Dell Documents Reveal Microsoft's Pre-launch Vista Errors 220
twitter writes "The New York Times has a piercing analysis of documents from the Vista capable lawsuit. The documents show that Microsoft seems to have put a wrench in Vista's driver situation only at the last minute. 'Late OS code changes broke drivers and applications, forcing key commodities to miss launch or limp out with issues,' said one slide in a Dell presentation dated March 25, 2007, about two months after Vista's launch at retail and availability on new PCs.' We have all heard the lazy vendors don't believe Vista will launch excuses but few of us have heard Steven Sinofsky, chief of Windows development, second and third opinions. 'Massive changes in the underpinnings for video and audio really led to a poor experience at RTM,' he said. 'This change led to incompatibilities. For example, you don't get Aero with an XP driver, but your card might not (ever) have a Vista driver.' Finally, said Sinofsky, other changes in Vista blocked Windows XP drivers altogether. 'This is across the board for printers, scanners, WAN, accessories and so on. Many of the associated applets don't run within the constraints of the security model or the new video/audio driver models.'
Hasn't MS always done things like this? (Score:2, Informative)
Re: (Score:2)
All 640k of it!
Re: (Score:3, Informative)
This means even the vendors putting Vista ?Ready sticker son computers knew those computes wouldn't run Vista all that well.
Re: (Score:2)
Re: (Score:3, Informative)
When OEMs were releasing their Windows XP computers with 128MB of RAM (256MB with SP2), I always said it was criminal.
I had a computer with 128MB ram that came with windows 98. I installed XP on it, and was much happier with XP than I had been with the previous OS. Sure, it was a piece of crap (speed/memory was not the worst of its problems), but I liked XP on that machine. I agree that a customer that bought a computer with 128MB ram and XP was probably getting a raw deal, but in that situation, I would have preferred XP to any prior windows OS.
Re: (Score:2)
(Generally, with less than 512MB RAM, I find Win2K's lower memory footprint makes up for it being not nearly as optimized as XP.)
how to advocate free software (Score:2, Insightful)
But why? (Score:5, Interesting)
Re: (Score:3, Interesting)
Re: (Score:3, Informative)
Re:But why? (Score:5, Insightful)
My father does a lot of video work, and any time he tries to access or move a video file he has crazy wait times while Vista chews on something.
What the hell changed between XP, which he has since gone back to using, and Vista that so radically changed the handling of video files? From his reading on various websites (none of which I can vouch for) the OS is checking for some sort of signatures in the files to figure out if he has permission to perform the selected task.
I have no idea if this is true or not, but either way, he had to ditch Vista and return to XP in order to do things like edit the video he shoots of conferences and events.
So the DRM issue has at least some anecdotal evidence in its favor. Either that or Vista is completely incapable of handling files over a certain size with any sort of grace.
Re:But why? (Score:5, Informative)
Re: (Score:3, Interesting)
Anyways, signatures don't give you permission to deal with files, they just state their origins. No different than in real life - stuff with my signature on it passed by me. Embedded metadata, of so
Re:But why? (Score:5, Informative)
While it goes into details about a lot of other stuff, there's the explanation of Vista's (apparent) slow disk performance.
Re: (Score:2)
1. Take a large media file and copy it from one drive to another.
2. Rename said media file in the original location so that it does not appear like a video file.
3. Copy the renamed video file to the same drive.
4. Measure how long it takes to process both scenarios and report which took longer.
Re: (Score:2)
I don't have Vista any more, but I tried it for a few months on a video editing machine.
I don't think it would be possible to run the sort of test you've suggested because Vista performance is so variable. At any time, it'll stutter, slow down or appear to hang for no reason that's apparent from the usage pattern. You'd need to shut down a heap of stuff to get consistency, and if you do that you won't be getting typical results.
Re: (Score:2)
As other replies have mentioned, Vista had problems with copying files - large or small quantities. Microsoft seemed to blame that on how the little scrolling bar dialog was updated, but whatever.
I've been using a beta of SP1 (heard about it from hear, nonetheless!) and that issue has been fixed. Purely anecdotal, I haven't done benchmarks or anything, but moving a GB of files around no longer takes a lunch break.
As others have mentioned, turn off thumbnail generation. Don't know how to do that off-h
Re: (Score:2)
As for lack of support, where else are the media companies going to go if MS says no to DRM? People aren't goin
Re:But why? (Score:5, Interesting)
Not sure, but I found the following, from Microsoft themselves, astounding:
From the Visual Studio 2005 Service Pack 1 Release Notes [microsoft.com]:
Installation Issues - Windows Vista
Setup dialog box fails to appear:
The verification that occurs under User Account Control (UAC) with all installations delays the appearance of the initial setup dialog box. Delays of more than one hour have been reported.
Re: (Score:3, Interesting)
Re: (Score:3, Insightful)
There is no good reason to not give the user an indication of what is going on. If the system design requires that, then the system design is faulty. Faulty system design is not a 'good reason'.
Digital signature checking (Score:3, Insightful)
The reason it takes so long is that it maps the entire file linearly into memory to hash it. Such a large mapping gets demand-loaded, which for a linear scan is very slow.
The workaround is actually quite simple.
Re:But why? (Score:5, Funny)
Re: (Score:3, Interesting)
That's my whole point. If MS blocked all DRM from existing in Vista, they would only be harming themselves. Media companies would take the "fuck you, we still have standalone players" (and lots of them) approach, and would-be customers would whine continually that they can't play back tons of media. A
Re:But why? (Score:5, Informative)
First off, it's well known that the redone video and audio drivers were required for the new DRM. That right there is a change: pre-Vista, the OS wasn't designed explicitly for DRM. Now it is.
Secondly, the new designs shave a good 10%-50% off performance. Audio acceleration is gone. EAX effects are no longer possible. Recording the audio output of programs is no longer possible. All in the name of DRM.
ALL layers are now encrypted. This, not surprisingly, slows down the OS. By a lot. It also greatly reduces battery life. Where before, playing a music file might involve a single decryption step to send the data to the audio player, it now must be re-encrypted before being sent to the card, then re-decrypted before being converted to analog. All because an enterprising user might otherwise snoop on the bus to "steal" the audio data.
In short, Vista is 10%-50% slower solely to allow for DRM. The kernel was redesigned with DRM in mind, not user experience. Battery life was halved in extreme cases - again, solely for DRM.
It's not trolling, there are simple facts that have been exposed time and time again. Look it up on Google. Vista is much, much, much worse than XP when it comes to DRM.
Re: (Score:2, Informative)
No layers are encrypted. The only encryption/decryption applies to protected media, which is already encrypted and requires decryption. If the media is not protected
Re:-1 Overrated?? Moderation Abuse (Score:4, Interesting)
Namely, to prevents developers from creating unauthorized audio/video drivers that can create analog outputs to all media.
Re:-1 Overrated?? Moderation Abuse (Score:5, Insightful)
No it's not. It's just more Microsoft marketing-speak. For example:
The new video and audio drivers have nothing to do with DRM.
No, if Microsoft doesn't support it, we can ALL kiss DRM'd high-def media good-bye, and good riddance. Microsoft had been a key supporter and booster of computer DRM despite their customers' distaste for it. Don't try to pretend they are anything but complicit partners with the studios in this.
Re: (Score:3, Informative)
As for the major patch the occurred very soon after the release of Vista, I was likely a DRM patch, to fix a typical M$ failure.
They remade NtCreateProcess for DRM (Score:4, Interesting)
Prior to Vista, NT had a "create process" mechanism differing in design from most other operating systems. NtCreateProcess() creates an empty process with nothing in it other than the new
This worked well until Vista. In Vista, their DRM system had a problem: they didn't want anyone to be able to debug audiodg.exe, but the parent process had to be able to debug it in order to start it. The solution? Redesign the entire process creation system such that the kernel does all the initial process creation procedures so that the parent does not have control over the child if it is a "protected process". Hence, NtCreateUserProcess() was born.
For those that don't believe that this change was for DRM, I offer proof [msdn.com] in the form of a Microsoft kernel developer on video explaining it.
Re: (Score:3, Interesting)
Re: (Score:2, Insightful)
I'm sure that if MS dropped DRM support there would be an uprising the likes of which have not been seen since /. deleted the Scientology post.
Are you serious?
Re: (Score:3, Insightful)
Re:But why? (Score:4, Insightful)
You've got it backwards. If Microsoft never supported the DRM, the RIAA and MPAA wouldn't have put it on the disks, because of the lack of support. It's not the customers that would be complaining, it's the "media partners." And those bastards? They can have some cheese with their whine.
Re: (Score:2)
If it wasn't supported by Windows, they wouldn't do it. As you may be aware, Windows is a very large proportion of the market.
Re: (Score:3, Interesting)
I don't use Vista for anything media production related so I haven't delved into this.. but it caught my eye a few weeks ago.
Not forced, no technical reason (Score:5, Insightful)
If ndiswrapper can run XP drivers in Linux, then surely MS could have run XP drivers with no problems at all.
Re:Not forced, no technical reason (Score:5, Insightful)
You know, everyone goes around saying "open source only copies, never innovates" and yet you have an (admittedly kludgy) solution to the problem of driver availability that have been forced by uncooperative hardware vendors that does work and does allow older hardware to function. Microsoft has all the kernel sources at their disposal and doesn't have to reverse engineer to get something like ndiswrapper running, and yet instead they shut out a lot of older hardware in one fell swoop.
There just doesn't seem to be much logic to what Redmond does. I can understand the vampiric murderous monopoly that wants to destroy any and all competition, but the design choices they make are bizarre. It's not as if Windows is some elegant masterpiece that they don't want to clutter kludges to keep older things running. Christ, the operating system has been like that since Windows 95.
The really sad thing is that it is closed source, so no one will ever be able to create that sort of an environment to get this hardware working.
Re: (Score:3, Insightful)
Re: (Score:2)
I don't understand it either. Why not build a wrapper and sandbox it? If there are security concerns, that ought to solve it.
No idea about the specifics of the situation, but in the general case this will not work for drivers. If your driver can issue a DMA request to the device, it can access anywhere in your physical memory. If you want a safe, secure, sandbox, then you have to send all device register write writes via something that validates them. Since every device has different DMA commands, your sandbox needs special code for every single driver it runs. At this point, it's cheaper and easier just to rewrite the drive
Re: (Score:2, Interesting)
Re: (Score:3, Insightful)
Microsoft doesn't have this particular problem. They do have the XP kernel sources in hand, and for them building a compatibility layer would be much much easier than for some FOSS developer. And why shouldn't Microsoft? They're the ones pushing thi
Re: (Score:3, Interesting)
Chances are there is no wrapper because the tuner is already supported natively by ivtv (for hardware-encoding MPEG-2 cards) or v4l (for framegrabber cards).
If your card isn't supported, blame the manufacturer and get a supported card instead. I recommend the Hauppage PVR-x50/500 series for
Re: (Score:3, Insightful)
That's like sayin' ya bought a WinModem 'cause it was cheap, and it's the fault of Linux that it's not supported...
Out of curiousity, what make/model is your TV card?
Re: (Score:2)
Seriously, though, NDIS is contained, but a tuner card interacts with the multimedia stack. Not anywhere as contained. It could be done, but the "wrapper" would carry the bulk of WINE with it.
And, it would have to wrap something -- the current multimedia architecture in Linux isn't stable enough to target.
Yes, I would have expected Microsoft to have a compatibility layer -- given that such a layer is the ONLY way to use certain devices moving forward. After all
kernel space is not an issue. (Score:2)
The extra engineering might pay off. People with older (or even current) machines with XP drivers might have bought Visa. hardware vendors would have had an easier job and the whole anti-Vista wave might not have happened.
Re: (Score:2)
Security (Score:5, Interesting)
When rebuilding a system from the ground up for security, these issues need to be hashed out first. The fact that the security and driver models were changing significantly shortly before launch is a sign of bad design. Or at the very least horrible project management. If Vista was in the works for over 5 years, and it was designed properly from the start, 3rd parties should have had plenty of time (years) to conform to new models.
Re: (Score:3, Insightful)
Re:Security (Score:5, Insightful)
The fact of the matter, and this is only getting driven home ever more with these revelations, is that Vista was released prematurely, before adequate time to test and correct various issues could be taken. Microsoft and the manufacturers needed to get this beta operating system to market to try to force new computer purchases. The unholy OEM alliance between the big manufacturers and Microsoft is coming home to roost.
Not only that, but it's a gas to watch the chaos that surrounded the final months before Vista's premature birth.
Re: (Score:2)
Re: (Score:3, Insightful)
We're not talking about a broken link on a web site. We're talking about an OS that goes on hundreds of millions of computers. Make a few mistakes in the some of the details, but don't fsck up the model.
Re: (Score:2)
It does indicate "horrible project management" when errors like this jeopardise a multi-billion dollar, 5 year project.
Besides, there's plenty of other evidence [blogspot.com] of horrible project management of Vista.
So that nets us an estimate of 24 people involved in this feature. Also each team was separated by 6 layers of management from the leads, so let's add them in too, giving us 24 + (6 * 3) + 1 (the shared manager) 43 total people with a voice in this feature.
The feature? Vista's shutdown menu...
Re:Security (Score:5, Insightful)
Re: (Score:2)
Re: (Score:2)
Re: (Score:3, Interesting)
Re: (Score:2)
Re: (Score:2)
In any case, I was talking about security, not stability. Shift the goal posts somewhere else.
Re:Security (Score:5, Insightful)
This is why monopolies are so damned bad. There's nobody for them to reasonably compete against, so they have their own market ecosphere which doesn't have to make any damned sense at all save from the marketers and bean counters point of views. Everyone keeps telling me how the best and brightest end up in Redmond, and yet over and over and over again we keep seeing the same bad architectual decisions that keep biting everyone in the ass in the same way. We see the absolute lack of forward thinking, of at least trying to create a development model that can accomodate security, utility and extensibility.
I don't think anyone expects five year old hardware to run the latest MS operating system (although, ironically, you can often do that with Linux and FreeBSD), but there have been hardware problems with stuff that was a couple of years old when Vista was released. Yes, the manufacturers deserve a lot of the blame, but Microsoft has handed the perfect excuse for engineered obsolescence. The manufacturers are only reflecting Microsoft's own disdain for doing right by the customer.
Yes, they've got the average consumer by the balls, but the average consumer isn't the big profit center. It's all those corporate installs, with their hundreds and thousands and tens of thousands of licenses for XP and Office. They're the ones that don't want to have to partake in massive upgrade or replacement programs for two year old computers, replace three year old printers, and then go through the retraining and administration unknowns that come along with this new operating system.
I work for a smaller company, but we've put off any consideration of upgrading from XP until 2010, by which point it's possible that Windows 7 will be out. By that point, supporting old hardware won't be much of an issue, and retraining won't be, because all those poor suckers who bought their Vista machine at a big box store will hopefully have learned the basics.
Vastly Different Models (Score:5, Interesting)
The driver situation wasn't any better when XP was launched. If anything it was much worse because all of a sudden consumer-grade hardware vendors had to jump to supporting the NT kernel rather than the 9x kernel, which finally locked down the memory isolation so that a user-mode app could not access kernel resources. It took years for the big companies like Creative Labs, nVidia or ATI to get half-decent drivers out for XP. The situation for Vista is already much better than it was for XP.
Re: (Score:2)
(Windows XP hardware support sucked at launch, and not just because of the switch from 98's driver model. I seem to recall reading warnings about various Windows 2000 drivers that should in theory work not actually being compatible with XP. Sound familiar?)
Re: (Score:2)
Re: (Score:2)
It was much the same for Win2k. Some NT4 drivers worked, some did not, and some were shaky.
More reason to avoid release dates. (Score:5, Insightful)
I can understand a statement such as "We hope for our product to be ready by [date]" or "We're aiming for a possible launch window of [date]", but to say "Our product will be available on this date" only puts pressure on those lower down the totem pole, and can result in a lot of lost features or quality assurance.
Conversely, this should not be used to infer the Duke Nukem Forever will be an awesome game if it is ever released.
Re:More reason to avoid release dates. (Score:4, Insightful)
Or you can try that when you have to pay your bills, "Yeah, we'll make our best effort to pay that mortgage on the 10th."
The world doesn't work that way.
Re: (Score:3, Insightful)
The world doesn't work that way.
Yes, having a firm release date may snag more customers, but you have to look at the end product and decide if the backlash will outweigh the praise.
Re: (Score:2)
Let's say you decide to start a company building a great product.
Do you know what the absolute drop-dead release date is? It's the day your funding runs out (whether it be personal or VC).
You think you can ask your mortgage company, your phone company, your VC to essentially extend you an indefinite loan because your software is not ready to be sold yet?
Re:More reason to avoid release dates. (Score:4, Informative)
Internal schedules, release dates, etc., those should always be present. If you don't have any internal dates, there's no motivation for your workers, as they'll just "get it done when it's done".
Published release dates are what can cause the problems. If you tell your employees "We need this by March", that's one thing. That's also something you can pass along to your business partners. But when you come out and tell the public "Our product will be out in March", and the product falls excessively short of expectations, or does not even make it out of the gate, that's when you create a problem, all for the sake of marketing.
Re: (Score:2)
Shareholders means 'people who own the company', and they get to demand answers to trifling questions about when the company is actually going to start selling new products.
By law, this information can't be kept secret.
Re: (Score:2)
Especially since we've recently been given an estimated release date.
Re: (Score:2, Funny)
Re:More reason to avoid release dates. (Score:4, Insightful)
Three Kinds (Score:2)
Microsoft really screwed the pooch on this one. If they were going to break Windows, they should have broken it completely, and wrote a new OS that resembled Windows, but didn't have a single thing in common. IBM had a similar situation in OS/2 (and PS/2) and screwed it up completely, by marketing OS/2 completely wrong (Windows on Roids).
If I had 100 Million Dollars startup, I'd have a complete OS and New HW platf
Re:More reason to avoid release dates. (Score:5, Insightful)
Re:More reason to avoid release dates. (Score:5, Insightful)
Remember, though, MS had sold all those software upgrade contracts with the stated timeline of having the new version out before they expired, this is why Vista was released to business before the user version was available.
Delivering an item on time and not "when it's ready" can be worth gobs more money to people who like to be able to contain risk. Look at how poorly Apple fares in the business market, for many reasons, but a big one is that they're pretty secretive about their development roadmap and you can't make million-dollar decisions based on Apple's stated trajectory (notice the recent deafening silence over the Xserve RAID EOL and iPhone SDK delay).
Not to say secrecy doesn't pay dividends in consumer segment, but consumers have always been the barnacle on the MS ship.
Microsoft... (Score:5, Insightful)
It is too bad that so many people who would benefit from reading
Vista: Windows 7 Beta (Score:2)
According to the e-mails made public this week, Microsoft will apply the lessons it learned with Vista the next time around. "There is really nothing we can do in the short term," noted Joan Kalkman, the general manager of OEM and embedded worldwide marketing, in a message written a week after Sinofsky's. "In the long term we have worked hard to establish and have committed to an OEM Theme for Win[dows] 7 planning.
"This was rejected for Vista. Having this theme puts accountability and early thinkin
Summary completely misleading (Score:4, Informative)
Or... (Score:5, Interesting)
This is the part that bugs me about this. It might be true that MS considers vista without aero to be fine when they shouldn't. However no one is forcing dell to use the stickers, Dell and the like used them on computers they knew couldn't run vista fully. They do it because they knew people would buy the computer thinking it would run vista.
When it turned out vista was crippled on the machine insted of Dell going "Ooops sorry, heres some store credit" (or whatever) they went "Don't look at us, MS made us do it! blame them!" As if MS was the one who built the computer.
Re:Or... (Score:5, Insightful)
The reason I ask is because it's possible that Dell's contracts with Microsoft did, in fact, obligate them to promote Vista by the stickers on computers. For instance their bulk discounts are tied to various deals, such as having "Dell recommends Windows Vista" on their website.
Also worth noting is that many OEMs were shipping systems with "Vista Ready" stickers long before Vista was finalized. They had no way of knowing how well Vista would ultimately run on the machines, other than what Microsoft was telling them. Still, the OEMs share the blame to the extent that it was irresponsible of them to trust Microsoft and put stickers on systems without being sure that their claims were correct.
Re: (Score:2)
However I knew that none of the machines that said they could run Vista was read
Re: (Score:2)
My Postmortem on Vista (Score:3, Interesting)
FYI: Postmortem also has an informal definition meaning "an analysis or review of a finished event".
Re:My Postmortem on Vista (Score:5, Interesting)
Sure it'll sell just like Windows ME did, purely because of OEM licenses. They'll use that to inflate sales figures, even where people are downgrading back to XP, but we now know just how fucked up things were in 2006.
Re: (Score:2)
Thats consistant with the pre-Vista fuck up day (and Vista's original intent), even if we don't count NT. 95, 98, ME/2000, XP, Vista's original schedule... all 2-3 years apart, give or take. And back then, a few months after a windows launch, the next one was hyped. Heck, its that way in the Linux world too, you
Re:My Postmortem on Vista (Score:5, Interesting)
Much like Windows 2000 was what NT4 should've been, I expect "Windows 7" to be what Vista should've been. Of course, an argument could be made that even what Vista "should've been" isn't what we actually WANT. Personally, I think MS should bite-the-bullet and just abandon backwards compatibility as part of the "base" operating system. Just run everything in a VM, much like Apple did with the Classic MacOS.
Re: (Score:2)
Windows 2000 was a lot more polished, but coming from a *nix background, I really didn't give a damn about polish.
Re: (Score:2)
Re: (Score:2)
2 minute Vista install (Score:5, Funny)
http://youtube.com/watch?v=FVbf9tOGwno [youtube.com]
how dare it not be perfect pre launch!! (Score:2)
It's not just drivers (Score:2)
Dogfood or not, this really seems like a case of the left hand not knowing what the ride hand was doing. Or they overshot. Or they tried too hard and failed to copy OS X. Or they just need to start over. IMHO Sandbox XP->98 apps and start fresh with a whole new
Needless to say, the article exagerates (Score:3, Interesting)
Deja Vu (Score:5, Informative)
Late OS code changes broke drivers...
This reminds me of the painful driver development from NT4 to Windows 2000. A few years before release MS was pushing us to port NT4 drivers to Win2K. We jumped on it quickly and had working drivers, but as the years rolled by changes would be made that broke the earlier work. This rinse and repeat continued to the *very* end. Years of wasted time and resources for no reason.
What I learned from that is to start looking at new Windows driver documentation a few months before release and then wait until the actual release before changing or writing any code. You just don't know what fundamental changes will occur until the discs are on retail shelves.
You sure as hell can't trust what MS tells you as a developer about interface changes and release dates.
Re: (Score:3, Insightful)
Re: (Score:2)
Nice idea, too bad it would not have worked. There are thousands of poorly written but business critical apps that run under win32 / MFC / ActiveX / Com and whatever else has dribbled out of Microsoft's faucets over the years.
Trying to support all of that nonsense in a VM would be a nightmare for small
Re: (Score:3, Insightful)
What has been said, is that stability and security should not be compromised to provide backwards compatibility. Drivers are not usually at the high end of security risks (that's for charming things like ActiveX).
With the fast hardware and cheap RAM we have now a days, there's no reason to start moving to a VM model, allowing for compatibility *and* security. But Microsoft and its unholy OEM alliance with manufacturers are not interested in that, but rather in forcing upgrades of both
bullshit (Score:2)
Re:Pot and Kettle (Score:5, Informative)
Examples please. All my devices work the same or better under SuSE or Ubuntu.
All my purchased Linux (Loki) games still work.
I can't say that for my $300 Microsoft Office 6.0 purchase under Windows. I can't say that either for the Windows games I've purchased over the years.
Enjoy,