Open-Source Hardware Makers Unite To Start Certifying Products (infoworld.com)

An anonymous reader quotes InfoWorld on the new certifications from the Open Source Hardware Association: The goal of certification is to clearly identify open-source hardware separate from the mish-mash of other hardware products. The certification allows hardware designs to be replicated. For certification, OSHWA requires hardware creators to publish a bill-of-materials list, software, schematics, design files, and other documents required to make derivative products. Those requirements could apply to circuit boards, 3D printed cases, electronics, processors, and any other hardware that meets OSHWA's definition of open-source hardware...OSHWA will host a directory for all certified products, something that doesn't exist today because the community is so fragmented.
After signing a legally-binding agreement, hardware makers are allowed to use the Open Hardware mark, which one of their board members believes will help foster a stronger sense of community among hardware makers. "People want to be associated with open source."

Comments

Can't Enforce Copyright on Hardware

[Bruce Perens]

Open Hardware is a good thing to make, but we need to be aware of its limitations. Regardless of the license used, creating a device from the plans is not copyright infringement.

Let's make sure everyone understands that. You can manufacture an open hardware design, regardless of the license, and share nothing, and it is not a crime.

It is, however, potentially a copyright infringement if you publish the plans in violation of the license.

This is because of this text in copyright law. This is the US version but there are similar things in many nations.

17 CFR 102(b) (b) In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.

The reason for this is that functional things such as hardware are protected by patent rather than copyright.

We should also consider what would happen if Open Hardware licenses could be enforced using copyright. Suddenly, any published schematic in a book or online publication would be protected using copyright and the copyright enforced on hardware manufacturers, including all of those in books that exist today. Which would have a major chilling effect on the Open Hardware industry and hardware production in general. We do not want this to happen.

Thus, in general we should not use copyright-based licenses on hardware, lest the courts begin to consider this to be normal practice and create case law that supports it. Courts and legislators do this, it's how we got software patents and other nightmares of today. Let's not encourage them.

Re:Can't Enforce Copyright on Hardware

[Bruce Perens]

You appear to be conflating the term "hardware" with "hardware design", which weakens your argument and makes it hard to discern what your argument actually is.

I'm sorry. In general I work with lawyers and other specialists on this stuff and I guess I missed the level that a nonspecialist can understand.

Right now, all schematic designs and other designs of functional things, including schematic designs, mechanical engineering designs, shapes of 3D objects, typefaces, and a long list of others I won't get into, are not protected by copyright. So, we can use them all freely except to the extent that they are protected by patent, and patents have terms of 21 years (15 for design patents). Copyright has a term much longer than a human lifetime. So, if copyright applies to designs of functional things, that means it's effectively forever and is much worse than the mess we have with patents today.

The problem is that if a large number of people use copyright-based licenses on functional things and behave as if they work, we might convince courts that they do work, and then we lose a large swath of freedoms that we have now.

Yes, ethically you feel you should do what the licensor wants. However, consider that if the licensor acts as if you are restricted by law when you actually are not, and the licensor knows it, the licensor is deceiving you, and we could say that's fraud or unethical behavior.

Re: Can't Enforce Copyright on Hardware

[Bruce Perens]

The answer to this is that the functional parts of your software are not copyrightable, but the expressive parts are. So you can not copyright an algorithm or an API (complexity in Oracle v. Google ignored), or a data structure, or the names and definitions of constants and variables, or a function name and its arguments and return, or anything dictated by the need to inter-operate or physical law, but you can copyright your particular way of writing code where there is more than one choice of how you write it.

if you want to understand this, start by reading Judge Walker's finding in CAI v. Altai

Can't Enforce Copyright on 3D-printed Objects

[Bruce Perens]

Use of copyright on intermediate copies is effective for using copyright to restrict the use of software. The end product of software is the execution of the software in the CPU, and is not possible to separate the execution from the intermediate copying which precedes it.

So, some legal theorists suggested using intermediate copying to restrict the production of 3D-printed objects using copyright, even though the objects themselves can not be protected with copyright under 17 USC 102(b) and similar law.

Use of copyright on intermediate copies is not, however, effective for restriction of copying of two-dimensional or three dimensional shapes, because the end product is a rendering or physical object which can be measured independently of the program that created it, and embodies all of the attributes of the shape. In the case of fonts, one need only render the font in a license-compliant manner, and then trace the outline of the resulting glyph into another program. This is a well-established way of bringing typefaces into Open Source from proprietary fonts, without copyright infringement.

In the case of 3D objects, any means of fitting a mesh or other geometric representation to the created shape would provide a means to bring that shape into another program in a manner that does not infringe on the copyright of the program or data which is used to create the shape. This would include various methods of scanning, optical ones or even exotic things such as CT and MRI. It would also be possible to record the physical movement of the printer or the light beam in producing the object, and map that back to a shape.

So, we have well-established precedent and I'd feel very comfortable testifying about 3D objects (and of course typefaces) not being capable of protection using copyright, in a relevant case.

Design patents, on the other hand, would work fine.

Re:Yet another lie

[AmiMoJo]

You can't poke bits into RAM by hand either, but it's not really important.

As someone who makes open source hardware I won't bother with this though. Slap a CC licence on the hardware, GPL/BSD on the firmware.

Re:Trust No One

[Tough Love]

This sounds like a good early step towards a more technologically free future.

Momentum has been building in the open source hardware space since way back. Opencores.org started in 1999 and now has a library of cores, some of which are in commercial use. That project is now forked as librecores.org, paralleling the Openoffice/Libreoffice split, to move project control into the hands of community contributors. Several initiatives are aimed at freeing up the FPGA toochain, including this toolchain project and this FPGA development board for Raspberry PI. Low cost ASIC manufacturing is available through educational institutions and commercial prototyping services are well within the reach of crowd-funded projects. Though it ramps up more slowly than the now-dominant open source software sector due to the higher cost base and more firmly entrenched proprietary barriers, it now seems clear that open source hardware is set to be the Next Big Thing.

I have no idea whether OSHWA is an important part of this landscape at the moment, but it's hard to see how this initiative could hurt. For the time being, the FOSSi Foundation appears considerably more substantial.

Re:CPU without management engine

[Bruce Perens]

The design of the board may be Open Source - I've not seen the license, etc., but the chips are very definitely not open. In general people who claim to make Open Hardware do use closed chips. Currently there is a RiscV chip which could claim to be Open but it doesn't have sufficient facilities like on-board program memory on-chip for most practical embedded implementations yet, and not the MMU, etc., required for desktops. I'm sure it will get there.