On Fri, 28 Feb 2003, Jon Smirl wrote: >> Does DRI have a future with neither NVIDIA nor ATI >> participating? > >I really don't understand ATI's position on Linux >drivers. They have better hardware but they are losing >because of their drivers. I can't think of a better >solution than having a couple hundred highly skilled, >performance obsessed, unpaid hackers fixing their code >for them.
This is not an argument against such a possible source code release, but I'm merely trying to be realistic here. I don't see 100 unpaid hackers hacking feverishly on anything X related right now. Why would 100 unpaid hackers come out of the woodwork all of a sudden? Quite unrealistic. What would really happen, would be the existing DRI team members, and others closely working on DRI _right_now_ would look at it likely, and probably work on it, and possibly a small number of new people too. I doubt there would be even 5 new people touching the code if that. >ATI and Nvidia have both disassembled each other's >drivers so there are no secrets between the two. Is there evidence to support this claim, or is it merely conjecture? >Another argument could be that a another chip manufacturer could >clone a chip and use the drivers for free. I don't see that >happening given the complexity of the chips. Another argument is that the binary drivers of both companies most likely contain 3rd party intellectual property, and patented techniques, some of which they themselves may have patented, and others which they have possibly licensed from other companies, perhaps many companies. Would all of those companies want to have their IP instantly open sourced and free for all to see/use? While I would love nothing more than to see that, I highly doubt that either ATI or Nvidia have the legal rights to open source the entire source code of all of their drivers. Parts of them perhaps, but I doubt all of them, and even then parts of them would be patent encumbered still. >Finally ATI could be afraid of patent issues by opening the >source. But this would just make the patent issues slightly >easier to find, they'd still have the same problems with closed >source. If they were being accused the lawyers would get the >source anyway. Not ATI - any vendor. You assume perhaps that there are perhaps patents being violated and these companies want to hide something, and by opening their code they could get sued. A more likely hypothesis is that they have licensed patented technology themselves and have the right to use it in their products, drivers, etc. but do not have the right to give it to other people or distribute the source code. That is not at all unrealistic to assume. There is also the threat that the code might contain something unknowingly is patented, such as a technique totally invented on one's own, that just coincidentally was invented by someone else first and patented. Opening up source code can cause other companies to jump in and see which of their patents you might have violated so they can sue you. Again, this isn't an argument against opening of such code, but rather an attempt to explain some real life reasons why some companies do not do so. >What about bad fixes to the code? ATI can just control the CVS >and only apply patches that they are happy with. ATI should >continue with their paid developers, just make their changes >public too. I'm sure if there weren't other reasons already that this would not be a problem. >Linux programmers like to fix things when they are broken. I >just removed the ATI radeon drivers from my system and went back >to the DRI ones. About once a day the ATI driver will lock up. >If I had the source I would have poked around and tried to fix >it. Without the source I threw the drivers in the trash. I agree with you, and many people on this list would want to twiddle with the source as well. We're the minority however. Most people do not want to hack on source code of software that doesn't work for them, even fewer on device drivers, and even fewer on video drivers. >ATI is really underestimating the skills of some of the hackers >in the Linux community. Some really good code comes out of >those long Siberian/Finnish winters. They are also missing the >opportunity to lead in a fast growing market. Just look of the >cover of Business Week. I completely disagree here. For the above paragraph to be true, ATI would have to be sitting there in Markham thinking "Well, we would release the source code of our binary drivers, but the open source community isn't talented enough, so we wont." Not only do I think that is not the case of what they think, I think it is not even in the equation. I have no idea what ATI, Nvidia, Kyro, or any other video hardware vendor's explicitly detailed reasons are for not releasing the complete source code to their drivers is, however I do understand enough to be able to piece together several legal and other reasons why they would. None of those reasons result in me thinking any of these vendors doubts the open source community is capable of hacking on source code. I would bet money, that the number 1, 2, and 3 reasons are entirely legal reasons surrounding intellectual property and other issues. While I totally am on the side of the fence that would love to see source code released of such drivers from _all_ vendors, I don't expect them to fully release the code of their drivers as I understand the legal and some other aspects. At least ATI contributes both source code and specifications to the open source community, and Nvidia contributes largely to the 2D only "nv" driver even though they do not providing hardware specs. Over time perhaps more and more will be released from different vendors. Time will tell I guess. One thing is certain though, and that is that no amount of community begging, petitioning, and other similar requests are going to get hardware companies to violate intellectual property laws just to oil a squeaky wheel. Instead, lets thank them for what they do give us, and politely request things of them that are realistically something they might be willing to, and legally able to provide us in the future, rather than a nonrealistic white card, which they're not likely to be legally able to do. Just my thoughts. -- Mike A. Harris ftp://people.redhat.com/mharris OS Systems Engineer - XFree86 maintainer - Red Hat ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel