takes at least 30 seconds for the first time plugging in each new device, sometimes even multiple times per device depending on which USB port you plug it into!
Not sure why these comments were voted down… I speak from experience, that I’m sure many have shared… whatever, seems pretty obvious that there are some MS astroturfers roaming the lands…
60 minutes for updates? Even 10 minutes is horribly long.
You should get an SSD. I think the longest Windows 7 update process I’ve experienced since going SSD is 3 minutes.
“This is, of course, a vital component of getting Windows NT ready for tablets.”
Why?
I have only rebooted my tablet on updates… since last Xmas. A fast boot is not really necessary for tablets.
KRR
Yep…. The main concern for a tablet OS is being able to wake up from sleep instantly and have a very fast resume-sleep cycle., while giving the apps enough time to do something useful…
it’s vital because it will need frequent reboot, and sometimes it will reboot by itself…
“Frequent reboot”?
Maybe once a month if there is an OS level update every month.
The last time I had to reboot Windows 7 for a non-update reason was when my antivirus’s web scanner got wedged and wouldn’t let me load pages.
It did bluescreen maybe two months ago because of the ATI video drivers.
What frequent reboots are you expecting Windows 8 to require?
I have a under-powered netbook called an Acer Aspire One 522, which has only a 1GHz 64bit dual-core CPU with Radeon HD 6250 graphics and a high-resolution 1280 x 720 pixel display.
http://liliputing.com/2011/03/acer-aspire-one-522-netbook-review.ht…
I have set up this machine to dual-boot Windows 7 (which it came with) and Kubuntu 11.04 (which is my self-installed value-enhancer). Only very occasionally do I bother to boot Windows 7, because it takes soooooo long.
Anyway, the time before last when I had booted Windows 7, the machine advised that updates were available, and recognising that updates on Windows 7 are absolutely crucial to the continued good functioning of the machine, I allowed them to proceed when I shut down the machine. It took ages to shut down, so I had assumed the machine had installed the updates. Silly me.
When I next booted the machine, the updates that were only queued did actually install. I waited, waited and waited for the machine to start, all the while it was telling me … “Do not turn off this machine” or similar message. It took well over an hour, and no less than three re-boots, before I could begin to use the desktop.
Well over an hour to boot, and three reboots required in only two Windows login sessions … that has GOT to be some kind of record doesn’t it?
Meanwhile, on the exact same machine, the maximum time to boot Kubuntu 11.04 has been about 15 seconds.
Edited 2011-09-12 23:09 UTC
Except when you happen to update your *buntu, after which your graphics driver is borked, sound driver needs to be reconfigured and Samba has gone bananas and needs to be tamed again.
Seriously, you sound like pretty precious Linux has NO kernel updates (that require a reboot and maybe recompiling a driver or two), no X updates (that require restarting X, a process equivalent to a reboot for any GUI-using user) and even if these updates happen (which they don’t because it’s Linux, which is by definition perfect) then they happen automagically and don’t bork your KDE session or require a reboot.
Sigh!
I don’t upgrade Linux operating systems in place … instead I replace the entire OS when I transition from one release to the next. I don’t use proprietary drivers, for graphics cards I use either the open source Radeon drivers from Xorg or the Intel drivers from Intel. The Acer Aspire One 522 works beautifully with the open source Radeon driver.
On Linux, when updates are due, one just lets the auto-updater run in the background. It is not required that you stop working.
Occasionally, even for Linux, the kernel or some core component is updated. In this case the system shows a little yellow icon in the system tray, advising that a re-boot is required. At my convenience, I save what I was doing, close down the applications I was using, and perform the requested re-boot.
As I said, this re-boot takes less than 15 seconds for Kubuntu 11.04 on my under-powered Acer Aspire One 522.
I have separated the user’s home partition from the the operating system partition on this machine. When the time comes after October this year, I will probably upgrade the Kubuntu operating system (at my convenience, when the machine is not being used for anything) to Kubuntu 11.10. I will wipe Kubuntu 11.04, re-format the OS partition (but not obviously the user’s home partition), and install Kubuntu 11.10 in the OS partition. The whole operation will take only about 30 minutes.
At no stage will I need to re-compile anything. I will test my system first by booting Kubuntu 11.10 from a USB stick before I commit it to the hard drive, so I will ensure that nothing will be borked.
Edited 2011-09-13 23:58 UTC
I seriously doubt they’ve got boot times down to two seconds, as that’s all my Windows 7 netbook takes to wake from sleep (aka standby). I think what was meant was waking from hibernate, not sleep. I don’t get the boot time obsession in either case. My netbook takes a whopping 15 seconds to boot. Wow, that’s just, like, so very much time to wait. I hardly reboot the thing anyway, when I’m done with it I just close the lid and let it sleep. When I need it again, I flip open the lid and it’s back, so long as I’ve still got battery power in it anyway. And yes, this netbook has a mechanical hd, not an ssd. If I can already boot the thing in 15 seconds, I really don’t see why boot times are so important. It’s not a super-tweaked version of Windows, though I did install Ultimate and get rid of Starter.
It’s important because Android and Apple would use it as a selling point if theirs was better. From what I’ve been able to read, this also will affect the length of time to shutdown the machine and will increase battery life. It’s full of win.
As I understand it, they save the state of all the hardware detection and services. So they seperate the initialization of the hardware in 2 parts ? Instead of searching for hardware, they just initialize it ?
Yes, that’s how I read it as well. It should mean you only need full reboot after changing hardware or drivers
Oh, I hope not. That’s fine and dandy on a tablet, but not on a desktop or netbook. Some laptop and netbook vendors tried this, they call it fastboot or quickboot or some other variant on that name. Do you know how many support calls I’ve gotten when that thing fails to initialize the new ram people install? You’d be amazed how many people think they know what they’re doing and then don’t bother to check the quickboot setting (you have to reset it so it can rescan the hardware). But hey, I guess it’ll keep people like me in business. 🙂
Well, if you have to use sleep instead of proper shutdowns, I guess it shows why boot times matter.
Not really. I use Vista and although shutdowns and cold starts are significantly long (3 to 5 minutes in total depending on whether there is an anti-virus running or not), hibernate/resume is also pretty long (around 2 minutes, sometimes more). For me, the main interest in hibernate compared to full shutdown is that work sessions are carried through several days or weeks.
Sure, but he was talking about a netbook which he keeps putting to sleep (closing the lid), not in hibernation. Recovering in seconds is one crucial advantage of sleep on modern OSs.
It also matters what you consider completing the “boot” process. My current computer puts up a login window in about 15sec and I can login but I can’t use the damn thing for another ~45sec-1min while it finishes the “boot” process and logs me in. If I wait about 2 minutes after the login window shows up and then login, it only takes about 5-10 sec to get to my desktop. This indicates that the OS is still booting even after the login window is presented to the user.
Is there a standard that is used to tell when booting is complete?
KRR
Wow, how much do you have running at startup? There’s no reason it should be taking that long unless you’re running Vista before sp1 or unless you have far too many startup services.
If your primary use of a computer is Facebook/Skype, then you will not get to those services within 30 seconds.
Actually I have 1 second login time on my Windows 7 desktop … pretty simple to achieve.
Windows + R .. then type msconfig … you can disable most unneeded startup items in there. I only run Adobe Flex, SQL Server, Skype and AV.
That is BS! Skype is not Usable in 1 second after it starts.
Oh Comon!
1 second for my desktop to be usable, you know when you have like more than 1 core, more than one programs can loaded at the same time ;-).
Skype is still doing it thing in the background … If I click for Chrome it starts up pretty much instantly. Okay skype isn’t signed it but I can start using my desktop.
(Skype seems to take exactly the same amount of time on my HTC desire as it does on Windows to sign in … I think the majority is network latency)
I just killed skype and tried started it from scratch and it took 5 seconds to sign in … most of this appeared to be network latency while syncing contacts and conversations.
My machine gets a WEI of 6.3 on Windows 7.
For a more “normal machine” I have a Dell D430, WEI on Windows 7 is 2.1 … We are talking about 30-45 seconds from cold boot. I don’t think that is bad for a machine has a 4200rpm iPod hardrive, and a 1.2ghz Core 2 ULV processor, and is running SQL Server 2008 R2 as a service.
Edited 2011-09-10 22:45 UTC
sleep/wakeup:
Windows 7: fast
Fedora 15: pretty fast
Mac OS 10.6: very fast
after wakeup, access to wifi:
Windows 7: very slow
Fedora 15: very slow
Mac OS 10.6: instant access !
so wasup with wifi ?
after wakeup, access to wifi:
Windows 7: very slow
Very fast here. Just tested (regular HP laptop, nothing fancy at all).
Wake up, enter password, click link and it works. No delays at all.
My guess – it depends on manufacturer/drivers, etc.
I think the answer to your question may be here:
http://www.osnews.com/story/24943/Rapid_DHCP_Or_How_Do_Macs_Get_on_…
My netbook running GRML (Debian Sid based) wakes up to the CDM (Console Display Manager) screen in about 2 seconds, and the moment I hit “Enter” after typing my password, I’m already connected to my wifi network ,assuming it’s the same network I was connected to when I put the netbook to sleep.
Just did a quick Sleep/Resume. With 2 seconds to password prompt. I type in my password in 1 second. WiFi was connected at the moment I saw the desktop no more than 3.5 sec after I pressed Fn+F4 (resume)…
Writing from Ubuntu 10.04 on ThinkPad T42(all original parts)
My computer totally blows away your computer!
It resumes from hibernate instantly as I walk towards it. It has a time dilation chip which digitally spins faster than light and is therefore able to go back in time to start resuming from hibernate the instant I think of waking it.
Also my dog is meaner then your dog and my tin foil had is pointier then your tin foil hat!
So there!
Ha! My PC is a time traveller! It came from 2006, stopped in April 2010 to get Ubuntu 10.04 and skipped another year!
PS: It also knows when I will want it to be on 😀
well, my disc is bigger than your disc
On PC laptop you have crappy “builtin usb” wifi dongle. Check lsusb for info about that.
Mac OS X 10.6 is not an instant reconnect to wifi after sleep. I experience the delay daily. Its anywhere between 5- 10 seconds. Sometimes I have to completely disable wifi and bring it back up.
hmm, wonder if will it be enough to “hiberboot” when going trough the tedious endless chain of reboots during windows update process.. :p
Doesn’t look like new technology to me… more like: logoff && pm-hibernate
Still think this would work better on linux and mac than on windows, since windows is an os that needs to be ¤restarted¤ pretty often in comparison to its competition. (not even including the horrible windows update in the math..)
I believe its a good idea though
Are you actually a current Windows user? I’m still on Vista and Windows Update is activated: reboots are not as frequent as you guys make it out to be. The Windows XP days are long gone.
vodoomoth, well I don’t consider my self a windows user, but technically I am.
I was in the past, and am still, but only when at work.
Now I’m responsible for 400+ windows computers and a half a dozen windows servers. So I install, configure, and maintain windows computers/servers daily. Mostly win7 and some xp. Even got a 95 running for a bizarre program :p
I love the speed, freedom, stability and versatility I got with linux. My home is MS free zone,
and I use windows only when I am getting paid.
Luckily I have to manage a few linux desktops and a couple of linux and bsd servers too both on and off site. Thats a nice break from all the windows-work.
Some days ago I was scripting and had to use icacls command, turned out I had a old and bugged version that wouldnt be updated through windows update. Command was partly defect and I had to go through hell sending emails to MS and crap to make them send me a download link to a patch.. wth? what is MS afraid of?! Patches should be publicly available.
chown+chmod ftw, sooo much better!
open source <3 accessible, powerful and free:)
I hate windows update, but that’s not the only part of windows that suck. The list would probably be shorter if listing what does not suck.
The only reason I can see for a techie to use windows must be to run directX games.. For the rest, linux eats MS for lunch.
Thanks, I hadn’t had a decent laugh today until now.
ilovebeer, glad you enjoyed it.
Here is for another laugh
I laughed at the microsoft commercials,
but feel free to laugh at author instead :p
http://linuxologist.com/1general/microsofts-best-buy-lies-about-lin…
Another new version of Windows, another new promise of faster boot up times. I thought you had used Windows long enough to notice the scam by now, Thom.
Notice how they are improving the part of boot-up that is not really truly the problem with Windows. It is all the disk trashing that goes on when you log on along with their brain-dead FIFO I/O system that causes people to count boot-up times in minutes on budget computers.
Basically what is happening is that all disk access is serialized in the kernel without any regard to what process is doing the request. This means that if you have just a few background processes using the disk, the kernel buffers are filled with that and the 1 KB file the user is waiting for now takes 5 seconds to fetch.
Nothing they have announced here addresses the real problem (poorly designed I/O priority and that countless apps all start at the same time) and so the next version of Windows will effectively boot just as slow as the last one did on budget PCs.
I’ve used most iterations of Windows back to Windows 95, never spent my money on the absolute top-of-the-line hardware, and never had a boot time longer then 30 seconds to usable desktop.
My advice to you is to go through the services and set whatever doesn’t need to be automatic to manual. You should notice a big difference with that alone.
Great post dpJudas!
my favorite part in Windows 9x was closing start menu – if you click on start menu and start to browse for favorite program BEFORE windows finished complete boot process then there is a big chance that start menu will CLOSE in moment when windows complete with booting process interrupting you in browsing through start menu.
Microsoft NEVER know how to build quality product – their products are Spaghetti code with crap UI.
like this: http://www.youtube.com/watch?v=vdCvVVFJdns
True. This still happens with Vista SP 2 and you don’t even need to be in the boot process: click an icon to launch a program and immediately after, open any menu or popup window (like the Start menu) and browse it. As soon as the launched program’s main window shows up on the screen, the popup window is closed. Just tried it with Nero and the Start menu. I guess Windows people still haven’t made their minds about where to insert new main windows when popup windows are opened.
I can see you’ve never used Visual Studio or even the basic Windows Mail. I hate Vista as much as I can hate anything but despite its insufferable slowness, it works flawlessly.
Microsoft makes bad systems software. At engineering, they fail. Their applications group has gotten pretty good, on the other hand, and is less easily maligned.
Cheap shot and unsubstantiated claim. Such a cheap shot and yet you don’t give an example. I have one counter-example: Microsoft’s SndVol.exe vs Linux sound system. And yes, a second one: video driver crash on Vista and video driver crash in X. I know which boat I’d rather be in.
Are you seriously suggesting that I need to provide *evidence* that MS isn’t good at writing systems software? Is your counter-argument *REALLY* going to be that they made one utility better than a similar utility that can be found elsewhere?
User-facing utilities like that are solidly on the application side of things. Perhaps you’re talking about alsa vs. the Windows sound subsystem? That would be more solidly on the systems side of things. But, even so, your argument is fallacious: “Because A does B, which is a member of class C, better than E does B, this proves that A is good at doing things of class C.” This doesn’t hold. It proves at least that A is better at B than E is, and at most that A is better at C than E is.
If you are trying to say “Audio sucks under Linux more than under Windows” you are saying something which is supportable but irrelevant. Very few people, or companies, fail at everything all the time. It was never my intention to say that Microsoft has *never* done *anything* right when it comes to systems software, just that they’re generally not good at it. One counter example doesn’t disprove this.
I know what you’re thinking! “COUNTER example? Where’s YOUR example to begin with?!” I hold up to you the last 30 years as evidence.
Honestly, I’m surprised. I was trying to be *complimentary* to Microsoft, which is something I am loath to do. I am forgiving MS its well known failings and being supportive by saying that there are, indeed, some things they do well. Take the compliment and try not to make yourself appear ignorant by trying to argue that Microsoft hasn’t repeatedly failed to produce quality software.
No you weren’t being supportive of Microsoft … it was a cheapshot.
Windows Vista Service Pack 2 and Windows 7 are so stable … fast (yes fast unless you have really crap hardware and most modern Linux distros will also run like crap on them).
Considering Windows has been ported to ARM pretty easily considering it is probably in excess of 100 million lines of code actually suggests they are fucking excellent system engineers.
Slag Microsoft off all you want … you won’t change the fact that many people use Windows everyday with zero problems … I have never needed to drop to the command line to fix a problem with Windows.
The fact is that Microsoft haters like yourself pretty much have nothing else to slag Microsoft off about other than “freedomz”, ancient business practices (which many of the competitors have done the same which are now proponents of Linux), it isn’t *nix based and that The OS doesn’t run that well on computers that are considered ancient now.
No it doesn’t. More code != better engineers. They moght be, but the number of lines of code says nothing about how good the engineering is. If anything, it’s the opposite.
“freedoms”. That’s clever. or not.
How this for a reason: I think it sucks and it hampers my work productivity.
That’s just a valid reason for me as “many people use it without problems”.
I didn’t say that more code != better engineers … what I said that a large code base got ported easily.
The big thing about NT is the HAL layer pretty much proves that the whole code is well modularized … which is what I alluded to.
How exactly? I have never seen a convincing argument that working with an Open Source bit of code is actually better than closed source … Crap code is crap code whether it is open or closed.
I don’t like shifting through fuck loads of someone else’s code to fix it. I rather it just bloody work in the first place … whether it is open or not is irrelevant.
Not really.
Oh BTW just so you know I buy OpenBSD releases. Got my copy of 4.9 right next to me.
I really don’t see how.
How easily you forget the 90s (and the 80s… oh my). Perhaps you just weren’t there?
Portability is a positive sign, but doesn’t prove much. Does that make it better designed? In terms of code structure, perhaps, but little else.
Windows NT has always been portable. Some parts of the kernel reflect good design, IMO, but there’s a lot of craziness running around there, too.
Wow, nice straw-man argument! I didn’t say anything of the kind, so I don’t know why you’d care to bring it up.
Aha, so you’ve never really had to use Windows? That explains a lot. FYI, there are a lot of things which can only be done via the command line on Windows… or (sometimes) are just much easier.
I am indeed a Microsoft hater, though how you arrive at that conclusion by reading this thread I cannot imagine. I hate Microsoft because they make things that are unpleasant, broken, break under me, are inscrutable, fragile and sometimes impossible to fix. It’s not fanaticism, it’s tired experience. (Full disclosure: I am also a GNU fanatic, but I am rational enough not to let my preferences color my facts to any excessive degree.)
Which “Ancient” practices? As far as I am aware, they’ve never stopped doing any of the bad things they do… except for when it no longer matters. Are you just saying “Don’t complain, you didn’t get screwed THIS month!”?
I do have a problem with Windows not being Unix-like: The problem is that it could be, but sometimes is arbitrarily not, and in some cases that bites me. I have a problem with getting bitten for (as far as I can tell) no good reason.
As for ancient computers… who is it you *think* you’re arguing with, here? Did I claim “Windows sucks because Windows 7 doesn’t run on my PII350!”? I would never be able to run Fedora 15 on it, either, but I don’t complain about that very much.
Basically your post boils down to
“I hate Microsoft … rah rah lalalalal … I had some random problems with Microsoft products that nobody else has … rah rah … something about anti-trust stuff that happened 15 years ago … rah rah … UNIX way is perfect (but forget about Plan9 … All Hail GNU and RMS”
Please … I think 90% of your problems are made up.
Edited 2011-09-12 21:50 UTC
Are you serious? Are you saying “The current software put out by Microsoft is finally at a quality level that isn’t laughably bad, so therefore you can’t say that Microsoft makes bad software.” Is that really your argument?
Are you trolling? The user ‘kovacm’ said that Microsoft never knew how to build quality. You replied and held up Visual Studio and Windows Mail as counter-examples. I replied and allowed that their *applications* group has gotten pretty good, but that their engineering wasn’t good. This isn’t me “bringing up WIndows being crap” – I’m just clarifying what kovacm was saying: VS and Windows Mail do not disprove it.
Portability *might* indicate well structured code, but well structured isn’t the be-all and end-all of quality.
I said, specifically, that only some of their software was crap. “Most people use it” is not a very strong argument: Most people eat at McDonalds, it’s “good enough,” but the food there is really not very good. I don’t see my point being disproven at all; Microsoft isn’t good at systems software, and no matter how popular it is that doesn’t change.
Don’t be offensive. Because I know something you don’t, I don’t know how to use Windows?
What has this got to do with *ANYTHING*? Seriously, you’ve baffled me here. “Notepad.exe isn’t a command line tool!!!” – well, no kidding.
In fact, there are tools for traceroute and ping. I am not talking about such trivial nonsense. Let me pull an example that’s fresh to mind (since I just did this): Set up an IIS7 virtual host with a single SSL cert so that you have two SSL sites running on port 443. Do it without invoking a command line tool.
You can say many bad things about open source tools and APIs, but in the end they are never impossible. You can always know. Furthermore, I am not here promoting Free Software. I am here describing Microsoft. I don’t care how many times you say “But look over here! It’s bad too!” it doesn’t detract one iota from my legitimate criticism.
Which .net is fantastic? The class libraries leave a lot to be desired, but the platform is okay. Clearly you think that this makes up for decades of worse things, but I don’t forgive so easily.
It shows my point of view, but it doesn’t automatically make what I say untrue. I do not let my bias alter any facts.
Bullshit. You’re either a troll or a fool: GNU is about MY freedom, not your profit. If you believe that a *side effect* of a GNU software world would be little money made by software then perhaps you have a point we can discuss, but to propose that this is the *purpose* is disingenuous and offensive. I have tried to stay away from hyperbolic, ridiculous claims about Microsoft and I think you could show a similar courtesy.
Irrelevant, as I have said, since I am not defending GNU.
Microsoft is well known for caring about the platform over all else, so this is no surprise. It’s a lesson others would do well to learn.
I think we’re using a very different definition of “broken.” Hint: It’s not about being to load up old executables.
Lovely to be young, I suppose?
Irrelevant, as I have said, since I am not defending arbitrary “open source” developers. “Microsoft writes good software, provably, because other people write bad software.” Even if I agreed with the premise that constantly breaking the platform is evidence of bad engineering (which I will not admit on the face of it) the argument is a logical fallacy and, thus, unsupportable.
Do you even read what I write? I’m beginning to think no… I’m complaining about places where Microsoft chose incompatibility when all else was equal (a thing they do rather often, really). There are a huge number of problems with Unix as OS design, which is why I am more of a Linux fan than a FreeBSD fan (for one) and quite fond of Haiku. Can you stop attacking straw men and focus on Microsoft?
Aha, the crux of the matter. “Works for me, you must be a liar.” I’m glad your spoon-fed software world as provided by Microsoft(tm) (so long as you do things their way) is so wonderful and glorious and idyllic that you cannot imagine why anyone would ever have a problem with it! Despite the fact that you said that some Microsoft software is crap a few paragraphs earlier…. can you pass me some of that kool-aid? I’d like to get it analyzed so I can become rich selling the essential ingredients.
I despised Windows long before I ever heard of Linux or GNU. There’s a lot to dislike. If you don’t, GREAT!!! I don’t wish to associate with you, or speak to you. Go and live your the life that your upstream vendor permits you to live. Work inside the limits that were arbitrarily chosen by people concerned with what’s fast and profitable and not what’s correct. Don’t let my assessment of the competence of the people who define your horizons affect you to even the smallest degree! It can’t possibly matter to you, just as the desperate way you cling to your little world doesn’t matter to me.
Have you yet realized that I am not trying to convince you, or debate you? I speak so that those who can hear will understand. If you don’t see what I mean the message isn’t for you. It is, in all probability, only for me. Sooner or later all conversations become metaphysical in nature and any further reply of mine would necessarily speak only of the vagaries of perception, the difficulty in proving that anything exists, and the questionable value of claiming to know things. That is the only useful thing left for you and I to talk about. We clearly stopped talking about the topic long ago.
Let put it this way … You make a claim that Microsoft Windows is a load of rubbish … don’t make any examples where it actually been a problem.
Spout a load of “the code must be free” … because you think you have the right to own someone else’s work, bceause it is software.
And you wonder why I speak to you in contempt.
I didn’t do that anywhere. I certainly prefer Free Software; I like the principles and support the movement. I don’t declare “The code must be free” – except to say that I prefer it for my systems. I only said that, with open source, no matter how bad the code is you can at least always find out what it’s doing. That’s useful and makes even bad open source software less frustrating than closed.
I don’t know where you get this notion. You are telling me what my motivation was for saying something I didn’t say. What do you think I’m talking about? It’s not Free Software, except as an aside because other people brought it up.
Truly baffled…
If you agree with GNU you believe that all code should be free and in “freedom” … if you don’t think this you are not in line with RMS ideology.
There is no halfway with RMS.
Oh right … we are talking Windows Server now not desktop … These aren’t common use cases.
Slagging off one thing while saying GNU is fantastic … of course I am going to counter by saying what is shit about GNU … and sometimes it is impossible … just look at the Braid Developers blog … X11 basically couldn’t do something so he gave up on a Linux port.
Anything => 2.0 … 1.1 was terrible in many ways and I still have to support it.
The class libraries are quite extensive and consistent … C# is basically “Java Improved” edition.
Neither do I, the difference is that I am a pragmatist … I want the job done … I don’t care about ideology until it gets in my way.
RMS created GPL and GNU because software engineering and the commercialism of software destroyed the hacker community at MIT …
RMS believes all code should be GPL or something that is GPL compatible … which leaves no room for proprietary software developers like me … he is the fundamentalist … and he hates developers like me.
Not to a pragmatist it isn’t.
Have you heard the expression “I’ll scratch your back if you scratch mine”.
My definition of broken is that it doesn’t work … what one are you using?
Your comments so far have come down to “microsoft is shit because I say it is”
I don’t care whether you think it is a logical fallacy or not. Windows works well on pretty much every PC out there … and has been ported to ARM with relative ease which is nothing else suggests they know what they are doing.
Your general attitude leads me to make these assumptions.
Microsoft make things easy for me that let me get my job done quicker. The marketing team are happy and my boss isn’t getting earache.
Kool-aid is like so early 2000 Linux fanboy speak … Look at the end of the day I was a Linux Server Admin for quite some time. I have used IRIX, OpenBSD, FreeBSD, Linux and Windows in production environments … Windows tbh has given me the least pain … OpenBSD second (I can just leave the box and it just does it thing).
Are you some sort of Linux hipster …
Oh wait because I have a differing opinion in you don’t wish to speak to me … fucking sad man.
Oh I am so above you because I have found the secret to life in GNU … FFS … you sound mental.
Getting the Job done is most important to a decent level of quality … nothing else matters to my superiors so I choose to get paid … bitches.
Lets get this straight you spoke a load of rubbish and I brought you up on it while using a few F-words and you are so upset you can’t deal with it.
GROW THE FUCK UP.
You said that *nothing* in Windows required using the command line. You didn’t say “No common cases in Windows when used as a Desktop” — nor was this implied in any way from context. Forgive me for not reading your mind.
You are free to criticise GNU software as much as you like, just don’t pretend that it de-legitimizes criticisms of Microsoft. That’s silly. (Bringing up Microsoft-vs-Linux-vs-GNU at all was pretty pointless.) As for the rest of your ellipses-laden comment… what are you babbling about?
I have some problems with what Microsoft seems to think is good API design (and I keep finding holes in coverage which are weird, to me). But, I can’t say much that’s bad. Java’s only virtue over .net is being older–well, it’s also GPL’d now, which I like, but that doesn’t help the class library.
Partially right. He created it because he was frustrated by not being permitted to fix something. Preserving the hacker culture was never the primary goal, from what I read.
I’m with you up to here.
I don’t believe this and am prepared to prove otherwise using logic (but not in this thread, as it’s off topic).
This is true.
This is speculation at best, libel at worst. To say he hates you is over-the-top drama; you perceive that he is on a course which will eventually destroy the basis on which you make your living. You feel upset by this and attribute malice where none is known to be intended. (If you spoke with RMS and he told you he hates you, personally, then I stand corrected.)
I don’t believe RMS’ course will eventually destroy your livelihood, nor do I believe that destroying the business of people like you is his intent. If it is a side effect I would be surprised if he were not regretful more than gleeful (but I won’t pretend to know anything about how he feels).
I again reiterate a confusion as to what RMS has to do with Microsoft’s software engineering practices.
Something is broken if it is stupid, braindead, backwards or idiotic. If you do a round trip to the server for every record, that’s broken no matter how many times it works. If you do something which performs poorly, makes poor assumptions, or isn’t robust, that’s broken. That a thing “works” is the least important factor in broken-ness. Something that’s not nicely done can be broken and still work every time.
I am not here to provide evidence. I don’t feel like it. You are free not to believe me, of course. I am not interested in convincing you.
You don’t care that your arguments are worthless? Windows *runs* on all kinds of hardware… that has been produced to run Windows. That doesn’t make it good. It was ported to ARM, but that’s not enough to *prove* its worth.
This is a problem (your problem). You see me say X, which you leads you to believe I am motivated by Y, which means I must fall in to group Z, which you have previously see vehemently argue in favor of position P, so you attack me for holding position P. You’ve done this sort of thing several times. It’s distracting and irrelevant.
No. I love speaking with people who have differing opinions. I dislike speaking with people who *like Microsoft and Windows*. This is a very specific personal hangup. If we were to switch to politics, religion, history, science, culture, economics or almost anything else you would find a very different respondent.
If you used any F-words I didn’t notice. I am in no way upset, nor have I been at any time. This brings me back to what I was saying: Further (useful…) discussion with you can only occur on the topic of the nature of reality. Is what you think you hear real, or what I intend to say? We are now entirely off-topic.
Okay, I noticed that time: Check, one F word.
I object to the assumption that I am young, therefore something is wrong with me and, if I would only reach your age, and this your opinions, I would no longer trouble you by not agreeing with you. If only one of us is young, and therefore inexperienced, and therefore not enlightened, and therefore not worth listening to… it’s you. I would not suggest such a thing of my own volition because it is as discourteous as it is irrelevant. I recommend in the future that you confine your rebuttals to things that are topical, logical and defensible.
Okay so you have a problem with me because I like a particular company … but you don’t have a particular problem with me if I have totally different view points in general ideology.
Seriously people are killed every day over religious ideology (and thankfully) nobody I don’t think has been killed over a difference in software licensing opinions.
Seriously you don’t like me because I like a certain companies products … WTF … My manager like Macs … and I don’t … guess what me and him do some “ribbing” over it and get on with our lives and jobs.
WTF?
Look at the end of the day … You have not given one concrete example … after asking you numerous times how Microsoft has made your life difficult.
And BTW I don’t consider anything pre 2000 relevant.
Your cmd line example which disproves me is not on a desktop version of Windows … which is what the original topic is about (talk about changing the goalposts) .. it is a server use case.
You then claim that somehow I am trapped in a warped world … and I am somehow trapped …
Guess what that is exactly What RMS wants people to believe that because I am using something “which works” but I can’t see the code e.g. GPL … I am somehow brainwashed.
I am a pragmatist … I use what works.
Microsoft shit works … but the GNU stuff doesn’t that is the criticism … what is so hard to understand.
No API/Software is perfect and bug free … but at least .NET is consistent since .NET 2.0. I don’t have any surprises …
There are is other evidence to support my theory.
http://penguinday.wordpress.com/2010/08/10/only-become-less/
Quotes with actual citations.
If you are a small company and GPL your software … your largest competitor (which has the resources) will just fork you code and offer support … how does that help small businesses … oh it doesn’t … GPL actually helps people like Oracle, IBM and Google … but it is Microsoft we have to hate RIGHT!
Oh it off topic now … the whole discussion isn’t about Windows 8 boot times .. you brought it off topic … now you aren’t willing to discuss off topic .. even though you brought it offtopic …
BTW my argument is in the last quote section.
I write decent code … this guy devalues the amount of effort I took in perfecting my craft, and says I should give it away or if I don’t I am a social menace … he can go f–k himself.
Well whatever it is … it won’t succeed
It to do with your love of the GNU.
Grown the f–k up … Windows works well over many devices so it doesn’t even meet your definition of broken … let alone mine which is more general.
Oh so you are here to make statements and not back them up with any evidence?
Look it isn’t my fault you can’t extrapolate.
You follow the same behaviour … how I am supposed to see different? If you don’t give any f–king examples of why Microsoft have wronged you … which you have still failed to supply a single example.
All you have done is parrot others.
I am pretty sure you don’t have a concrete example … which I can disprove with a simple google.
Edited 2011-09-13 23:20 UTC
Although you have edited this out now … is that what you really believe? Amazing! Why would anyone do that? Spite? It makes absolutely no sense.
In actual fact, GNU is organised as a consumer’s co-operative business.
http://en.wikipedia.org/wiki/Cooperative#Consumers.27_cooperative
The actual objective of a co-operative is to create quality goods for minimum cost.
There are essentailly two ways for enterprises to increase profits … one is to increase prices and the other is to reduce costs.
For the vast majority of the world’s businesses, who are consumers not producers of software, GNU software (being a consumer’s co-operative) is an EXCELLENT business opportunity to increase profits. Software is purely a cost to most businesses.
GNU is a great way for businesses to make more money from their use of software.
Edited 2011-09-12 23:35 UTC
It does actually RMS hates paid programmers, because the commercialisation of IT took away his hacker haven at IT in the 70s.
As for your other points … Most commercial products are a one off licensing cost … however getting support of Say Redhat Linux is an on going cost.
Saying that GNU is always cheaper is okay as long as a business is happy not to have any support … most require a SLA contract.`
In fact implementing Moodle has been significant expense compared to our in house solution.
Edited 2011-09-13 08:57 UTC
RMS is a programmer. The majority of open source programmers get paid for their contibutions. Where on earth do you get this rubbish from, anyway?
Agreed. Consumers of software should look at all the options available to them and make a rational choice. Looked at objectively, open source software is nearly always the most cost effective solution, despite TCO studies sponsored by Microsoft which ignore costs of downtime or malware protection, for example.
This is exactly why Toyota joined the Linux Foundation and are moving to open source solutions.
http://www.techspot.com/news/44579-toyota-joins-the-linux-foundatio…
Toyota made a rational choice. BTW, Toyota will still be employing programmers to write software for its vehicles.
Now, as for governments, an entirely different situation arises. Governments want sovreignity, they do not want to be locked in to a sole commercial supplier for vital infrastructure components, especially a supplier from a different country.
http://homembit.com/2011/09/microsofts-attack-on-brazilian-national…
Oh, hello lemur… I’m surprised to find anyone else slogging through this thread. I’m so sorry.
Uh, those two aren’t even remotely comparable. SndVol.exe is comparable to kmix, puvucontrol etc. The equivalent to the Linux sound system would be Windows Audio.
You seem to forget that the ‘disk thrashing’ will be a thing of the past when a new mandatory requirement of Windows 8 emerges.
-> In order to activate (if you have a Windows 8 FastBoot license) Fastboot, Windows 8 must be installed on a 12Gbps SSD of at least 1TB in size.
For everything else it is the 2-3 minutes disk thrash.
IMHO, Microsoft could do a lot to cut down the crud that gets started in an OOTB Windows 7 system. Just look at the services or the list in msconfig. There are far too many things running. If you (I mean you MS) want to get startup times down cut the crap out.
What I’d like to see is some proper benchmarks of startup. Not just with a clean freshly installed OS with no ‘extras’ installed. That is just not a realistic option. It is like stating that ‘This car can do 160mph and in thre small(very small print) {coming down pikes peak with a following wind}.
I’ve noticed also that Windows depends heavily on disk access speed. The proof is that as soon as you defragment the Windows system files the overall system responsiveness increases massively. This means, that while using the Windows – like opening explorer, coying files, etc – windows every time accesses its DLL-s and it likes to read them from the disk. INSANE. If you have your shell, kernel and user DLL files in 20 pieces, those pieces are grabbed together gazillion times a day and that mans slow response time. I don’t understand why MS does not use its cached copies of DLLs and accesses them from the disk.
Edited 2011-09-10 08:38 UTC
Windows’ I/O scheduler is just really awful. I mostly use Linux but use Windows at work and was amazed by the ever-present disk grinding and slow processing on any day when Symantec was scheduled to run a full scan. It’s priority is set to as low as possible, but still everything takes at least 4 times as long (no exaggeration here, I’ve timed it).
After reading LWN one day and seeing a quote from a Linux kernel dev saying, basically, “If telling all your I/O heavy software to go all at once is ever slower than telling it to run in sequence, that’s a kernel bug and we want to fix it,” I realized that the reason this slowness bothered me on Windows is that I don’t see it nearly as badly under Linux… and that the Windows I/O scheduling system must just be bad.
After that I removed all startup items from the registry and start menu and replaced them with a single batch file. In this file I inserted explicit sleeps of approximately the amount of time (as recorded by me) that I/O-heavy tasks would take to start up. Doing this, I found, improved overall startup time noticeably.
Verdict: Windows (at least up to the latest Vista) is so poorly written that ‘manually’ hinting to the I/O system dramatically improves performance. Who’s surprised?
Cut it with this crap … You read on a site that is pro-Linux that Windows was doing something wrong … That isn’t evidence that is propaganda.
Your problem was that anything by Symantec is a pile of crap.
Funnily enough I can run a Disk Scan (complete) with MSE and I was playing Deus Ex: Human Revolution at the same time and never even experienced a slow down in framerate.
Now now, be nice.
I didn’t spout off ridiculous “ME TOO”isms. I am relating my experiences and data I collected by observation over the course of many months. Symantec AV is pretty crap as well, of course, and this made things worse on days when it was running, but the overall effect was not altered. Just to be very clear: I observed the same pattern of slowness and speedup even without symantec grinding the disks horribly.
I’d like to address this point specifically, because you greatly misunderstand me. I read on LWN a Linux kernel developer talking about **LINUX**. In his opinion Linux has a bug if it cannot cope better with “A bunch of IO heavy tasks started simultaneously” than it can with “The same tasks started in sequence.” He was saying that Linux should do better than the user, always, in scheduling IO or it’s a bug in Linux. He never mentioned Windows.
What **I** did was begin thinking about the differences in the way I observed Linux and Windows reacting to disk contention, which lead me to run some experiments, which lead me to the conclusion that I began with above: Windows sucks at scheduling IO.
Edited 2011-09-12 20:09 UTC
I always find it interesting that Only Microsoft haters have massive problems with Windows …
Edited 2011-09-12 21:45 UTC
It’s cause and effect, really. Those who have problems hate Microsoft.
I think you look for problems tbh. I still haven’t had any real reason why you’ve had problems working With Microsoft products.
Edited 2011-09-12 22:38 UTC
You hear no specifics because I am weary. I see no future in bringing up complaints here and now. I just don’t care to enumerate all of the things that have bitten me. You can assume that I am just looking for reasons to complain, or that I have no basis in fact for my complaints, or that I just don’t have any idea what I’m talking about, or whatever you like. It’s not true, but if you don’t believe what I say I cannot do a thing about it by saying more.
Right, Let me get this correct, you basically made a accusation with nothing to back it up … after I asked for example you are unwilling to hight-light one problem
…
Somewhat related to this is my personal observations of Linux and Windows I/O behavior on a laptop with only 1 GB of RAM back in 2006.
Surprisingly, Windows XP seemed to perform much better under serious load than Linux 2.6.24. The main problem that Linux had is managing read activity under severe memory pressure. Reading data in always required writing data out to swap and/or releasing data from disk cache.
Linux almost always got this wrong, writing out data to swap or releasing buffer cache that was soon required, while Windows XP seemed to be able to keep a relevant working set in RAM for each application.
Linux also created absolutely horrid I/O patterns, writing a few KB to swap, then seeking back to read a few KB from a file, then seeking back to write a few more KB to swap, then needing to read a bit from other swap, etc. On a laptop drive this decreased I/O throughput to less than 50 KB/second while Windows XP seemed to be able to swap out a few megabytes at a time and then read entire files into cache. This may also have a lot to do with how Windows prioritizes the foreground application.
I sure hope the Linux developers have managed to fix the awful VM behavior. I know there have been a lot of fixes made to it.
I probably won’t see the differences since I no longer have any 1 GB machines.
Your experiences are not out of line with mine. I did stop using swap on Linux years ago on my desktop systems, or rather set swappiness so that it was only ever used in dire emergencies. This does make make things remarkably better. And, as you note, Windows uses swap often without as bad an effect.
It is my understanding that there’s a lot of tuning you can do if you want to try to make this sort of thing better, but I haven’t looked in to it.
I’m one of those people that doesn’t use sleep or hibernate, just because the normal shutdown startup of ubuntu 10.04 LTS is good enough for me.
Sure, I’ve seen computers running windows 7 that are ready to use quickly, but most of them are just horrible on startup. Some even need minutes after you can see the desktop before they become really usable.
Before I believe all those things I first have to see how windows 8 performs on the average computer with an average user after a couple of months..
Why does everybody seem to think that startup times are the only reason to use hibernate? I only care about resuming sessions where I left them.
I don’t need my session to resume, so I forgot to mention it, but I certainly agree that it can be a valid reason.
Just leave your computer running, don’t shutdown OR hibernate. You get instant-resume and your ‘session’ is just as you left it.
I have to laugh and non-laptop sleep and hibernate aficionados. Overcomplicating things, much?
Brilliant! How come nobody has ever figured that out?
Are you the kind of person who leaves their car running so that they feel comfortable in it when the time to commute has come? I’m not. And I guess nobody sane would suggest that.
Nobody sane would suggest that, because it’s crazy. Who said anything about comfort? You’re confusing things which are not alike. We’re talking about convenience and efficiency, not comfort, and we’re talking about computers, not cars. These things are different therefore what’s okay for one is not necessarily okay for another.
Are you the kind of person who puts oil in his computer keep it running well? I’m not, and I’m guessing nobody sane would suggest that.
We all know Windows is great at some things and crap at others. We all know Linux is great at some things and crap at others. Why has this thread turned into an argument between Windows/Linux fanboys over who’s wiener is smaller?
It’s 2011. I can’t believe people still whine about this stuff. It’s like 1990 refuses to die.