Comment by mindslight | Hacker News Reader

Comment by mindslight | Hacker News Reader

If someone is going to an install fest, their technical skills are not very strong. Especially in this day and age where plentiful devices means not rendering your only computer inoperable, but being able to follow along with online instructions! That user's effective freedom is very low, even if they're sitting in front of a fully documented Free machine!

But they're interested in taking a step in the right direction, even though they will not reap the full benefits for quite some time. This must not be squandered!

The problem here is that emphasizing the negative association still ties it to GNU/Linux - first, the presence of a visible "devil", and second the difficulty of being confronted with an overt looked-down-upon decision that is not fully understood.

Rather, people respond better to the carrot. They should be informed - they might indeed have a different machine that would be a better fit for Linux. And they should walk away knowing that the machine they have is not perfect from a Freedom standpoint.

But how are they supposed to care enough to influence a purchasing decision, especially of a new machine, if we make it difficult for them to become accustomed to the benefits of Freedom? A machine loaded with all proprietary blob drivers still showcases the results of Freedom quite well - similar to why there is such big business in taking Free software and locking it up with DRM/HTTP, but I digress.

In general, RMS really needs to reconsider the nuance of how he applies his philosophy to the modern world. While he has been extremely prescient and unyielding for the overall war, he keeps wanting to fight the original battle rather than many of the new fronts that have sprung up.

For example, he uses a T400 laptop - what storage device does it have? I guarantee it runs non-free firmware from flash. There is little difference between this and AMD's graphics firmware blob files, if the latter is only ever loaded under Free control. Condemning the latter because it puts the blob front and center in our faces is fundamentally a symptom of failing to flesh out the paradigm of what applied Freedom means, and without that we cannot hope to defend it!

> what storage device does it have? I guarantee it runs non-free firmware from flash.

RMS is pretty consistently clear that software intended to be upgraded by the end user must be free. Software/firmware that is not intended to ever be changed or upgraded by the end user is more or less the same as circuitry and outside the scope of the Free Software movement.

> There is little difference between this and AMD's graphics firmware blob files.

AMD's blobs are intended to be installed by the end user. Both the firmware that you upload to the card when you use it, or binary drivers that run on your main CPU. Those are both software that you receive and execute as an end user which is fundamentally different from the gates and firmware fused into your device.

Except that firmware isn't "fused" into your device, but simply on an undocumented flash chip - the days of windowed EPROM are long gone. So RMS is compromising here, in my opinion incorrectly.

If we stop talking about a device as being "Free" but rather use the FSF's verbiage of "respects your freedom", then I would actually argue that AMD's firmware blobs are more-RYF than SSD firmware. The former I can be pretty sure AMD did a decent job at the code signing, meaning they're the only party that can use the device to disrespect my freedom. Whereas the SSD firmware interface remains more opaque, so there's a decent chance that anybody could reverse engineer the SSD and surreptitiously modify the firmware on mine to undermine my freedom!

Of course you can insulate yourself from the SSD by using full disk encryption and considering your SATA controller a security boundary. And that's my entire point - we need to evolve these definitions to incorporate this type of nuance.

I don't see a difference between RYF-ness of CPU microcode and SSD firmware. The main practical difference in the context of TFA is that distros typically bundle CPU microcode but not SSD firmware, hence the question is unlikely to come up in that context for SSD firmware, but that doesn't mean that it's not a problem.

Btw, the AMD CPU microcode is fused into the CPU, otherwise it wouldn't boot; there's an additional SRAM area where the microcode update is applied; for details see https://media.ccc.de/v/35c3-9614-inside_the_amd_microcode_ro...

To elaborate on your last point, I'd always use software encryption with SSDs, because with the opaque firmware wear-leveling it's essentially impossible to be sure that anything is actually physically deleted from SSDs, and if it isn't physically deleted it can be read by a custom firmware.

I was referring to AMD's graphics cards. They have a binary firmware blob that basically must be loaded, but the loading itself is done by Free software. I don't see how this is any different from a RYF perspective than if AMD had put another flash chip on the BOM for storing it cold. That is, assuming a competent signature scheme for both.

So the card itself doesn't RYF, but it can be used to display the output from a RYF computer. And unfortunately barring a better graphics option based on open firmware, these are the compromises we have to make. RMS recognizes this - I just think the manner in which he framed the compromise is a bit unnuanced and out of date. Rather than finding reasons to ignore least-worst blobs, we should be talking about boundaries between Free/non-free components.

SSDs are an interesting case, because the hard drive interface abstraction is so simple and longstanding, we just kind of assume it's a good boundary. But if you want to pop back into abstract Freedom land, imagine what the market would look like if vendors weren't able to market around decommoditizing software features. For example, if the FTL were done by Free software (perhaps on the main CPU), there would be no worries about certain lines of drives getting corrupted due to power failures!

(And yes, totally agree about FDE. I actually just changed my router back to being a general purpose Linux box, and it felt quite odd installing that with no FDE).

I'm sorry, no idea how I managed to jump from AMD GPUs to AMD CPUs :-)

> RMS is pretty consistently clear that software intended to be upgraded by the end user must be free. Software/firmware that is not intended to ever be changed or upgraded by the end user is more or less the same as circuitry and outside the scope of the Free Software movement.

That seems somewhat circular in definition on the edges. I can and have updated firmware on hard-drives (always opaque binary blobs from the vendor, but definitely changes, updates, and bug fixes not directly analogous to circuitry).

RMS has a role to play precisely because he is unyielding and refuses to be pragmatic.

Most of us here are pragmatists. We are trying to get things done as best as we can. Therefore we shouldn't follow his advice. But we should take it into consideration. "This is what could be"

That's not really what I'm saying. In so far as I invoked pragmatism, it's only about what's pragmatic to get people to adopt Free software.

Even RMS engages in pragmatism, hence calling "Free" his T400 with non-Free storage!

From https://www.fsf.org/resources/hw/endorsement/criteria

> However, there is an exception for secondary embedded processors. The exception applies to software delivered inside auxiliary and low-level processors and FPGAs, within which software installation is not intended after the user obtains the product. This can include, for instance, microcode inside a processor, firmware built into an I/O device, or the gate pattern of an FPGA. The software in such secondary processors does not count as product software.

> second the difficulty of being confronted with an overt looked-down-upon decision that is not fully understood.

But that's the point. The goal of the Free Software movement is not to get every person running free software, but to educated people on _why_ they should care, and why they should insist on running free software for themselves.

Installing Ubuntu on a bunch of people's computers and saying "see how much better linux is than Windows" misses the point. I think that's the main flaw in these so called "install fests."

The Free Software movement argues from a moral standpoint, not a standpoint of technical superiority. Otherwise you set yourself up for a constant back and forth of "but this thing was easier on windows."

That's not the point.

Reply on news.ycombinator.com