I am a developer on the Windows team, I think I understand quite a bit.

Perhaps you should re-read the post to which you replied. A password prompt is useless if the OS doesn't check actions (like attempts to access a file) to compare the account the user authenticated with against the permissions or access control lists on the file or other securable object. Do you really think that having user accounts and passwords doesn't involve any additional work after you've logged in? Quite the contrary, nearly every action you perform results in access control checks, on any operating system you choose.

If you don't understand that then you don't understand how software works.

Admittedly I confused you with the person who posted the original DRM comment. You have my apologies for that. But still, you said:

"It seems to be both by "common-sense" and "real figuring" (like in the article), all the DRM stuff that was added in Vista does have some impact. Even on those who wouldn't be using DRM'ed files."

That seems to me like you're propogating the same disinformation. Even the phrase "all the DRM stuff that was added to Vista" is a loaded statement that misrepresents the facts. In truth, there wasn't that much DRM stuff added to Vista, just some improvements on support that was already in Windows XP.

Sorry, your post was a little confusing so I wasn't sure what you were trying to get across there =P

Why would support for DRM restrictions impact performance or stability? That's just ridiculous. That's like saying having a password on your computer will make it slower. Well, there's more code to handle password authentication, and yeah the OS needs to do the appropriate checks to make sure you're authenticated. But does it make the system less stable? Of course not. Does it slow it down in a perceivable way? Of course not.

More importantly, "support for DRM'd media" and "support for the restrictions of DRM'd media" are the same thing. If you want to play BluRay movies, you have to accept the limitations imposed by the BluRay standard. That's just the way it is, at least in the world of legally licensed software.

I have no complaints about it. Most of my machines never play any such material, and so no DRM-related code ever runs on them. My media center PC has a BluRay drive connected to my Samsung LCD via a DVI to HDMI converter, and I play BluRay movies with PowerDVD at 1080p and use an optical output connection to my receiver. I don't know what DRM or other content protection measures PowerDVD employs or makes use of, but I do know that I have never experienced any problems at all with that setup.

That issue had nothing to do with DRM, it had to do with CPU scheduling / network throttling behavior designed to provide a better media experience, and a bug / oversight in the associated code.

It's totally bogus. None of it makes any sense or applies to the actual Windows Vista OS, and none of its predictions about Vista ever came true.

It said Vista would end unified drivers. It did not.
It said device drivers would become more complex / less reliable, and they have not. In fact, quite the opposite is true.
It says that drivers or devices coudl be and would be "revoked." This has never happened and there's no basis for claiming it's even possible.
It says that somehow hardware will cost more, but Vista (and Win7) runs on the same hardware as XP or Mac OS.
It makes all sorts of ridiculous claims about how Vista requires devices drivers to "poll" for some undescribed something or other that will make things slow, but such requirements do not exist. Heck, Vista supports the very same drivers that worked on Windows XP for most hardware.

It quotes hearsay about mfpmp.exe using CPU cycles while playing MP3 files, while not referring at all to the PMP documentation or noting that these are the CPU cycles necessary to *decode the MP3 file* - as mfpmp.exe is the host process for decoders. In the past those same cycles would have been used inside wmplayer.exe. No new CPU usage was added, it was only moved.

I find it especially ironic that you're here evangelizing such disinformation while admitting that you don't actually use the OS and haven't done any research at all into the authenticity of the claims made by the single document you're using as the basis for your claims. It's ironic because these are very much like the so-called Fear, Uncertainty, Denial attacks which Slashdotters used to blast Microsoft for making against its competition. I guess what's good for the goose is good for the gander, eh?

The OS does no checking on files you work with. Windows Media Player will of course check media files you play to see if they are DRM'd, just as every media player does. This is part of the process by which it determines the media format type, i.e. is it an MP3? A WMA? A DRM'd WMA? This works exactly the same on Vista as on XP, and is the same sort of check iTunes and other apps do. There is no overhead here beyond that required to determine the difference between an MP3 and a WMA, and it only occurs at the time a file is loaded.

What Vista provides above and beyond XP is something called Protected Media Path. This is similar to Windows XP's Certified Output Protection Protocol, but more sophisticated.

Protected Media Path allows for a great many possible restrictions on both audio and video output, including those defined by HDCP. And since it is a protected process, it allows decoders to be run in a context that prevents them from being manipulated or attacked (i.e. having their memory scanned for secret keys and such). Whether there is any overhead for using PMP to host your decoder is debateable, if there is any it should be negligible at worst. It's the same code running, whether it runs in your app's process or in mfpmp.exe instead. Yes there's IPC overhead, but lots of media player do out-of-proc hosting anyway so that a bad decoder won't crash the media player and to dodge certain security issues like heap spraying attacks.

What's important, though, is that this is an API that applications *CAN* use, not something that is imposed on applications or users. The Protected Media Path code will only get loaded and used if an application specifically calls MFCreatePMPMediaSession. It will cause the application-provided code to be hosted inside the protected process (mfpmp.exe). You won't see mfpmp.exe running unless an application has specifically invoked it via that API call - which would most likely happen because you are playing a BluRay disc with a player that has decided to make use of PMP.

Windows Media Player on Vista does use PMP for all media decoding, and suffers no ill effects from it. However, you do have a problem with it, just use Winamp or some other player that doesn't invoke PMP.

Like I said, having the support for DRM or output protections like HDCP won't affect you at all if you don't use media (or applications) that request or require that support. It just allows developers to run their code in a protected space and place what restrictions they want on their content. It make no determination about whether such restrictions are necessary, wise, or just. It's just an API. APIs can be used for good or for evil, if you have a problem with how a BluRay app uses the PMP API, then complain to Sony or the app developer =)

Again you are mistaken.

You cannot store arbitrary data in GPU memory without incurring significant penalties, in particular if you ever need to read from something stored in video memory as you'll stall the GPU pipe. The cases where it is beneficial are when that data is in a format that allows the GPU to actually do work on the data, where the data won't ever need to be read back into system memory. For rendering a desktop, the 3D acceleration capabilities are useful because they allow the window surfaces to be stored as textures in a DirectX format supported by the GPU, which the GPU can then scale, animate, and composite natively.

You are incorrect in numerous ways.

First, Windows Vista does not support 2D acceleration, as the new WDDM 1.0 driver model doesn't allow it. That means all the GDI acceleration in Windows XP is gone. However, for the desktop and window manager this is offset by the fact that the new window manager makes use of 3D acceleration, which is very richly supported by the new driver model.

Second, many tasks will feel faster with composition enabled. For instance, dragging a window around the screen on Windows XP will cause a great deal of CPU usage. On Windows Vista with the DWM running, there is virtually none.

Finally, since the desktop is composited, it allows for a higher quality and more robust user experience. When a window is hung, the window manager can continue to show the last state of the window (and can even "frost" it over to show that it is not responding), and can still allow it to be dragged around even though the window itself has stopped pumping messages.

Note that Windows 7's new WDDM 1.1 driver model brings back a lot of the 2D GDI acceleration that was missing in Vista.

Indeed. How many times has the Finder seen a major rewrite since 10.0?

In Windows 7 about half of the Explorer was rewritten from scratch. I should know, I work on it ;)

There have been fully 64-bit versions of Windows for several years now. Microsoft is clearly leading the way with 64-bit adoption. Just go to Best Buy and look at all the new HP computers running 64-bit versions of Windows. Most buyers will have no idea, which is exactly the way it should be. Windows 7 includes improvements to help 64-bit adoption as well, such as including top-quality 64-bit codecs in the box for popular media formats, and reducing (virtually eliminating) the memory footprint overhead of a clean 64-bit install versus a clean 32-bit one.

Contrast this with Apple who have announced "64-bit" as a killer new feature for 3 consecutive OS releases. Who knows, maybe Snow Leopard will actually be 64-bit at last.

Windows Vista's performance "problems" have nothing to do with DRM. If you aren't playing back a DRM'd file, then there is no DRM-specific code running, and no penalty of any kind. The idea that Vista had any more DRM code running than Windows XP was a myth propogated mostly buy people who knew it wasn't true, and others who were gullible and believed anything that sounded bad about Vista.

If you don't want DRM, don't buy any DRM'd media. Having support for DRM'd media in the OS (like BluRay / HDCP / etc) has absolutely ZERO impact on people who don't use DRM'd media.

Vista had its issues and they are well understood, there is no reason to make up myths to blame them on.

WinFS was not a user feature. It was an implementation detail, and for the purposes it tried to serve in the Longhorn days, it failed. The innovations on the server side live on in SQL Server, the .NET Entities Framework, etc. The innovations on the Winodws shell side live on in the Windows shell. Windows 7 includes the features that *would have* been built on WinFS. Libraries, metadata-based views, rich pivots and instant search filters (listing all the Authors that are valid for the current location or query, etc), among others. While the backing database isn't WinFS, it is a great deal lighter and faster. WinFS was a managed code behemoth which would have made Vista's performance and compatibility problems look like nothing at all.

Just shut up. That couldn't possibly be farther from the truth.

The topic here is Windows 7, not Mac OS X.