Conversation

this screen makes the remaining life escape my body

when i was a teenager i used to have the iwlwifi ucode for the card in my laptop always stashed onto a good thumb drive.. anyone else?

5
0
1

@domi I don’t like that nonfree software is included on the installation media at all.

I am a smart person and I make sure I buy hardware that works with free software, so I don’t need it.

2
0
0

and of course it didn’t work and I had to copy the file manually to /lib/firmware… i would complain but they did provide ABSOLUTELY NO WARRANTY so i guess that’s fair

0
0
0

@domi I didn't, but that was one of my main reasons for avoiding debian. I believe a year or two ago they finally decided to make the firmware-including image the default one.

1
0
1

@ar i’m only installing it because i’m in Driver Hell

1
0
0

@domi to fix this i have to install non-free firmware using a netinstall ISO

0
0
0

@domi Driving to Hel? In this weather? SMH ablobcatwink

0
0
1

@domi realtek network cards are notoriously bad on linux

1
0
0

@Starcross no they aren’t

they’re cheap garbage on all platforms, it’s not any worse on linux

0
0
0
@SuperDicq @domi isn't this screen a thing literally because the non-free software *isn't* included?

like, if non-free software was included on the installation media, the screen would read: non-free firmware is required to use your network card. are you fine with non-free firmware getting installed?
1
0
1

@mei @domi This is an old version. Nowadays they include it.

1
0
0
@SuperDicq @domi ah, I see. not sure why you dislike that though. like, wouldn't the option of using Debian on the existing hardware of a user mean that they can easily get most of the benefits without having to jump through excessive hoops, therefore making it more likely that they'll later buy hardware that doesn't require non-free firmware?

I guess this is the sort of thing where one's stance probably depends if using free software is a pragmatic choice, or more of a religious thing
2
0
1

@mei @domi It has nothing to do about religion.

People should be required to jump through a few hoops to install proprietary software or they will have no incentive to start using 100% free software.

1
0
0

@SuperDicq @domi @mei it’s not religion! proceeds to say it’s religion

2
0
1

@Lili @mei hold on, i have a good test

hey @SuperDicq - what do you think of Atheros cards that don’t require firmware to be loaded? Remember, the card still runs non-free code, the only difference is that you don’t see it

1
0
1

@domi @Lili @mei It’s misleading to say ā€œonly difference is that you don’t see itā€.

The actual important difference here is that you conveniently forgot to mention is that the firmware is not upgradable and read-only.

It’s not ideal obviously. I would prefer to have a fully upgradable card with free software. But at least these Atheros cards are more fair, because there is no injustice between user and developer (neither of them can change the software anymore, it’s baked into the hardware)

2
0
0

@Lili @domi @mei I don’t understand how you read religion from my post. I did not refer to any made up deities or fictional books.

If you want me to talk about religion instead I can talk about the Church of Emacs if you wish saintignucius

0
0
0

@aetios @domi @Lili @mei Een discussie hebben betekent niet gelijk dat je ruzie hebt, wesley

1
0
0
@SuperDicq @domi @Lili @mei ik wilde gewoon een grappig plaatje posten
0
0
0

@SuperDicq @domi @Lili @mei …but isn't it more freedom if you can choose which version you want to load?

1
0
1

@lnl @domi @Lili @mei No, because all the versions that are signed to run on the hardware are currently proprietary.

If someone at some point had made free replacement firmware this statement would’ve been true.

1
0
0

@SuperDicq @domi @Lili @mei obviously it's proprietary in both cases, it'd be best if the firmware was free, etc. My point is, it doesn't become more proprietary if you have to load the firmware yourself?

1
0
1

@lnl @domi @Lili @mei With Intel where the software is signed by the company and you can’t change it but they can. That is an unethical and malicious handcuff done on purpose by Intel.

With those atheros cards, with the firmware being not upgradable it becomes just simply a hardware limitation.

And that I find that infinitely more acceptable than the conscious decision Intel made to handcuff users.

1
0
0

@SuperDicq @domi @Lili @mei this isn't a remote upgrade they're applying to your card, you are not required to run any newer version. You can use the first release forever. There's nothing preventing you from rolling back to an earlier version

1
0
0

@lnl @domi @Lili @mei There is in reality no running the first release forever or upgrading or downgrading versions.

The firmware has to get flashed on every boot. You have to install the proprietary software every single time you boot your computer.

Unlike the atheros card where the firmware is on the card and I never have to install proprietary software.

1
0
0
@SuperDicq @domi @Lili @lnl so, let's consider a concrete situation in which you might want to exercise your freedom: for example, let's say that the firmware of your network card has a bug, and you would like to fix it.

now, let's consider four cases:

1. the firmware is proprietary, in ROM on the network card, you cannot modify it. in this case, you're out of luck – the bug is not getting fixed.
2. the firmware is proprietary, loaded during boot, signed with the key of the vendor. in this case, the bug is only getting fixed if you can convince the vendor to release a firmware update.
3. the firmware is proprietary, loaded during boot, but not authenticated in any way. in this case, you need to either convince the vendor to help you as before, or you can fire up your favorite decompiler and get to work figuring out how the firmware works, and how you'd need to modify it to fix the bug yourself.
4. the firmware is free, and loaded during boot. you can also fix the bug yourself, but with less effort required.

it seems to me that this order goes from the least freedom, to the most freedom. is there a flaw in my logic?
3
0
1

@mei @SuperDicq @Lili @lnl on some devices that don’t allow in-band firmware upgrades, you can still do one through a hardware flasher. much more involved, but possible

then you also have cases where the signature verification has been haxxed or the key leaked. you can make your own firmware, but not thanks to the manufacturer (which puts you in a curious position where you give money to someone that acted in a way opposite to your goals. or something)

1
0
0

@mei @domi @Lili @lnl You’ve ordered these potential situations based on a different question than the one that I am asking.

Your question is a solution oriented one ā€œHow likely can I fix a potential bug?ā€

My question is a philosophical one ā€œWhich situation is the least unethical?ā€

2 and 3 are equally bad from an ethical standpoint, someone else is in control of your computing, because they are able to modify the software on your computer while not giving you the permission to allow you to do the same. That is unethical because they refuse to give you control when they could.

Number 1 is simply an annoying situation, but the fact that nobody can do anything about it makes it not an unjustice. It’s just an unfortunate situation with nobody to blame. It’s simply a hardware limitation, it’s not rights being denied from you on purpose.

0
0
0

@domi @mei @Lili @lnl Yes, if the way to sign your own code got leaked and someone wrote free firmware for one of those network cards it would be acceptable to use that. But so far that has not yet happened.

2
0
0
@SuperDicq @domi @Lili @lnl it feels to me like you are treating free software as a dogma, with the loophole of "it's okay if i'm not installing the firmware myself" being much like the various workarounds the Rabbis have invented for the laws of Judaism that are too onerous to honestly obey
2
0
1

@mei @domi @Lili @lnl That’s not how I see it at all. Firmware on a network card rom chip is not ā€œinstalledā€ by anyone. It is baked on the chip and it is a part of the hardware.

I think that if you focus only on the pragmatic and solution oriented questions, like engineers often do, you might forget to ask important ethical questions.

Philosophy and ethics don’t have much to do with religion.

The free software movement is a social movement because it has a clear real world end-goal (to make all software respect the user’s freedom), unlike religion which is just simply fairytails for adults.

1
0
0
@SuperDicq @domi @mei @lnl It's installed by a chip designer without sharing a source for such chip, then by the robot operator/whoever clicked 'make a chip' in the factory.

By the way, ath9k has updatable EEPROM, which means it's non-free by default since you don't know what this actually does, nobody seems to bother about that even on Librebooted (not to be confused with what Leah Rowe called the later project that includes X230 and others) machines https://github.com/CodeFetch/art-collection
1
0
1

@Lili @domi @mei @lnl I’m not sure what you’re trying to get at here. I am very well aware that they never published the source code.

But releasing that source code isn’t necessary, as I wouldn’t be able to build and flash that code anyway due to hardware limitations.

0
0
0

Inga stands with šŸ‡ŗšŸ‡¦ šŸ‡µšŸ‡ø

@mei @Lili @domi @SuperDicq @lnl
Let's consider four cases:
1. Read-only iphone, absolutely locked down and simply not flasheable on a physical level, users are forced to use the exact OS and apps it came with.
2. Bricked regular iphone which you have to flash with (some) iOS before using it. Maybe some day the firmware verification will be broken (as it was for earlier iphones) and you'll be able to flash something else onto it.
3. Regular android phone, with unlockable bootloader.
4. Pinephone, but there are still some proprietary firmware blobs for the radio.

It's glaringly obvious that 1 gives user the most freedom! /s

0
0
0

@mei @domi @SuperDicq I dislike it for a different reason: instead of providing an image with non-free-firmware alongside the free-only image (which was option 6 on the ballot), the free-only image was replaced with the one including non-free-firmware (option 5). This means that if you don’t need the non-free firmware, you still have to waste bandwidth downloading it and remember to boot with firmware=never. There’s also the fact that non-free firmware being available for your system doesn’t necessarily mean you need it (e.g. if you have AR9462 and don’t use bluetooth, you don’t need to load anything), but I don’t remember if Debian asks you whether to load each firmware blob or just loads them all without asking (if it asks it’s not an issue).

2
0
0

@noisytoot @mei @domi Yes, I agree that free should always be default. And non-free should be a discouraged optional thing for those that need it.

1
0
0
@noisytoot @domi @mei @SuperDicq You have intentionally been confused - there is nothing firm about such proprietary software that takes the users freedom.

Much of such software is peripheral software that happens to run on a auxiliary processors, but that's merely a technical point - and a lot of it runs on the main processor.

The Debian installer detects to see if the hardware could run some proprietary software and then permanently installs it for autoloading without asking, or even listing what was installed and also enables the "non-free-firmware" proprietary software repository.

The result of that proprietary software repository being enabled makes apt do things like go and automatically install proprietary software from nvidia (that includes proprietary malware binaries that run on the main processor) on installing an unrelated package, even if the computer doesn't have a nvidia GPU.

>(if it asks it’s not an issue).
If proprietary software has been included in the recommended installer - no asking has taken place - rather there is an attempt to force the proprietary software onto you.
0
0
0
@SuperDicq @domi @mei @noisytoot "need" is the wrong word - in a lot of cases the proprietary software is not needed and is rather convenient to have.

Nvidia GPUs work fine with the Nouveau driver and without proprietary peripheral software too even for handcuffed GPUs (it seems the nouveau developers have broken such functionality for later versions of Linux?), for GPUs that have been designed to use more than trivial re-clocking, the proprietary nvidia driver conveniently tends to be more performant.
1
0
0

@Suiseiseki @domi @mei @noisytoot It’s more common than you think that someone needs proprietary software thanks to most WiFi cards not working otherwise.

2
0
1

@Suiseiseki @domi @mei @noisytoot Which I mean is understandable if you’re working with hardware bought by someone else or donated to you or anything like that.

But if you buy your own hardware and you didn’t check before buying if your hardware is going to work with free software that is kind of your own fault.

0
0
0
@SuperDicq @domi @mei @noisytoot Such Wi-Fi cards don't work properly even with proprietary software.

You don't strictly need Wi-Fi - you can use 1000BASE-T.
1
0
0

@Suiseiseki @domi @mei @noisytoot Unfortunately most laptops don’t come with ethernet ports anymore and fully rely on internet connectivity using a proprietary WiFi card.

1
0
1
@SuperDicq @domi @mei @noisytoot Yes, inconveniently you need to use a usb-1000BASE-T adapter, or a decent usb-Wi-Fi card.
0
0
1
@SuperDicq @domi @Lili @mei @lnl It's RSA and you need the RSA private key, but intel will never publish that.

Intel did slightly screw up with the RSA signature incompetently, but that just allows for arbitrary code execution only after the proprietary software is loaded - which is not useful for a free replacement.

Intel of course still refuses to provide the needed key for the user to be able to fix security problems intel can't be bothered to fix - meaning the attackers win and the users lose.
0
0
0