Hashing and down sampling cannot possibly be covered for similar reason. A down sampled movie is still a movie, an hash cannot possibly be used to obtain anything of the copyrighted content.

You would need to make the argument that the hash of the movie was somehow relevant to the original content of what made the movie that movie.

It’s not a question of being the same thing. If you use a single frame from a movie in a painting that’s still a derivative work.

It’s not about format, it’s a question of the source of originality. Hashes are a mechanical transformation without creative input, they use the entire work, and are frankly as close to the definition of derivative work as it gets.

Normally the way around this is a fair use exception, but I doubt that applies here.

I disagree on this definition of derivative. A still frame is still clearly part of the movie, on the other hand if I buy an harry potter book, scan it by hand, print it on paper then burn the print and sell the ashes, that is clearly not a derivative work.

A derivative work need to use the copyrighted work for what it is.

The hash of a movie is no different that the weight of a printed book or its number of pages, it would be absurd for an author to file a DMCA takedown against all books that weight the same amount of milligrams.

> on the other hand if I buy an harry potter book, scan it by hand, print it on paper then burn the print and sell the ashes, that is clearly not a derivative work.

For this analogy, you have to talk not about selling the ashes, but about asserting copyright of your own on those ashes. Your printout of the book would unquestionably be covered by the original author's copyright. The fire would certainly be judged to have made your pile of ashes a transformative use, and any remaining legible letters in your pile of ashes would be in every way de minimus, and you haven't hurt the market for the original book, so you have an extremely safe fair use defense.

Then a more similar case. I make a oil painting of the weight in milligrams of the book (either using or not the ashes in the process).

My point is that the object of the copyright is neither the physical book nor the bitstream in which the movie is encoded, but the book itself and the movie itself.

For example writing a fanfiction of a movie is not transformative because the plot, story, setting, and characters are usually all copyrighted and a fanfiction (generally) use them as-is.

The physical medium is subject to copyright to the extent it can act as a medium for the copyrighted work.

I am not saying that it is legal to sell those first ashes as a product if marketed with the recipe. So I am also not arguing that hosting hashes of movies with the purpose of facilitating piracy is legal (I actually think it is not), but I do not think the reason has anything to do with it being transformative of not.

(I just now realize that ashes/hashes looked like the set up for a pun... it was actually purely coincidental)

Fanfiction of Calvin and Hobs with Calvin the 45 year old parent of an 8 year old girl could easily be transformative without reusing any plot elements. However, if 10 people are all going to end up with the same Hash from the same bit stream you can’t argue it’s a creative endeavor. It’s also not a tiny excerpt as changing even a single bit of the work out result in a different hash. Thus the size of the hash is irrelevant, it’s still copying from the full work.

The only possibility I see is to argue the hash as a simple fact without copyright protection, But, the contents of a book is also a piece of factual information yet copyright is not irrelevant.

PS: As much as you want to use burning as an analogy, a burned book can’t be used to verify a download of the original work where popcorn time’s hashes can.

The hash matches an infinite subset of infinite works. The same thing can glibly be said of any digital representation but the difference is that an algorithm exists to turn the number represented by the flac file back into sound that human beings can enjoy. None such exists for a hash it can never be turned back into the original nor even a piece of same.

Copyright isn't about mathematical derivatives its about protecting the licensed distributor as the sole source of the work. A single screen of a movie or a single page at least is part of the work. Even the first letter of the first word of a book represents a larger portion of the work. It's one nothingth of the work because you can't even derive that first letter from it. Judges are probably substantially interested in the purpose of the law. In math not so much.

> used to verify a download of the original work

I have nothing against arguing that hosting hashes is illegal. It is not that far from facilitating piracy. What I am quite sure is that the hash itself is not covered by copyright.

If you can prove that I am hosting a file whose hash is equal to the hash of a mp4 file that encodes your movie then you can (probabilistically) prove that I am hosting that same file. This is also exactly how bittorrent authentication work if I remember correctly. (my only defense would be to exhibit a non-your-movie- file that still has the same hash)

But this would be no different (or at least comparable) than offering an easy to use list of download links. It would still have nothing to do with whether those links/hashes are derivative or not.

A hash won't violate copyright based on two tests:

1. Transformative use. The uses of e.g. an MD5 hash of a movie file are not similar in any way to the uses of the movie itself, however represented.

2. Market substitution. The availability of hashes cannot possibly damage the commercial market for the movie. Nobody who wanted the movie is going to content themselves with the hashcode instead.

So hashes are always going to be in the clear. To the extent that Popcorn Time violates copyright by providing lists of hashes, that is because of a consideration of the Popcorn Time ecosystem generally, not because the hashes are violations in themselves. They aren't.

Use of a Hash by a torrenting site to enable people to download the movie is clearly damaging the market for the original work.

Clearly some uses of a Hash qualify for fair use, but it’s hard to argue that Popcorn Time is one of them.

I don't think we're disagreeing on that point. All I'm saying is that the nature of the hash as being derived from the movie file is irrelevant. If people organized movie torrenting by assigning version 4 UUIDs to movie files, then Popcorn Time would infringe just as much by publishing those UUIDs as it does by publishing hashes. There's no "derivative work" argument being made on either side. If there was, the pro-Popcorn-Time side would win.

Is a list of movie titles and durations also copyright infringement? They require the entire work to determine, obtaining the lengths is a mechanical process with no creative input.

IANAL, but this comes close to the API/phone book/map arguments where you can copyright a specific expression (a specific map, phone book, with a curated selection) but the facts themselves are just that, facts, and not copyrightable.

The hash is not the movie, cannot be transformed into the movie. You can search it on your favourite piracy site to find the movie, but a similar functionality could easily exist for title + length

Purely factional information is not offered copyright protection. However, the specific contents of a book is also a fact yet copyright is still valid. That might seem like a contradiction, but the law is interpreted by humans who make judgement calls.

Use of a hash by a database to index files is likely fair use, but a torrent site aiming to promote copyright infringement is a different argument.

PS: The Harry Potter movies also can’t be transformed back into the books, but they are clearly a derivative work.

The digital files unlike hashes can be transformed into coherent sound and video reproduction of the actors portrayal of the work.

A comedy set in South America where off screen they hear someone say obliviate and they start talking to someone who can’t remember anything is still a derivative work. Fair uses may apply or it might not, but it’s not simply a question of how much of the original is reproduced, but where information comes from.

I don't think you can claim to own a plot and a hash is less than a skit or a plot point. Infinitely less in fact because whereas even a single character on a page is some small fraction of the work a hash is a nothingth of the work. It can not be used to reproduce nothing. It is data about the work.

Someone already came up with the correct analog. Its metadata like the ISBN or number of pages. Both datum are in fact derived from the actual work but nobody argues that the number of pages is a derivative of the work.

You can't really talk about legal implications and pretend that intent doesn't matter at all. It's rarely a big part of the equation but it's never completely absent.

My claim is only that the hashes are not subject to copyright. Not that hosting them with the purpose of facilitating piracy is legal.

My guess is that in this context the hash is claimed to be comparable to a link.

Say I hash your movie and my hash is modulo 16. Do you now own the number that happens to come out? How can you prove that it was your movie that generated the hash and not some other piece of data?

Copyright is simply a question of source. Unlike patents, it’s very possible for two independent people to come up with the exact same sentence and those people to independently own copyrights to the exact same words. Proving independent creation may be difficult, but it’s still possible.

Similarly, proving where something comes from can be difficult, but let’s not pretend people are independently coming up with 100’s of bit hashes.

PS: As to ownership, both the creator of a derivative work and the creator of the original source have specific rights which in theory expire at different times. Fair use exceptions can also supersede those rights.

> Unlike patents, it’s very possible for two independent people to come up with the exact same sentence and those people to independently own copyrights to the exact same words. Proving independent creation may be difficult, but it’s still possible.

In cases where this is possible, it's often not possible to hold a copyright over the phrase for reasons of lack of creativity. The main difficulty isn't proving independent creation, it's proving that a third party didn't also come up with the phrase independently.

> Copyright is simply a question of source. Unlike patents, it’s very possible for two independent people to come up with the exact same sentence and those people to independently own copyrights to the exact same words.

The canonical essay on this matter, I think, is: https://ansuz.sooke.bc.ca/entry/23 .

> Hashes of copywrited works are derivative works

if a hash is a derivative work because it is based on the original work, what hope is there that a copying a function call from an API is not a derivative work?

Maybe someone could make a clean room version of the hash.