On Tue, Feb 24, 2004 at 04:08:45PM +0000, MJ Ray wrote: > On 2004-02-24 15:08:20 +0000 Sven Luther <[EMAIL PROTECTED]> > wrote: > >Err, my anecdotes are first hand experience. So, i think it is rather > >more than vague sentiments. > > Please publish the comparative analysis of the magic non-free effect, > if you have one.
You are stuborn, are you not ? Please read the mail archive of this list, i have often stated my experience with the ocaml package there. But then, if you cannot be bothered to read it, i think your opinion on this is not worth reading too. > >The question is, once we decide it is non-free, and we have to remove > >it > >from main, where will it go, and if we decide to keep it in main > >despite > >its non-freeness, then what about other non-free documentation we are > >going to kick from the non-free archive ? Especially the ocaml-docs i > >care about ? > > This is the "who would keep non-free.org?" question. I have nothing > new to write on that yet. And where will we put the threshold of what goes into main, and what will not. I can assure you that if we drop non-free, the freeness of the main stuff will disminish. > >>Yes, these sort of NDAs are always a problem. The problems are at > >>least as > >>bad as the symptoms. Why accomodate the symptoms? > >Because i want to run kick-ass 3D on my debian box ? Because i want to > >use the cheap ADSL pci modem i have in my box ? > > Removing non-free from the archive doesn't make these any more or less > possible than they were before. Maybe it makes it a tiny bit harder, > but Debian is centred around a free operating system, not on making > all half-related user tasks easy. I don't have enough money to order > as much beer online with my Debian box as I'd like... can we make that > easier? ;-) How would being able to run nvidia graphic card in 3D mode be in any way comparable to ordering beer ? The Operating System is the piece of software actually responsible to speak to the low level hardware stuff, and allow you to use it. Supporting a graphic card does indeed fall into this catgoery. And despite all your words, you have no solution to this, no chance to get a free alternative out, so where does this let you ? And then, please go and look at our user mailing list, and the number of questions asked by user using third party graphic driver, either for nvidia and ati hardware, and you will see what it cost us to drop non-free. And even when Branden writes a small reply that he won't support the proprietary nvidia driver on debian-x, it is so much less time he has to improve the xfree86 packages. > >Yeah, sure, but it is only a minority now, i tell you i see a trend > >for > >more and more non-free drivers. What will be our position to those. > >And > >even now, it is only a minority because the majority of non-free > >drivers > >are not packaged. > > I think it makes it even more important that we are clear and > unambiguous in the message: "non-free is not part of the Debian > operating system". But forgetting what we told in section 5, about supporting our users which need non-free. > >>If you support removal when there is a free alternative, it sounds > >>like a > >>weaker form of the Suffield drop GR would get your support. How > >That is what i have been telling since the begining. We have to see > >non-free on a case by case basis, not on a shot first, ask later as > >the non-free removal people suggest. > > Again, you make unjustified allegations about people who disagree with > you. Please stop. Which unjustified allegations ? You are telling us to remove non-free, and that a hypothetical non-free.org will take its place, and fullfill our needs. Is that not what i said ? > The cut must come eventually, else we will be held hostage by a small > minority. If we try to cut too soon, the vote will tell. The problem is that i believe it is not a 'cut' that should come in one go, it should never, it will be an ever shifting 'cut' moment, which will be different for every individual piece of software. > >But for that, we don't really need a GR or changing anything, it can > >even be done with aj's proposal. > > I agree. > > >>to word it, though? I can't think of a way that avoids the "when > >>alternatives exist" condition problem I mentioned in other emails, > >>though. > >There is no need of a new wording. Just point out the case where > >alternative exists, and by all mean, let's remove them. > > You and I have no way to remove them. Sure, we have. We can fill a bug report against them asking for their removal, where you make the case that it should be removed because so-and-so is a viable alternative. You have to convince the maintainer though, and thus handle a reasonable technical discussion with him, which i believe most of the remove non-free proponent are not yet able or willing to do. > >>So, there is still some extra work for each package in non-free and > >>not > >>just "freebie". The OP exaggerated. > >Well, wrong, there is extra work for _new_ packages in non-free. > > Surely each upload is checked a bit? Nope, only new uploads, the rest is automated, or the whole system would fold. > >Yeah, sure, please go ahead, and reverse engineer the nvidia hardware, > > I work on as much as I can, but you are better skilled to do that > sooner than I can. I have absolutely no hardware reverse enginneering skill, and lake the hardware for doing it. > >Yeah. But this third party discouragement is one of the strenght of > >debian, and its seal of quality too. If you want to go beyond that, > >then > >you start having loads of problems. > > This third-party discouragement limits debian unnecessarily, in > development, adoption and collaboration. We will not scale infinitely > and we must address that. Yeah, sure, but encouraging non-free third parties is not the way to go. The way to go could be for achieving a, let's say a debian-approved label or something such, with maintainer of third party stuff who provide as high quality stuff as what we provide. But encouraging them to go the non-free route will be counter productive, and will even be less productive in the fact that they will not make any more effort to have it in debian, and in the long run we lose. Also, i doubt they will build their stuff for the more exotic arches, let alone for anything beyond x86. > >>Does the proposal do that any better than the Suffield drop GR not > >>passing? > >I think not. It is just a keep the status quo thingy. But it would at > >least guarantee a certain time delay before we have to discuss about > >this again. > > How does it do that? I looked, but didn't spot anything. Because i hope you would be ashamed of yourself if you were going to propose this same thing again a week or so after the vote is held, and i hope also you would then fail to raise enough seconds. Friendly, Sven Luther
