One immediate (and simple) solution is what we're already working with now under our 2D environments -- the mouse cursor is contrained to be in the 2D plane that is the screen surface, and whatever is clicked on is whatever can be seen at the point the mouse is at. I.E. whatever is front-most.
In our current 2D environments we sometimes have to use the "Lower Window" command, which un-obscures other windows which can can then manipulate. Not as effective as having 3D input, but it works. Further, some users aren't going to be comfortable using a 3D cursor.
So you start with what everyone already has and is comfortable with, and then you add additional input options as time goes on.
Re:HOW did you get the voodoo 3 to run q3test!?
on
John Carmack on Linux
·
· Score: 3
I used the Glide drivers from the www.linux3d.org site, and the Mesa out of CVS. Compile Mesa with linux-glide support, and I was there. Oh, and be sure to install the 3dfx module in your kernel, otherwise you have to run as root to get the 3D accelerated displays.
An excellent resource for information is the 3dfx.glide.linux news group, available on the news server news.3dfx.com. There are lots of messages from people detailing how to get things running on the various distributions and versions.
But please note that Matrox haven't fully released their specs -- the setup engine has not been documented, and that's where a large part of the hardware acceleration comes in.
My opinion: I'm willing to accept closed drivers to extremely hi-end cards, particularily 3D hardware. In this space, the intelectual property is truely all these companies have. Giving away the register specs will reveal way too much detail on how they're doing things.
Although Open processes are often superior to closed, they are not always. 3D accelerated hardware solutions involve hardware, and thus a whole new area of hard costs have to be managed which OSS projects don't tend to deal with well.
But then, I run Word Perfect from time to time too, and I've never been given the chance to look at it's source either.
Just thought it worth commenting on... John mentions that he's testing Q3Test on a G200 Linux box, and that it's only "bordering on playable with all quality options set to the minimum on a fast computer."
As a comparison of what's possible right now, I've got a Voodoo3 in a Celeron 300A@450, and am able to get 35 to 50fps in Q3Test at 1024x768 with all the quality options set to their maximum! (...under Linux, of course.)
Check out the www.linux3d.org site for more details on the Glide driver which makes this possible. It's not Open, but it's free and it's here now.
Disclaimer: I don't work for 3DFX, I'm just in awe of one of their cards.
I just put a V3 in my machine. WOW!!! Daryll Strauss has released a new X server for the whole like of Voodoo cards, and it screams!!! Although Glide (and thus, Mesa) isn't available for it yet, the talk is we'll likely see betas out in a few weeks.
Having 16 megs on your video card gives whole new meaning to textures! It can also double-buffer displays at upto 2046 by 1536.
Disclaimer: I don't work for 3dfx -- I'm just in awe of one of their cards.
A good article, very solid in the economic facts -- unfortunately the implementation of the revenue stream won't work in the field. The open source community would not allow the type of control COSS would impose.
What would work though is a much simplier solution: support companies (e.g. LinuxCare) could begin supporting one or more OSS packages, and contribute a portion of the revenue to pay for one or more developers of those packages.
It's quite common in the Closed Software world to have a user call up the ISV and be charged $50 to $150 per incident. After the 1-800 call costs, the support human(s) and the knowedge base there's still lots of profit there.
A company doing this would gain respect within the community, something the article quite correctly points out is very important in this culture. The project gets full time developers to keep things moving and it remains Free.
There you go. A revenue stream without taxation nor per-use/per-seat restrictions and entirely demand-side driven. Not everyone who contributed to a project gets monetary compensation, but frankly that's never going to happen. Nor is it every contributer's goal.
This is another example of the mainstream press not understanding the process.
It might not be pretty when these blow-ups occur, but by having it all out in the open I think the matters are settled quicker and generally for good. When things must be settled publically, it's much harder to try to push a hidden agenda 'cause someone is going to see through it and call you on it.
Conflict is normal. The press just isn't used to seeing it behind the nice, warm and fuzzy facade which wraps traditional development processes.
Damn. It's amazing what a difference a word makes... That should read "CORBA-like".
In our current 2D environments we sometimes have to use the "Lower Window" command, which un-obscures other windows which can can then manipulate. Not as effective as having 3D input, but it works. Further, some users aren't going to be comfortable using a 3D cursor.
So you start with what everyone already has and is comfortable with, and then you add additional input options as time goes on.
An excellent resource for information is the 3dfx.glide.linux news group, available on the news server news.3dfx.com. There are lots of messages from people detailing how to get things running on the various distributions and versions.
Enjoy -- it's pretty sweet!!!
My opinion: I'm willing to accept closed drivers to extremely hi-end cards, particularily 3D hardware. In this space, the intelectual property is truely all these companies have. Giving away the register specs will reveal way too much detail on how they're doing things.
Although Open processes are often superior to closed, they are not always. 3D accelerated hardware solutions involve hardware, and thus a whole new area of hard costs have to be managed which OSS projects don't tend to deal with well.
But then, I run Word Perfect from time to time too, and I've never been given the chance to look at it's source either.
As a comparison of what's possible right now, I've got a Voodoo3 in a Celeron 300A@450, and am able to get 35 to 50fps in Q3Test at 1024x768 with all the quality options set to their maximum! (...under Linux, of course.)
Check out the www.linux3d.org site for more details on the Glide driver which makes this possible. It's not Open, but it's free and it's here now.
Disclaimer: I don't work for 3DFX, I'm just in awe of one of their cards.
Having 16 megs on your video card gives whole new meaning to textures! It can also double-buffer displays at upto 2046 by 1536.
Disclaimer: I don't work for 3dfx -- I'm just in awe of one of their cards.
What would work though is a much simplier solution: support companies (e.g. LinuxCare) could begin supporting one or more OSS packages, and contribute a portion of the revenue to pay for one or more developers of those packages.
It's quite common in the Closed Software world to have a user call up the ISV and be charged $50 to $150 per incident. After the 1-800 call costs, the support human(s) and the knowedge base there's still lots of profit there.
A company doing this would gain respect within the community, something the article quite correctly points out is very important in this culture. The project gets full time developers to keep things moving and it remains Free.
There you go. A revenue stream without taxation nor per-use/per-seat restrictions and entirely demand-side driven. Not everyone who contributed to a project gets monetary compensation, but frankly that's never going to happen. Nor is it every contributer's goal.
It might not be pretty when these blow-ups occur, but by having it all out in the open I think the matters are settled quicker and generally for good. When things must be settled publically, it's much harder to try to push a hidden agenda 'cause someone is going to see through it and call you on it.
Conflict is normal. The press just isn't used to seeing it behind the nice, warm and fuzzy facade which wraps traditional development processes.