Hacker News Reader

Comment by Hermel | Hacker News Reader

I'm pretty sure that if you rewrote Harry Potter from scratch with new names and wording (redoing the implementation) and published that, you would still get in trouble. So copyright is somewhat broader than you think.

View on HN

Of course there are grey areas, and it's appropriate that they're decided by the courts. I'm sure nobody would argue that a lossy copy of a work protected by copyright can still infringe, even if the lossiness was deliberate and intended to differentiate it from the original to a minimal degree.

Copyright of literary characters is one of the harder ones to define, and perhaps a system more like patents would be more appropriate for them, but that's neither here nor there.

Let's stick to computer science for a while. Imagine I have come up with a sorting algorithm for integers with significantly better time complexity than existing options. I can describe that in any number of programming languages, in natural language, in pictographs, etc. It's a skill that people can learn, or a tool that could be embedded in silicon and used for financial gain. Doesn't that sound more like an invention than a creative work? Shouldn't patents be the best fit?

Don't get me wrong, copyright and patent systems the world over hace significant class, but for us as software engineers to declare that patents shouldn't apply to our work because it is somehow fundamentally different to the other science and engineering disciplines is just arrogance.

The problem is that algorithms and math are explicitly forbidden from being patented. This is why most software patents include some sort of clause specifying that the work is done using a computer.

The patent office has simply done a lousy job when it comes to software patents.

I hadn't heard it explained this way before, but it explains so much.

So, for example, if I came up with a better sorting algorithm, I couldn't patent Luca Sort, but I could patent using a computer to run the Luca Sort algorithm? Or am I misunderstanding?

That's not how it works in Europe; just saying "done on a computer" is explicitly not enough. There's a concept called "technical effect" that's been worked over in the courts and well developed in the last decade (I'll leave you to look it up, eg https://en.m.wikipedia.org/wiki/Software_patents_under_the_E...).

I'm a bit out of the loop but whilst USA started letting pretty much anything in there seem to have tightened up over the years.

There's a parallel with business methods too. In both cases there been a move towards a middleground - harmonisation. Personally I feel the USA position is far worse and this harmonisation is not good; liberalisation of soft-pat and business method patents on Europe would be a big loss IMO.

>So, for example, if I came up with a better sorting algorithm, I couldn't patent Luca Sort, but I could patent using a computer to run the Luca Sort algorithm? Or am I misunderstanding?

That was how the law was until recently. But the supreme court found that "just do it on a computer" no longer counts.

That's the Alice case, which is described here:

https://en.m.wikipedia.org/wiki/Alice_Corp._v._CLS_Bank_Inte...

I don't believe you are correct. You can't patent math, but you can patent algorithms. This is analogous to the material world, where you can't patent physics, but you can patent engineering.

An algorithm isn't math, it's a series of instructions. That is something that is patentable, as long as it involves an inventive step.

A more efficient sorting algorithm that is non-obvious would almost certainly be patentable.

You're partially correct. Algorithms can't be patented. However, in practice, patent drafters work around this by adding phrases such as "A memory with computer readable instructions to [Insert Algorithm]" to turn the algorithm into an invented product. The bottom line is that the invention is the memory and the processor system whose novelty is that it includes instructions which implement the particular steps of the algorithm. This practically covers all implementations of the algorithm on Von Neumann machines, in effect covering all current implementations of the algorithm.

Anything you can do in software I can do with a set of hydraulic value). Your software will be faster (a video game on my valves will take several lifetimes to display the opening), but it will work (until something leaks). thus you can translate your program to hydraulics and it is patentable.

It's gotten harder to do this since Alice.

https://en.m.wikipedia.org/wiki/Alice_Corp._v._CLS_Bank_Inte...

https://en.wikipedia.org/wiki/Gottschalk_v._Benson

https://en.wikipedia.org/wiki/Parker_v._Flook

https://en.wikipedia.org/wiki/Diamond_v._Diehr

https://en.wikipedia.org/wiki/Bilski_v._Kappos

https://www.google.com/patents/US4558302

As you can see, that is a patent for an “apparatus and method”, not an algorithm per se.

As a holder of dozens of patents, many of them on processes derived from algorithms, I fully understand (as well as anyone can) the implications of current case law. In order to patent an algorithm, you have to re-cast it as an operation to be performed on a general purpose computer. As long as you do that, you have the patent. I fully agree that there's nothing to stop you from calculating GIF's with pencil and paper, or on an abacus, or in your head -- is that a relevant distinction, or merely an academic one?

Additionally, the algorithm has to be non-trivial, which I take to mean multi-step, for it to be patentable when it's transformed into an operation performed on a general-purpose computer. Simply multiplying by a constant, even when done on a general purpose computer, is not patentable.

The first sentence on the Wikipedia page of algorithm:

"In mathematics and computer science, an algorithm is a self-contained sequence of actions to be performed. Algorithms can perform calculation, data processing and automated reasoning tasks. "

An algorithm is a mathematical construct that happens to be useful with computers. See the first volume of The Art of Computer Programming for a nice formal definition of algorithm.

Larry Cotter, yer a wizzzard Larry. You're going to Fogsnorts under the greatest principle ever, Bumblebore.

Oh, I don't know. JK Rowling rewrote Charlie and the Chocolate Factory and added wizards and no-one's complained. :p

You must be kidding, Harry Potter is nothing like Charlie and the Chocolate Factory, unless you consider magical candy to be the most important part of the story. :)

Philosopher's Stone has quite a few direct correspondences. The letter/golden ticket, the factory/Hogwarts (I could go on, but it would require me to reread it, not that that would be a bad thing). The unjustly discriminated against kid is another Dahl trope (although not one in CF per se). And more generally, the style is very similar. Beyond the first book it diverges, however.

Copyright considers things a unique work if less than something like 10% is copied. Harry Potter is in no danger. Even changing the golden ticket to a letter is enough to get around infringement there.

Of course consult a lawyer if the 10% figure above it important to your copy - there are lots of details that I don't know.

I only felt the need to establish enough similarity to demonstrate that the joke was funny, not so much the Dahl estate could take Rowling's hard-earned money. :)

https://en.wikipedia.org/wiki/Tanya_Grotter

The literary characters and other proper nouns in a series may be covered by trademark. Copyright only covers the specific expression of the text. Patent, of course, does not apply to fiction plots.

So if you scrupulously avoided character names and quotes from the original work, you can indeed rewrite the same plot and story arcs in your own words, and not be in conflict with the law.

That will be of no solace when lawyers for the Harry Potter media empire pursue you anyway. Just as you did not break the law in a technical, letter-of-the-law sense, so is it also possible for an entity with great wealth to make someone with lesser wealth miserable, without breaking the law.

It is far easier to rip off a nobody, or someone who is dead. For instance, rewrite Amber without Corwin or Merlin in it; make superficial modifications to all named characters and locations. Maybe you get some fans that never read the original; maybe you get some that did read it, and like your version better.

That didn't happen when Disney ripped off Kimba the White Lion and made The Lion King.

It also didn't happen when Terry Brooks reskinned the lord of the rings as "The Sword of Shannara".

Or when EL James rewrote Twilight without the vampires and called it 50 Shades of Grey.

Or when Disney animated Doc Hollywood and called it Cars

Cars was a Pixar movie.

Pixar was already owned by Disney when Cars was released.

Wikipedia[1] says otherwise:

> [Cars was] Pixar's final independently-produced motion picture before its purchase by Disney in May 2006

[1] https://en.wikipedia.org/wiki/Cars_(film)

I was under the impression that the characters in Twilight never actually have sex, while the characters in 50 Shades of Grey have a lot?

Nope, it's not a violation of copyright.

Copyright law apparently makes an explicit exception for translations. (i.e. translations are subject to copyright despite usually sharing literally no wording with the original.)

It isn't clear to me whether that only covers works in a foreign language that claim to be translations of some other work, or whether it would apply to any foreign-language work that duplicates the plot of an earlier copyrighted work.

The thing to remember with copyright is it covers copying. If you read a book and rewrite it in another language you didn't copy it, you copied the ideas, perhaps, but copyright covers expression and not ideas.

Creating a translation of a work is a derivative work itself. This applies to porting software from C++ to Java as well as translating of Harry Potter to French.

Re-writing a story and translation are different things entirely.

If I describe at an abstracted level the working of a C++ program to you and you implement it in Java then you have not copied the work, nor even created a derivative, ergo clean-room implementation being a thing.

Like I said, the copyright law makes an explicit exception for "translations". Copyright also covers translations, despite the fact that they share none of the "expression" of the original.

Nope, it's too wishy-washy for you to judge the potential for violation in a forum post. The 4 criteria have to be interpreted with respect to the situation at hand, and hypotheticals are almost impossible to interpret rationally.

Republishing single sentences has previously been found to be a copyright violation, and copyright has been granted on as few as three notes. On the other hand (to pick a recent example), 50 Shades of Grey, originally written as a Twilight fan-fiction (derivative by definition) was never pursued for copyright violation (AFAIK).

Reply on news.ycombinator.com