Questions on WGF 1.0
- Thread starter azopi
- Start date
Chalnoth said:
If in terms of standalone graphics sollutions even a 5200 will do fine, I don't see why a NV4x could signify any sort of shortcoming for Longhorn.
In fact I'd say that both major IHVs were aware a long time ago of future requirements and had designed the widest majority of their current and former generation (SM2.0 and upwards) sollutions with those in mind.
I couldn't care as much if an OS environment is 2D or 3D in the end. What I could make more use of is heavier multitasking or if you prefer quite more faster and advanced task switching. The only other thing I would worry about are faster and/or more flexible hard drives....
Unknown Soldier
Veteran
Ailuros said:
True .. but then I expect that from the OS anyway. That is why they release new OS's to make things faster and easier. The interface to that though is important and this is where the 3D part comes in. If it's friendly, easy to use and overall looks great then I'll use it. If not then the classic desktop is what i'll use.
US
I think you're misunderstanding my point. My point was that if Microsoft wants to put forth a new iteration of their OS that is out there and the only reason for people to upgrade to said OS from previous Windows versions is that it provides a more compelling user experience, then I'm fine with it. If it provides new software interfaces that are not then exposed to previous Windows versions, such that I am forced into upgrading when new software comes out exclusively for this OS version, that's when I get upset.Ailuros said:
Well, no, they release new OS's to make more money. There are two ways, however, to get people to purchase the new OS:Unknown Soldier said:
1. Make it a truly compelling improvement to the user experience such that people want to upgrade.
2. Allow incompatibilities with previous OS versions such that new software that is written for the new OS cannot "just run" on previous versions, forcing people who want to run new software to upgrade.
One is okay. Two would piss me off to no end.
Chalnoth said:
Would you be happy to pay for the support of it through previous OS versions?
Anyway, looks like a Beta is being aimed for June '05
http://news.com.com/Microsoft+Longh...by+June/2100-1012_3-5566423.html?tag=nefd.top
Actually, if Microsoft forces an upgrade, what I'm liable to do is switch my system entirely over to Linux and start using something like WineX (A Windows emulator that has subscription fees, user support, and DirectX emulation). Assuming the performance is reasonable, this will actually be more convenient for me as I do all of my work in Linux.
But, to get support in older OS's would you be willing to pay for it?
No. Microsoft has released updates to DirectX that have worked with all of their operating systems back to Windows 95. I see no reason why that must change now. Once again, if the product stands on its own as something that would be useful to upgrade to, but is not necessary to upgrade to, then I consider it fine.DaveBaumann said:
Because it costs money to do these things?
Chalnoth said:
The DirectX DDK is primary still the same as in the old days. Because of this it was easy to write runtimes that work with the different operating systems that have a driver model that is compatible to this DDK.
With Longhorn the driver model is changed. As a result of this Microsoft have to rewrite all DX runtimes down to the first version to make sure that games that use this DX version will still run with Longhorn. WGF 2.0 is build on this new driver model. If the want to make it run with older versions of windows they have to write a complete new runtime that works with the old model. And they have to extend the DDK again. Even than it is not sure that WGF 2.0 will run very well with the old systems because the old driver model is very slow compared to the new the build for Longhorn.
Chalnoth said:
Actually no they haven't, DX9 doesn't support Win95... its oddly to do with test matrices, to many OS options costs them a fortune to test so they don't support any OS they no longer see as viable for high end games.
I.e. exactly who runs HL2 on a Win95 powered box?
Well, that's a reason why Longhorn should cost money, not why Microsoft shouldn't continue supporting API updates for previous OS's.Demirug said:
Well, why not? There are a few big reasons to add support for the new interface to older versions of the OS:
1. It's not as monopolistic (this is my big reason for wanting it to happen).
2. It'll ensure wider software implementation of the new features.
Very slow? Heh, I doubt there will be any significant difference for most situations.
Well, I guess that makes sense. But we are talking about an OS that was around 8-9 years old at the release of DX9. I'm more concerned with required upgrades in a 1-3 year span.DeanoC said:
Edit: But I do have to add that I was rather pissed when my build of Windows 95 had no support for USB, way back in the day.
From what I gather so far the changes to the API (from DX9c to WGF 2.0) are they are listerning to the feedback from game developers and MS is working with the IHV's to to solve some of the issues/problems with the older DX API models.
A couple of the such things that sticks out in my mind from the presentations was fast small batch support and things like decent VPU/GPU profling.
A couple of the such things that sticks out in my mind from the presentations was fast small batch support and things like decent VPU/GPU profling.
Well, sure there are improvements that need to be made to DX. Doesn't mean I should be forced to pay Microsoft if I want to play new games. If I'm going to give Microsoft any money, it should be because I want to make use of their product, not so that I can play a game made by somebody else.
Its up to the developer which API they support, hence which platform its aimed at. To get significant new "game feature" functionality on a console you have to buy a new console, there's little different here if this transpires.
When you buy a console, you are purchasing specific components, such as the hardware to run games and the control interfaces. Frequently the cost of these components to the consumer is actually rather low compared to the value of the hardware. After all, console makers make most of their money through licensing fees from game sales.DaveBaumann said:
In this same vein, I have no problem purchasing new hardware. I do it all the time. But I don't feel I should need to purchase a new version of Windows because it is no longer API-compatible with current versions.
Now, what I think would make more sense is extend DirectX to have the API interfaces available to Longhorn, but through the current DirectX DDK interface. This would probably suffer from the same performance problems Direct3D currently suffers from, for instance, and so there may be added value in upgrading to Longhorn. This is fine. But a forced upgrade would not be, in my mind.
Edit: Note that lower performance is a slippery slope. It should not be artifically low. That is, in no instance should an application that can be made to run in the exact same way in "DirectX Next" or DirectX 9 run faster in DirectX 9, and the added features in DXN would be optimized well enough that they lead to roughly the same relative performance characteristics as they do in Longhorn compared to a DX9 program running on that OS.
Chalnoth said:
Its pretty much inheriting a driver model similar to OpenGL with its more kernel side based design (though unlike OGL, MS are writing most of it with only a small kernel driver provided by the IHV)... Given Longhorn is probably doing lots of calls, it could make a big difference speed-wise