New Steam survey results

[Groo The Wanderer]

It's not like MS hasn't questionable things before but I'm really curious about this. Could you elaborate on this topic?

Anyway, what prevents you from having an <insert fave Linux/BSD> firewall between your Windows box and the router that drops those packets?

Sure. MS was caught stealing personally identifiable info in XP, they backpedaled, and did it again when they felt eyes were off. For Vista, they changed that bad behavior, they just encrypted it all so you couldn't tell what was being transmitted, and they flat out will not tell you what it is. I have asked.

http://news.softpedia.com/news/Forg...s-Harvest-User-Data-for-Microsoft-58752.shtml

There are a few dozen other links like this you can find if you want, but this is one of the easiest reads. All you do is install a windows box and put a logging firewall between you and the net. Log every packet and connection to known MS domains. If you want to be really scientific, install a clean box, turn off power management so it doesn't sleep, turn off all automatic updates and such checks, and let it sit in a corner for a month or three. There should be no traffic to MS.

Unfortunately, there is, and there is lots of it, to places, from processes that have no explanation of what they are sending and why. It is much harder to trace on the windows side, but if packets are going back and forth, packets are going back and forth.

What is in them? Beats me, modern encryption, even when implemented by MS, works well enough to keep me out. If you ask MS, they will deny it. If you show them the traces (I haven't done this, but others have), they will point to the EULA and say it is OK. They will NOT however tell you what you need to know.

To me, this is unacceptable. I require the ability, not want, but require, to control what comes in and out of my box. People talk to me because they know it will be confidential. I don't mind that MS needs some data, I understand that. I have to know what is going, and be able to manually decline to provide it. If that causes problems, I am OK with that as long as it is my choice.

MS also pushes things to your PC, and simply will not apologize or put rules into place to say they will not do it again. The main link is on Windows Secrets, but that is down right now, so I can't link it. There is a summary here:
http://www.seattlepi.com/business/331579_windowsupdates14.html

Note the MS response. That should scare you silly. They have been caught several times since doing this on other DLLs and programs. Scary. Really scary.

As for firewalling externally, that is what I would personally do, but MS is crafty there too. With the Broken OS and the Broken OS SP7, MS forces you to check in with an activation server. Last time I checked, it was once every 30 days, and after 90, it turned off. I know MS changed the terms on this several times, so it might be out of date.

If you firewall off the MS domains/IPs, Windows stops working. Wonderful, eh? If you open it up for short periods of time, MS dumps all the data as soon as you open it up. Crafty buggers. No, just buggers.

What they are saying is that if you don't allow them to steal your data, they will turn you off. That part they can do in the EULA for basically any reason, and you have no recourse. Why anyone would touch this festering pile of an OS is beyond me, but some people really are that stupid.

XP does work if you get a version that is pre-WGA, and don't install any of the WGA patches it asks for. This is what I do on the one XP box I have for testing. Yes, you could crack Vista/7, it is trivial, but why do I want to commit a felony to promote their products? Screw that, I put my time and effort into Linux now where I can make a difference. MS is no longer worth the trouble.

-Charlie

 

[Groo The Wanderer]

So I see you had no actual retort for the ring level zero nonsense you're riling against. Well done.

And yes, I've taken a packet sniffer to several thousand Vista machines. They don't "send home" anything like what you're describing. And unless you have some rational and factual data to back up your claim (which you will post here for peer review), then I think you're deluded at best -- or being purposefully antagonistic, or just entirely stupid.

Sorry to use big words at you, I will try to connect the dots in smaller steps. Back in the days of NT 3-4, OS/2 was not a competitor, it was an outlier. Had they merged, that may have changed things, but that story is pretty well played out.

The competitor for NT was Netware, in 3.x and 4.x guise. Unix wasn't really a competitor on the small scale server world, or at least I never saw it, and I have been running small/mid-size networks since the mid-80s. Things were mostly NW, or mixed NW and NT environments.

On the server side, NW didn't have a gui, it had a bare minimum green/text screen for the few admin jobs you could do locally. It didn't have a GUI, and didn't have any 3rd party graphics drivers in ring 0. I am not sure how the various unices of the day did things, but I would guess that they weren't in Ring 0. Anyone know?

So, is that clear enough for you? Need a diagram?

-Charlie

 

[Silent_Buddha]

Sorry to use big words at you, I will try to connect the dots in smaller steps. Back in the days of NT 3-4, OS/2 was not a competitor, it was an outlier. Had they merged, that may have changed things, but that story is pretty well played out.

The competitor for NT was Netware, in 3.x and 4.x guise. Unix wasn't really a competitor on the small scale server world, or at least I never saw it, and I have been running small/mid-size networks since the mid-80s. Things were mostly NW, or mixed NW and NT environments.

NT was always primarily a workstation OS with the server OS being a sideshoot. As such OS/2 wasn't an outlier it was the main competitor at the time. As time went on, MS got more serious about NT being used as a server, but the server was always secondary to the primary role as workstation.

BTW - isn't this sidetrack starting go waaay off-topic? Maybe split it into its own thread?

Regards,
SB

 

[neliz]

Nobody reads the link I posted above.

Clearly, Windows NT wasn't a competitor to OS/2 untill MS decided to start Win32, before that NT was supposed to RUN OS/2.

The decision to drop OS/2 for Windows forever damaged the relationship between the two companies. "But we had executive approval, and started the port," Lucovsky said. "So instead of working on an OS/2 subsystem for NT, we picked up Win32." At that moment, he said, the product became Windows NT.

http://www.winsupersite.com/reviews/winserver2k3_gold1.asp

And yes, a windows topic please!

 

[Albuquerque]

The competitor for NT was Netware, in 3.x and 4.x guise.

And yet, drivers in NetWare 3 ran entirely in Ring 0. Netware 4 offered the admin the ability to 'decide', but Ring 0 was the preferred location for performance.

Oh wait, didn't you say competitors weren't doing this? Are you wrong yet again?

The net result is, Windows (and Microsoft) were doing things to remain competitive by looking at their competitors. They've continued making changes to remain competitive, even in the face of outlying whiners who can't deal with change.

Here's hoping that more change is in store for "Windows 8".

And I'll wait until there's a valid "Windows" topic to dice up your arguments about the internet traffic.

 

[aaronspink]

There are several people in this thread complaining about things that Microsoft hasn't fixed -- but could if they so choose. And look what happened when they took something that was fundamentally broken and fixed it? We get people complaining about "some improvement but at the cost of compatibility."

you don't get credit for explicitly breaking your model then realizing that you broke it when you admitted up front you were breaking it then finally going back and fixing it cause you broke it. If you understand the history, this all makes sense...

Doing these things required new driver model, new kernel interfaces, and a LOT of work. Backporting to XP could've certainly happened, say as part of a whole new service pack. But then we'd be also needing a whole new slew of drivers at the same time. Talk about a huge pile of fail when you take an already functional OS and replace such a massive chunk of it, and then require all the 3rd parties to follow suit.

Yes, but DX10 != WDDM.

 

[Silent_Buddha]

Nobody reads the link I posted above.

Clearly, Windows NT wasn't a competitor to OS/2 untill MS decided to start Win32, before that NT was supposed to RUN OS/2.


http://www.winsupersite.com/reviews/winserver2k3_gold1.asp

And yes, a windows topic please!

OK my last post on this unless it gets split into a new thread...

Thanks for that Neliz, I didn't know much of the details prior to Windows NT being released to the public. Fascinating.

But while it wasn't orignally meant as a competitor to OS/2, it turned into a competitor during developement, so for all intents and purposes they were once NT was released.

Regards,
SB

 

[Silent_Buddha]

Yes, but DX10 != WDDM.

Yes but Dx10 functionality required WDDM.

Regards,
SB

 

[aaronspink]

Actually, if you recall (and you probably don't, because you were wrong about competition not doing it...) NT4 was going to be the point where OS2 and NT met in the middle and became "OS2 v3." Remember? And if you remember that, you'll remember that OS2 placed video drivers in Ring 0 long before NT3.5 was around... As operating systems take quite a while to design and implement, and the original design for OS2 v3 (later NT4) called for an amalgamation of the NT3.5 and OS2 operating systems, I have a feeling that inclusion of driver models into Ring 0 were a combination of decisions from MS and IBM. Likely for performance reasons, sure, but since Ring swaps from user to kernel mode take many hundreds (or thousands) of cycles to complete, and we're talking about the days of 386 and 486 hardware, I think it was less of a "benchmark" win and more of a "making the OS fundamentally responsive" win.

BUNK!

MS was done with OS/2 before NT even shipped.

 

[Albuquerque]

you don't get credit for explicitly breaking your model then realizing that you broke it when you admitted up front you were breaking it then finally going back and fixing it cause you broke it. If you understand the history, this all makes sense...

Unless you're going to explain the history in a way that I don't already know, every mass-deployed competitor to NT3.5 and NT4.0 at this time was running drivers in Ring 0 to gain performance benefit. I see no reason why Microsoft "broke" anything by moving their drivers into Ring 0 like all their competitors were doing. This change made perfect sense at the time; there's little reason to throw them under the bus because they couldn't see that 10 years later this might actually be a security concern.

When they fully realized the error, they solved it. Isn't that what we were just talking about? People bitching about them solving problems? The Windows API has a big pile of holes, and one of these days when they gut it and "fix it", people are going to be bitching again. Should they have never invented the API? I think that's a bit of a stretch...

Yes, but DX10 != WDDM.

So show me DX10 without WDDM. When you can, your response will have logical merit. If you cannot, then your response is as vapid and intellectually dishonest as Charlie's nonsense.

 

[Albuquerque]

BUNK!

MS was done with OS/2 before NT even shipped.

NT4 was indeed going to be the summation of OS2 v2 and NT3.5. Unless you're trying to say something else? There's a massive amount of documentation to this effect, specifically around Microsoft ditching OS2 for the Win32 process and the massive pissing match it started with IBM thereafter.

Again, I think you're being quite dishonest in your replies.

 

[aaronspink]

NT was always primarily a workstation OS with the server OS being a sideshoot. As such OS/2 wasn't an outlier it was the main competitor at the time. As time went on, MS got more serious about NT being used as a server, but the server was always secondary to the primary role as workstation.

NT was as much a workstation OS as VMS was a workstation OS.

 

[Albuquerque]

NT was as much a workstation OS as VMS was a workstation OS.

I think that description gives too much credit to NT Seriously though, as good as VMS was, it was also fairly "niche" as a product.

Windows was far more in the public eye, and NT's ability to run Windows applications in a preemtively multitasked and fully 32-bit environment was fantastic at the time. I vividly recall a very high number of NT4 workstations in college and at my first several IT jobs.

Granted, most "general users" were on Win31 or Win95/98, but NT4 was quite prevalent for anyone doing real work.

Edit: I still generally agree with you that, IMO, Windows NT was originally targeted as a server operating system rather than a workstation, although I also think that Microsoft realized this had at least moderate potential in workstation form.

 

[aaronspink]

Unless you're going to explain the history in a way that I don't already know, every mass-deployed competitor to NT3.5 and NT4.0 at this time was running drivers in Ring 0 to gain performance benefit. I see no reason why Microsoft "broke" anything by moving their drivers into Ring 0 like all their competitors were doing. This change made perfect sense at the time; there's little reason to throw them under the bus because they couldn't see that 10 years later this might actually be a security concern.

There were specific reasons that NT originally didn't have drivers in ring 0, it wasn't just some happy accident. The motivation minimally by performance. One of the major issues in the transition to ring 0 was to accelerate driver development as MS started to push NT into the mainstream.

When they fully realized the error, they solved it. Isn't that what we were just talking about? People bitching about them solving problems? The Windows API has a big pile of holes, and one of these days when they gut it and "fix it", people are going to be bitching again. Should they have never invented the API? I think that's a bit of a stretch...

They didn't "realize the error", they made a determined design change LATE in the version history to a less reliable model. They KNEW all the issues going in. Once again, you don't get credit for putting out forest fires that you start!

So show me DX10 without WDDM. When you can, your response will have logical merit. If you cannot, then your response is as vapid and intellectually dishonest as Charlie's nonsense.

Take the SMs from DX10, run in DX9 environment. WDDM has little to do with actually running the SM code.

 

[neliz]

Edit: I still generally agree with you that, IMO, Windows NT was originally targeted as a server operating system rather than a workstation, although I also think that Microsoft realized this had at least moderate potential in workstation form.

Again, from the supersite

Novell was ambivalent about the NT desktop – they didn't know if they wanted to build a client. We offered our assistance, but they kept messing around and ... well. We did our own. And it just blew them away. Ours was the better Netware client, and customers used ours for years, even after they finally did one. That client enabled the NT desktop, because Netware was the prevalent server in the market. We wouldn't have been able to sell NT desktops otherwise."

That's their take on the story, their good netware support basically created a use for NT Workstation.

 

[ShaidarHaran]

LOL, an entire page about OS/2, early NT derivatives and NetWare in the Steam survey thread.

 

[aaronspink]

NT4 was indeed going to be the summation of OS2 v2 and NT3.5. Unless you're trying to say something else? There's a massive amount of documentation to this effect, specifically around Microsoft ditching OS2 for the Win32 process and the massive pissing match it started with IBM thereafter.

Not from anyone that I know at MS/IBM at the time. The MS/IBM split was all but signed by the lawyers at the time on the NT / OS/2 v2 release.

Certainly, before the original NT release, MS had determined they were going to have nothing further to do with OS/2.

 

[aaronspink]

I think that description gives too much credit to NT Seriously though, as good as VMS was, it was also fairly "niche" as a product.

Um, not really. Same people. pieces of the same code. Interesting agreements to support a random CPU architecture that was owned by the company that also owned VMS and had access to the NT source code and plenty of lawyers.

I vividly recall a very high number of NT4 workstations in college and at my first several IT jobs.

MS was basically paying the cost of the hardware in order for universities to install NT workstations. Its basic marketing, free workstations or pay for them?

 

[neliz]

MS was basically paying the cost of the hardware in order for universities to install NT workstations. Its basic marketing, free workstations or pay for them?

I'm not advocating piracy here, but everyone able to fill in 040-1111111 in the CD-Key box was the happy owner of a fully working workstation or (enterprise/terminal) server.
Even for our machines that were under a 5000 client license used that code.

Piracy was easy back then, M$ widespread.

 

[Albuquerque]

There were specific reasons that NT originally didn't have drivers in ring 0, it wasn't just some happy accident. The motivation minimally by performance. One of the major issues in the transition to ring 0 was to accelerate driver development as MS started to push NT into the mainstream.

So it sounds like you have better (ie, insider?) information as to why the move was made. Can you elaborate further, because I'd very much be interested in knowing. I'm very serious, this is the kind of stuff that I like to learn about, and there isn't much information that directly tells us why MS made the move, other than the inference that all their competitors had moved drivers to Ring 0 for (supposedly) performance reasons.

They didn't "realize the error", they made a determined design change LATE in the version history to a less reliable model. They KNEW all the issues going in. Once again, you don't get credit for putting out forest fires that you start!

Again, I don't see how their decision to move drivers to Ring 0 was somehow "bad" given the era. We have plenty of examples of their competition having no issues with it. Maybe when I hear your expanded story about why they made the change late in the development cycle (I'm taking your word on this, although I cannot find any documentation that specifies either way) then perhaps I'll have a different take on it.

For now, with the information that I have, I can only conclude that they did what was right for the time period when the development occurred.

Take the SMs from DX10, run in DX9 environment. WDDM has little to do with actually running the SM code.

DX10 is not entirely comprised or limited to SM. If it were that simple, you could immediately run DX10 rendering paths on XP right now without issue, with no exceptions.