Russ Allbery <[email protected]> writes: > Simon Josefsson <[email protected]> writes: >> Russ Allbery <[email protected]> writes: > >>> I believe that's because you are limiting the definition of "firmware" >>> to define most of the firmware in the system to not be firmware, in >>> ways that I find indefensible, so that you can make this claim. > >> Ok, now this become a bit more clear to me. It seems there are two >> things at play here: > >> 1) Hardware-bound firmware shipped with the physical hardware. > >> 2) Supplied firmware provided to the hardware by the operating system. > >> If I understand you correctly, you believe these two are the same thing, >> and you are using the term "firmware" to refer to both and, further, >> that you believe it is not possible to separate these two kind of >> firmwares. > > Correct,
Okay good! > because both methods of shipping the firmware have exactly the same > effect on the user's freedom, which is the thing that matters here. I disagree with that. > The point of the free software movement, and presumably also the free > hardware movement, is the user's freedom to control their own computing. > This is the lens through which one therefore must analyze this in a > discussion in a free software context. I don't think that is universally true, and the FSF, myself and others disagree with that. >> If so, I think we found some fundamental aspect to disagree on. > > Well, it took years and multiple threads with you to get the point across, > but better late than never, I guess? My perception is that earlier discussions more followed the "you are wrong" followed by "no, YOU are wrong" pattern. At least now it should be possible to acknowledge that there are two possible opinions on this, tracing it down to different concepts of the term "firmware" where I believe it is important to separate between two forms, and you believe it is indefensible to separate between them. >> To me this is comparable to saying "owning music on a CD" or "streaming >> music from a service" is the same, > > I have no idea why you would say this. In the typical situation with both > firmware cases, the user is in physical possession of the hardware and the > firmware, but cannot modify the firmware or exercise the other standard > free software freedoms. This is precisely the problem with non-free > software in general. In one case, the firmware is shipped separately, but > this makes no more practical difference than whether the computer came in > one or two boxes. There is no "streaming" involved in the general firmware > problem. (There are probably some devices with some sort of call-home or > remote-disable capabilities, but that's an entirely different problem than > what we're talking about here.) Interesting -- I see 'apt-get --update dist-upgrade' as the streaming I was thinking of. I believe the comparison works well when considering the legal rights for a user. When buying a piece of hardware, at least where I live, I have some consumer laws and QA process that grant me some rights and protection. That includes the entire product I bought, including the pre-loaded firmware that is embedded into the machine. When I install some streamed non-free firmware, I do that under a software license that usually give up all the consumer protection I had before. Sometimes it even further restrict what I can use the device for (inspecting the non-free firmware, for example). It may also restrict the field of use, or which geographic territories I can use the software in. Read the license of some of the stuff in non-free-firmware. Which may change tomorrow. To me all that seems similar to how different the rights for a user is after buying a CD with music on it, and when you stream it from Spotify. This view is also consistent with why hardware manufactures like OS-supplied non-free firmware: it is cheaper to produce. Which is the same for Spotify vs CD. >> It doesn't seem reasonable to blame the lack of adequate hardware on >> software people. That ought to be the responsibility of hardware >> people, and that's where help is needed. > > I'm not blaming anyone for anything. (Well, I'm blaming the hardware > companies for using non-free firmware, but I don't think we disagree about > that.) > > I am pointing out that removing a user's ability to update their non-free > firmware does not in any way give that user more freedom. It does the > exact opposite. I don't think anyone has suggested to remove that ability. I also disagree: the example above illustrate how the user gets less freedom with OS-supplied firmware than with pre-loaded firmware. > The goal of the free software movement should be to provide that user with > the full source code to all of their firmware and enable them to change it > in any way that they want. I don't think the free software movement has the responsibility or the solution for that. Some free hardware community is needed for this. > We cannot currently do that for basically all commerically available > hardware, which is sad. The best that we can do is offer them the > ability to manage the non-free software their system is crippled with, > so at least they can exercise what control they have to choose whether > and when to change the versions of that loaded non-free firmware. This > is not true software freedom, but it's *more* freedom than freezing > the firmware at whatever copies are currently loaded into the > hardware. This should be obvious: the user has a few choices instead > of no choices. My disagreement here follows from our established different opinion. I believe the user has less freedom, since both Debian and the user is now subjugated by the non-free software licenses that come with the OS-supplied non-free firmware. That wasn't the case before. > This of course also provides some concrete practical advantages to the > user, such as being able to install security updates from the non-free > software vendor if they choose to do so. It also means that we archive the > historical updates independent of the vendor, so users can still obtain > what was previously available for their devices even if the hardware > manufacturer has disappeared or has pulled older versions, perhaps because > they allow the user more control and freedom than the manufacturer had > wanted. That those updates are also non-free sucks, but the role Debian is > serving here is to provide the user with as many choices as we can, given > the regrettable situation. Sure -- and users of Debian Deblob can make use of all those non-free firmware, and have those choices. It just doesn't have to be pre-loaded on the installer images, but an opt-in decision by the user. > The approach that you are describing as somehow more free removes the > support to the user for exercising those choices in return for... well, > making you personally feel better? I'm not trying to be flippant here; I > really can't see any benefit to the user whatsoever in your approach other > than that it somehow satisfies some bizarre ideological purity test that > doesn't practically help the user in any way. The user has full choice to opt-in to install non-free stuff on their machine, so I'm not following this point. Nobody has suggested to remove the ability for the user to install the firmware. The advantage is for the producer of the Debian images and the user to not be subjugated by the moving target of complying to the non-free license agreements. >> Hardware freedom concerns has to be resolved by the hardware community, >> it cannot be solved software people. > > Right. So in the meantime, as we all try to navigate the imperfect world > we are currently living in, we should provide our users with the absolute > maximum amount of freedom that we can. Because our priority is the freedom > of our users. Agreed. That's why it is important to retain an ability for users to install Debian without needing any non-DFSG parts. /Simon
signature.asc
Description: PGP signature
