libreplanet-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: FSF continuously harms Free Hardware


From: Jacob Hrbek
Subject: Re: FSF continuously harms Free Hardware
Date: Thu, 20 Jan 2022 18:17:36 +0000

> These three are equivalent because, in all three, we are equally helplsss.  In theory, in case 3 reverse engineering would be able to fix it.  But we can't do any reverse engineering -- we can encourage people to do such it. -- RMS

I agree that complicated components such as processors can be problematic as they are very often very difficult to review, change and fabricate but I don't align with it being a reason to consider Free Hardware Designs ("FHD") as not worthy of supporting, because as long as the designs are GPLv3-compatible then they can be changed to use more appropriate components.

In terms of processors RISC-V would be my quick answer for the definitive user-freedom as they are open-source and can be fabricated using a modified 3D printer or custom CNC machine (majority of CNC machines are GPLv3 compatible and easily changable for this task), but the process is complicated and i am not yet an expert on the subject to represent it.

That said I don't feel like the use of processors is a freedom issue in general thanks to projects such as Rockchip (Arduino, etc..) that make the proprietary ARM architecture malware-free to be considered safe while providing the required components to use the processor (firmware, bootloader, documentation, linux patches, toolkit, etc..) under GPLv3-compatible licenses [https://github.com/rockchip-linux] which then can be declared as safe and through an independent security review with thanks to the provided tools the results can be verified.

---

But again the main issue about which i want to make a case here is that FSF supports and endorses proprietary hardware designs at the expense of Free Hardware Designs and their developers + that GPLv3 is not made with hardware and is problematic for that use.

---

About Creality the required informations are not public (to my knowledge), but this is from what i was told, researched and found relevant:

Creality Ender-3 and Original Prusa i3 are both forks of Prusa Mendel (https://reprap.org/wiki/Prusa_Mendel) which is a fork of RepRap Mendel (https://reprap.org/wiki/Mendel).

Allegedly this video https://www.youtube.com/watch?v=tyVM3-v84I0 inspired Creality to take the Prusa Mendel design and re-engineer it for production to create Creality Ender-3 which was forked by Josef Průša to create the Original Prusa i3 (or prusa was first and creality forked it and optimized it for production cost).

So all of these printers and their forks were forked from RepRap Mendel so they have to follow GPLv3, but they all fail to credit all the authors.

So assuming that RepRap is the copyright holder, I found this on the subject:

I was provided a response of a RepRap representative on their forum explaining most of the issues with GPLv3 and problems with the enforcement of FHD: https://reprap.org/forum/read.php?33,40874

In short my summary of the discussion would be that GPLv3 is not worded and made for hardware as it lacks the required protections to enforce and maintain the values of four freedoms for hardware.

There is also a statement by OSHW/OSWR on GPL that i find relevant, but note that OSHW/OSWR are seemingly not aligned with 4 freedoms: https://ohwr.org/project/cernohl/wikis/faq#q-why-not-use-existing-licences-such-as-gpl-and-any-in-the-family-of-creative-commons-licences

And relevant KiCAD (FOSS EDA) forum post: https://forum.kicad.info/t/using-the-l-gpl-as-an-open-source-hardware-license/1925/2

On 1/19/22 05:16, Richard Stallman wrote:
You're right that the two are similar.  But there is a crucial
difference.  We can get around the problems at the level above the
processor level by writing software.  We can't deal with the problems
inside the processor that way.

Suppose a processor has malicious functionalities.  There are three
ways it is likely to be implementd:

1. By unchangeable circuits.

2. By firmware in ROM.

3. By secret firmware in RAM.

These three are equivalent because, in all three, we are equally
helplsss.  In theory, in case 3 reverse engineering would be able to
fix it.  But we can't do any reverse engineering -- we can encourage
people to do such it.

Thus, we treat all three cases the same.

--
Jacob Hrbek

Attachment: publickey - kreyren@rixotstudio.cz - 1677db82.asc
Description: application/pgp-keys

Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]