Hardware Isn't Generally Copyrightable (2012)

Versus: No one is going to magically re-create peter-pan line for line.

I think the reason for this is that copyright and patent law exists to promote the advancement of the arts and sciences. When copyright covers a work it also covers smaller portions of it. With machines there's often only a few efficient ways of making it work. If those were covered by copyright then everyone else would have to invent new ways of doing the same thing. You can get a patent on a novel machine though, but patents last for much less time than copyright does.

It's a bit like copyrighting an API and then everyone else has to rename their math.min and math.max functions.

Don't worry, it is!

Schematics are copyrighted as well.

But someone can take your schematic, lay out their own board, and produce a exact equivalent of your circuit, using all the same components.

I think one of the confusions is that we're so used to source code being copyrightable but that's a relatively recent invention when software came into mass use [3]. My opinion is that were it not for lobbying efforts to consider source code as artistic work, source code would probably be considered a "device" and not subject to copyright, just like hardware.

I don't want to put interpret Stallman's view incorrectly but I believe that this is the crux of his argument. If I interpret his view correctly, he believes software is more function than art, especially as it relates to fundamental infrastructure, and that copyright shouldn't be applied to such works or least should be restricted in the scope of application.

Also note that clothing is not copyrightable. This is why you see so many (legal) knock off clothes. The brand, or trademark, is something that's legally protected but not the designs themselves. I think the fashion industry is lobbying to try and change this and have their work protected by copyright [4].

IANAL

[1] https://en.wikipedia.org/wiki/Copyright

[2] https://www.bitlaw.com/copyright/unprotected.html#ideas

[3] http://digital-law-online.info/lpdi1.0/treatise17.html

[4] https://www.washingtonpost.com/blogs/arts-post/post/fashion-...

There is no monolithic "intellectual property", it's an catch-all, umbrella term that refers to a large body of independent laws on copyright, patent, trademark, and trade secrets, which are historically established under different contexts. A work can be uncopyrigtable, but still restricted from reproduction due to a patent or trade secret status. It's why the term "intellectual property" should be avoided.

> Why couldn't the actual design of a PCB be copyrighted, for instance? I mean, one could argue that there's as much intellectual property being created when you route a PCB than when you write a book.

The bottom line is, just because something takes intellectual efforts to create doesn't imply it must have copyright status.

Copyright is an artificial monopoly constructed by the governments. Usually, the scope of copyright laws is limited to the expression of creative and artistic works, and furthermore, not the ideas or systems themselves, but merely their expressions. Many types of "functional" works are explicitly not covered under the copyright law. Also, it's also important to remember that the ultimate purpose of copyright law is to promote public interest. Thus it's entirely reasonable that the government deliberately decides not to put something under the copyright laws if the public interest takes priority. It's equally possible that the government suddenly wants to put something under the copyright laws because it considers the public interest is better served by granting copyright restrictions to the authors.

A controversial example is the French copyright law. The lawmakers in France apparently decided that the public interest is better served by putting the nightview physical appearance of the Eiffel Tower under the copyright law. In France, if you take a photo of Eiffel Tower at night and post it to the Internet, it's possible that your photo infringes the copyright of the nightview design of the Eiffel Tower. [0]

A better example under the U.S. copyright laws is rasterized fonts in bitmaps and pixmaps, they are not copyrightable, because the fonts are "functional" - it's more of an apparatus/machine for printing texts than a work of art. In the past 200 years, the copyright was never updated to include fonts, likely due to the concern that the lawmakers consider that the rights to the public to use the printing press outweighs the interest of the font authors. However, a digital font, a.k.a the computer program that generates vectorized fonts, is copyrightable as a computer program, but not the pixels it generates. [1]

Another example: a few weeks ago there was a news story about how a volunteer recreated and 3D-printed a ventilator valve and got threatened by the vendor. But if the case is ever brought to a court of law (it's not), the defendant's lawyers may argue that a single physical valve mainly serves as a functional part of the machinery, and by itself, it does not meet the threshold of being a copyrightable work by itself. On the other hand, the original CAD drawing of such a valve is protected under copyright. However, it doesn't stop someone from legally reverse engineering the valve and independently reproducing its CAD drawing, even if the CAD drawing is identical to the original file, you can argue in court that an identical drawing is the only way to express the idea of the valve. Unless the valve has a patent. Unlike copyright, a patent restrict the ideas themselves, not only its expression.

And finally we come back to electronics design. Most aspects of hardware design are not copyrightable, due to the same reason that the layout of integrated circuit wasn't really copyrightable until the 1980s. Back the 1970s, it was possible to produce a chip with the same layout from your competitor. The reason is similar to the font program vs font output: Although the pattern of the layout itself is a piece of artwork thus protected, however, the photomask it produced was generally not copyrightable because it's only an industrial apparatus for producing the chip. Thus, although the artwork layout in itself cannot be copies and redistributed without authorization, but it was legal to produce an identical mask for the chip, thus, it was legal to produce a chip with the same layout straights from your competitor. The only way to prevent others from making an identical chip was patenting the design processes or technology in the chip, however, many chips only use general-purpose technology and process, so it cannot be patented. The lawmakers decided that the U.S. public interest will be better served by allowing the masks to be copyrightable, so the copyright law was modified. [3]

As we see, although many aspects of hardware design can be copyrighted, like a CAD drawing or its physical appearance (as industrial design), however, a lot of other aspects of hardware design was, and still is uncopyrightable due to the same line of reasoning. Ultimately, stopping one from making a functionally identical machinery isn't inside the scope of copyright laws, but patent laws.

Another conclusion is that, while it's good to explicitly release a hardware design under a FOSS license, a FOSS license applied to hardware design is pretty powerless than when it's applied to software. Especially, copylefting hardware itself (You replicated our machines, release the source too) is somewhat impossible.

[0] https://en.wikipedia.org/wiki/Eiffel_Tower#Illumination_copy...

[1] https://en.wikipedia.org/wiki/Eltra_Corp._v._Ringer

[2] https://en.wikipedia.org/wiki/Integrated_circuit_layout_desi...

Here's an overview of how it affects the auto industry.[1]

[1] https://digitalcommons.lmu.edu/cgi/viewcontent.cgi?&article=...

This seems relatively straightforward and while I understand the law is properly understood by understanding a large body of law and legal opinions on the interactions of same I see no particular reason to disbelieve it.

In particular nobody is obliged to accept an obviously absurd interpretation of the law because it would seem to follow from an exact reading of the law. It's not math nor a computer program interpreted in the minds of judges. Case in point the above argument on the "schematics" of a book.

This doesn't work because judges aren't robots and would laugh it out of a court room. The same way we should laugh new innovative interpretations of copyright out of the world stage.

Lets ask a deeper question. Why WOULD it be a good thing if copyright applied to machines? Copyright for all intents and purposes will probably never be allowed to lapse in reality. As it stands the people that end up owning everything else would end up being the people that own the ability to move forward under each and eventually every industry ad infinitum as they bought up basic rights in a given industry not for a finite amount of time but forever.

In short we are already in what must be nearly the most ridiculous time line possible. Lets not make it worse.

The process for transforming a notional chip mask into one that can be used to actually produce chips on a fabrication device that uses a particular wavelength is going to be covered by multiple patents, unless the chip fab is more than 20 years old, so you can't copy the post-transformation masks, either.

The whole design of the hardware is only patentable, and only to the extent that the patent claims apply to it.

So if you take an entire electronics device, reduce it down to a block/circuit diagram, then use that to create a new schematic, and use the schematic to produce another electronic device with identical function, you have not violated copyright in doing so, nor are the schematics and masks derivative works.

In the case of the book that describes a process of manufacturing, the process part of it is like a recipe. It is only the supplemental prose that is copyrighted. So you would have to strip out the introduction, descriptions of specific applications, etc.

That is why map-makers used to put fake towns or streets on maps, beyond the borders of the area that is purportedly mapped. So on a map of Pennsylvania, they might place a town called "Fakesburg" in the greyed-out part of Delaware that adjoins Pennsylvania (a literal grey area), and put a "Trapsville" in the greyed-out New York, the Pennsylvania portion of the paper being presented as factual, and the margins as copyrightable creativity. But the courts called this out as bullshit, so the mapmakers stopped. Some elements of copyright still exist in mapping, but only to the extent that the map diverges from literal facts. For instance, if the widths of roads are exaggerated, and colors added, for the purposes of making an infographic for a traffic throughput data set, the creative combination of facts could possibly be copyrightable, but little prevents someone else from independently combining the same two sets of facts into an all-new (but similar-looking) infographic.

In the case of hardware, the lowest-level circuit diagram is the factual map of the device. But there are many ways to lay it out in physical reality, as to how many layers to use, where to route the traces, which subunits to place in the blocks, how to label the components with printed text, etc. So you couldn't make a photographically identical copy, but you could make a copy that does exactly the same thing with a redesigned PCB.

Mathematical formulas are considered unpatentable and uncopyrightable and legally, programs have been recognized to be equivalent to mathematical formulas.

So only the formatting, variable name, decoration of code is the ground for copyright. My understanding is that if you were to run the code through an obfuscator, the result would be free of copyrightable items.

To my despair I never found anyone playful enough to help me explore how solid this reasoning was.

Even assuming that insurance companies act as naive rationalists believe they do, losses are socialized. In actual fact, insurance companies are able to price discriminate against policy holders likely to cost more than the average cost in the future, especially if likelihood of future losses are as public as court proceedings.

A classic example of this is CentOS and RHEL. You can fork RHEL all you want, but you cannot put "Red Hat" on your fork.

Free in hw is kind of hard, meaning, who pays for physically manufacturing?

Don't worry, the FOSS hardware scene is alive and well, although not as advanced as software yet.

> Does anyone know what happened?

It appears that the openhardware.org project has been discontinued in favor of other projects of the same nature, such as oshwa.org.

> (ii) 'layout-design (topography)' means the three-dimensional disposition, however expressed, of the elements, at least one of which is an active element, and of some or all of the interconnections of an integrated circuit, or such a three-dimensional disposition prepared for an integrated circuit intended for manufacture ...

So if you just express a schematics, with no actual placement, it is not copyrightable.

"Partnerships with large OEMs operate as signals that also generate barriers to imitation. As pointed out by one of our informants: “If a company like Samsung or TI is licensing from us, this means it may be difficult to develop our technology.” Multiple licenses create a psychological barrier to entry. In addition, the IP that is licensed needs to be complemented by ARM’s customer services and considerable expertise"

from p22 of here https://www.ifm.eng.cam.ac.uk/uploads/Research/CTM/working_p...

So I suppose somebody could make an ARM clone but they'd have to actually do it.

An actual licensee would surely be prohibited by the license though.