Slow and Open is better than Fast and Closed
|
| Author | Content |
|---|---|
| swbrown Feb 07, 2007 2:54 AM EDT |
Here's hoping we get at least Slow and Open drivers for NVidia cards soon. It'll allow for AIGLX to be a standard thing in distros rather than something you have to poke and tweak to make work. With that and the upcoming input redirection extension for xorg, it'll really open up the field for making creative new interfaces on our desktops. |
| devnet Feb 07, 2007 8:46 AM EDT |
I'm more of the school of thought... "Whatever works better for you...use it." That's a loaded statement because open slow drivers may be what makes you feel comfortable and therefore, they work better for you. Conversely, closed fast drivers may be your cup o tea...and in that case, they'll work better for you. It's up to each individual to choose. Choice is nice. I'm sure there are purists out there that think my opinion here is what is causing all the problems...but I'd have to ask just one simple question if they think this: How is limiting choice when using software (that is, only to free alternatives) a valid thing to do? It should be up to each individual as to what they choose. Just be glad that a person is actually trying Open Source at all in this large Microsoftian controlled monopoly of technology. Perhaps eventually, after they use it for a while, they'll come to the realization you came to...some quicker than others. But don't force their hand...you'll end up turning them off to any ideals you or FLOSS/FOSS can offer. |
| dinotrac Feb 07, 2007 8:55 AM EDT |
Of course, fast and open is better still... |
| DarrenR114 Feb 07, 2007 8:59 AM EDT |
Quoting: Of course, fast and open is better still... You know, some of us have dirty minds .... /me goes to wash brain out with soap. |
| jimf Feb 07, 2007 9:13 AM EDT |
I have trouble faulting people who use the proprietary drivers. If you're into games and stuff requiring fast 3d performance I can see why they 'need' them. Ethical considerations aside, that can be quite a sacrifice to make. However... I would suggest that users owe it to themselves to try working with the xorg driver first. You may find that it does just as well as the ATI one and is much more stable. Again, it all depends on the apps you use and the particular card you have. For me, the xorg radeon driver delivers superb performance in everything except games and I use few of those. What I may be loosing in 3D speed, I more than gain in 2D graphical clarity, and stability. For the radeon driver, 3d has also become available in the current xorg for many of the older cards. Xorg is really starting to live up to its potential. |
| jsusanka Feb 07, 2007 9:21 AM EDT |
I think we should make a proprietary card and have ati's engineer try to reverse engineer drivers for it and then do a comparison. I will take slow and open any day over fast and closed. guess that is why I still buy ati powercolor 7000 cards when I get a chance they work awesome and use open source drivers and have full 3d capabilities and are fast enough for me and well as fanless. I can also run multiple keyboards and monitors off of one computer because they have a dvi and a vga connection so I can have multiple users on one pc at the same time and is t just like they have separate computers. great for secretaries and call centers and help on the electric bills - although the monitors is what killing the electric bills. |
| bigg Feb 07, 2007 9:24 AM EDT |
Devnet: I'd go further and say "People are going to use what works better for them". As I've said elsewhere, free is a feature. It's a good feature to have, but if someone is really interested in using an ATI card, it may be better for them to use the proprietary drivers. Fast might be a better feature than free. > How is limiting choice when using software (that is, only to free alternatives) a valid thing to do? It should be up to each individual as to what they choose. That can be a good thing. If nobody uses proprietary drivers, the choice for a manufacturer is to sell with open drivers vs. not selling at all. A classic externality problem. I'm better off using proprietary software, but I'm better off if everyone else doesn't use proprietary software. That's why ESR's strategy is the dumbest thing I've ever heard. People will use proprietary software no matter what, so there is no way his strategy can eventually lead to open drivers. |
| DarrenR114 Feb 07, 2007 9:41 AM EDT |
Quoting: That's why ESR's strategy is the dumbest thing I've ever heard. People will use proprietary software no matter what, so there is no way his strategy can eventually lead to open drivers. I think it could work depending on how the manufacturers consider using FOSS to lower their bottom line. If they can see a way to lower their R&D and support costs by incorporating FOSS into their development process, then you can bet that's the route they'll go - especially if they see a market that can use their deliverables. Scenario in my mind: 1. nVidia continues to provide closed drivers for Linux. 2. More and more Lusers use the nVidia drivers with nVidia hardware AND register with the company to make sure nVidia knows they've got a growing Linux-using customer base. 3. nVidia looks at ways to cut production costs and noting the demographics of their customer base, sees that there is a X percent share of Linux users. 4. Seeing that it would be possible to cut the number of developers assigned to a given product and still support a larger customer-base, nVidia starts a FOSS project ala Netscape or StarOffice. Note that this purely conjecture on how it *could* happen. And don't expect any cooperation from the software developers employed by manufacturers such as nVidia - would you want to risk your job for the sake of someone else's "software freedom"? |
| jimf Feb 07, 2007 9:47 AM EDT |
Darren, I think that if Intel begins to release discrete graphics cards, we would see a big change in how this is all perceived by the other OEM's. And, almost every Linux user I talk to seems to be eager to go in that direction. The whole scene could get a lot more complex... |
| swbrown Feb 07, 2007 9:48 PM EDT |
> That's a loaded statement because open slow drivers may be what makes you feel comfortable and therefore, they work better for you. Conversely, closed fast drivers may be your cup o tea...and in that case, they'll work better for you. It's up to each individual to choose. The issue here is the "closed fast" drivers are holding us back - basically what you expect to happen with closed drivers as we have to convince the authors to support our new stuff rather than doing it ourselves. ATI users attempting to use the closed driver are mostly stuck using XGL rather than AIGLX because ATI hasn't bothered to add the required extensions. That's why they use the open driver - it supports new, vastly superior technology like AIGLX. The closed driver also has issues with Xen if I remember right. NVidia's closed driver also has issues with AIGLX - it causes strange rendering bugs like this: http://forum.beryl-project.org/viewtopic.php?f=36&t=3337 |
| dinotrac Feb 08, 2007 2:48 AM EDT |
>The issue here is the "closed fast" drivers are holding us back They are not holding anybody back. After all, slow and open is better than closed and fast, remember? |
| swbrown Feb 08, 2007 5:10 AM EDT |
> They are not holding anybody back. After all, slow and open is better than closed and fast, remember? You're taking that reply in the wrong context - the reply's point was that you can't simply choose the "closed fast drivers may be your cup o tea" wrt ATI with no side-effects because you are then held back from using modern software like AIGLX/Xen. It's not just a 'freedom vs. speed' choice, it restricts your use of other software as well. It's the classic problem with binary drivers. |
| bigg Feb 08, 2007 7:33 AM EDT |
> It's the classic problem with binary drivers. I don't know if I'd say it's a "problem" with binary drivers. It is a "property" of binary drivers that sucks. IMO the "problem" is that people buy hardware that only has proprietary drivers. If Novell or Canonical or Linspire decided to get into the hardware business, it would be solved instantly, just pick up the phone and buy a laptop from Novell. If there were at least some easy way to find out which pieces of hardware have open drivers, but there's not. It's an information problem. I believe ATI should be free to exclude itself from selling to Linux users if it wants. But I also believe Linux users have a right to know which hardware to buy. |
| dinotrac Feb 08, 2007 7:40 AM EDT |
>I don't know if I'd say it's a "problem" with binary drivers. It is a "property" of binary drivers that sucks. Of course, the two are entangled. If a vendor provides a lousy closed driver, there is nothing we can do about it. A lousy open driver -- with equally open specs -- can be turned into a good open driver. |
| bigg Feb 08, 2007 7:49 AM EDT |
Good point. I wonder if all the problems with Vista drivers have been limited to proprietary drivers. I'd be surprised if any open drivers have had the problems that companies like nvidia have had. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!