Hacker News new | past | comments | ask | show | jobs | submit login
EU court: Reverse engineering OK, API not protected by (c) and more (wildeboer.net)
116 points by jwildeboer on May 2, 2012 | hide | past | favorite | 40 comments



Assuming the US court does end up ruling (even partially) in Oracle's favour, this would serve to make the U.S. an even less attractive location to start a startup than it already is due to current and proposed violations of basic human and civil rights by the government in order to protect content copyright holders (i.e. the interests represented by the MPAA and RIAA)[Citing Kim Dotcom/Megaupload, SOPA/PIPA/CISPA]. I postulate that this might result in a perfect storm of conditions that result in the shifting of the centre of gravity of the startup, and eventually wider entrepreneurial/financial worlds off American shores for the first time in more than a century, perhaps to Europe, due to a more conducive legislative environment, or, more intriguingly, to India or perhaps Brazil (assuming they get their governments in order). The advantages of the Silicon Valley ecosystem are still massive, but they would, in such circumstances, erode over a (not necessarily very long) period of time.

Edit: The page appears to be down, so I've posted the full text on pastebin - http://pastebin.com/6G5vqLj5


The opposite could happen too. If you want to make sure your APIs are copyrighted then you would stay in the US.

For example it would make it harder for someone to offer a drop in replacement for your cloud based service by replicating your APIs exactly.

Personally I doubt APIs could be copyrighted, but some would find an advantage if they could.


Right, but I suspect those who would find advantage would be the incumbent market leaders in their respective fields. The truly disruptive startups, the innovators, would be weighted down considerably by such legislation, since they might unknowingly replicate an obvious part of a copyrighted API. The situation would quickly spiral out of control. If that were the case, the copyright-ability of APIs might result in the stagnation of software development itself, in the absolute worst case, though I suspect, as I have previously stated, that people would get around that by relocating to more friendly legislative environments.


I think this is the bigger risk. Copyrights are even more numerous than patents and are potentially much more difficult to avoid infringing. You could infringe an API in a totally different industry and never know it, and there are so many of them that checking first would be basically pointless.

At least with patents you can do some reasonable due-diligence (e.g. when building an audio codec, check for audio processing patents). You're not in the clear just by doing so (obviously, look at all the lawsuits) but it's at least some protection. API copyrights could end up being much more deadly IP landmines.


Unknowingly violating copyright is like unknowingly shoplifting; there are cases where you will have incorrect knowledge about licenses or somesuch (or you'll put a piece of fruit in your bag and not pay for it), but it's an extreme edge case. The problem with patents is that if you and someone else arrive at the same solution and the other person patents it before you put your solution out there, you are in violation of their parent. There is no such analogue for copyright. You and I can write functionally identical code for a decade and never once violate each other's copyright.


Unlike patents, you can't accidentally infringe Copyright. You can have 3 photo's of a buss that look vary similar, but only 1 was infringing because it tried to recreate another picture.


I see the point that it is difficult to infringe a copyright purely by accident.

I may have over- and misstated the problem as it exists at the moment. I am still concerned that the Oracle case will push us toward copyright applying to an idea rather than an expression, as discussed elsewhere in the thread.

I am bothered by the fact that an API, by its nature, to some extent, is just an idea. You call some function with parameters of certain types in a certain order and you get back some piece of information computed in a particular way from the parameters.

So the function prototypes themselves (the "API") are really just descriptions of an idea the way "a red bus driving across London bridge" is a description of an idea for a photograph. They are more terse, non-English (or whatever) descriptions, but descriptions nonetheless.

In this case it would be quite a bit easier to accidentally infringe a copyright because copyright would end up intersecting with patents.

For example, a photo sharing service might have an API endpoint called "photo" that accepts some image and meta data through a POST request. You are starting a competing service. You are aware that their API is copyrighted so you call your endpoint "picture". But the problem now is that you've done the same thing the photographer did with the picture of the bus in London, you've intentionally wiggled around a copyright to do something very similar.

Could you be sued? Maybe, seems that you could be if you happen to be in England, but I'm not a lawyer. The problem is that this makes a lawsuit seem much more likely to me than if the API itself weren't covered by copyright at all.

And we definitely don't need more IP lawsuits.


Sure you can - see http://en.wikipedia.org/wiki/George_Harrison#All_Things_Must... for example - George Harrison lost a copyright infringement case because he was deemed to have subconsciously plagiarized, no less. Care to claim that when you design a new API, it isn't influenced by APIs you're familiar with?


Referring to [1]? That's already a huge step worse than copyright, and just starts on the path to "Well look, I published an API to upload files with a progress callback. And you clearly wanted to do the same thing and even named it Upload, but you just moved the parameter orders around."

1: http://www.techdirt.com/articles/20120126/10515817552/uk-cou...


What I think your missing is that picture was as common as a the coke logo. If they where a random coffee shop somewhere the 'I have never seen that photo' defense would have worked just fine. Patent trolls work because they can lie in wait for people to 'infringe' but unpopular API's that few people use don't have the landmine effect. If you never saw them they don't exist.


That wasn't quite my point. Copyright should cover the expression only. Right now, I could go write my own book that's clearly inspired by Star Wars or Harry Potter. This sets a precedent that even creating a _similar_ book, photo, or API is infringement. This directly goes against the idea of copyright: to generate more works for society.


The case is quite unique in some ways because the alleged infringer disclosed that they had intentionally instructed a designer to design around the original.

Copyright, at least in the UK, has for a long time granted protection for copies which are not necessarily identical, but similar where they are a substantial reproduction of the original.

Admittedly in the case above the line between protecting an idea and the expression was pushed towards the idea, however, as stated, the specific facts of the case (in the sense that it was a blatant attempt to mimic but not infringe the original) made it more likely that the judge would find infringement.


They should have used the law on 'passing off' rather than copyright.

http://en.wikipedia.org/wiki/Passing_off


Why do you say that? They were successful in the instant case and in my experience, passing off actions are far harder to prove than simple copyright infringement actions. There was no misrepresentation here designed to divert custom, just a good old fashioned desire to use someone else's image.


Because the case was more suited to the law of passing off.

How much experience do you have of passing off cases versus simple copyright infringement? Because to me this looks like a bad decision, as whatever the intent the same image was not used, just one that was similar in content and execution.

And it certainly looks like misrepresentation to divert custom.


I have experience of acting for clients on both sides of passing off and copyright infringement claims.

There was clearly no passing off here in the way you tend to frame a passing off action as per Warnink (in the UK at least). What misrepresentation are you referring to?

The defendant was a tea company looking to use an image, the claimant was a gift company so it wasn't a case of a customer going to the tea company when they meant to go to the gift company. This is the misrepresentation which generally has to be in place for a passing off action to be successful.

Copyright infringement does not just cover identical images but substantial reproductions. The correct question is whether the copy was a substantial reproduction or whether it simply borrowed elements from the original, that is to say whether the line between inspiration and copying had been crossed.

\edit I would also add that in general passing off actions are far harder to prove than copyright infringement. To my mind, if a copyright infringement action is available this would always be preferable to bringing a passing off action.


Add this to the fact that the U.S. legislature is considering making copyrights potentially infinite(edit: infinitely extendable) in length, and you have a recipe for total disaster.


Do you have a source for that?


They can't make copyrights infinite. But the last several decades have shown that they can extend copyright terms indefinitely. And, thanks to Disney's lobbying, will do so every time that it looks like Steamboat Willie is going to become public domain. (That's the first film that Mickey Mouse was in.)


I do not think it will work. If, just for example, Amazon decide to copyright their API and keep their business in USA for copyright protection, another company, who decides to replicate Amazon's API fully will just do it in Europe. Customers probably will not care, since it is internet. Only possible downside here - latency between USA and Europe.


Although it should be borne in mind that the reciprocal protection of copyright works conferred by the Berne Convention should still technically apply meaning that, to use your example, Amazon could still take action against infringement of their API in whatever country was also a signatory to the Berne Convention.


Nice to know, as I'm currently reverse engineering API calls for a few organisations in the UK who don't supply their data in easily parsable formats. It's not for anything malicious; I'm actually trying to drive traffic to their sites and help them gain more visitors. Try explaining to non-technical people the benefits of an open API and why they should pay for one :\



Since the blog is down here http://curia.europa.eu/juris/document/document.jsf?text=&... is the judgment and some commentary: http://www.bloomberg.com/news/2012-05-02/copyright-can-t-blo...

most important bit:

Consequently, the answer to Questions 1 to 5 is that Article 1(2) of Directive 91/250 must be interpreted as meaning that neither the functionality of a computer program nor the programming language and the format of data files used in a computer program in order to exploit certain of its functions constitute a form of expression of that program and, as such, are not protected by copyright in computer programs for the purposes of that directive.


That's interesting because it goes even further than I'd have expected: that file formats themselves aren't protected. I guess that isn't surprising, since you'd want a market where programs could all compete to operate on the same files, but I don't remember that being talked about before (at least in reference to Oracle vs Google). And that does go to the format of the Java .class files.

I wonder what the EU courts would have to say about an encrypted file format...


The key thing to remember is that the case in front of Judge Alsup in the US Northern District of California is a jury case. It'll be up to people who have no idea how computers work to decide this. Having just served on a jury, I see the value, but in this sort of instance, I am frightened by the possible result.


No, the jury will not be deciding on whether or not APIs are copywriteable. Judge Alsup is deciding on that, if necessary. The jury will be deciding on Google's fair use arguments etc.


Thanks for the correction - though, quite honestly, having met Judge Alsup, I'm not sure this is any better. He hates technology.


So this is interesting. What happens if national (eg. polish) law specifically disallows reverse engineering? Which law takes precedence?


ECJ follows EU directives, which take precedence over national law. When individual countries refuse compliance (it happens quite often), an "infraction procedure" is open, which can potentially result in heavy fines for the country.

More often than not, the country could pass a law that complies with the directive in question, then "forget" to enforce it as soon as the original process is terminated in a satisfactory way; individuals would then appeal to the ECJ again and the process would be repeated, this time resulting almost certainly in fines. And so on ad infinitum, or (usually) until the political winds change at the national level.

In almost every case, EU laws and directives take precedence over national law, but the mechanisms to enforce this primacy are still somewhat vague and weak. This is by design: if they had been too clearly stated, most countries would not have signed the treaties. The "plan" was to state the overall principles, then find a way to enact them further down the line. It's a very pragmatic approach, and it's producing results in many areas, but clearly it's not something that, today, will satisfy all individuals in their lifetimes.


The rule is simple. EU directives must be implemented in national laws of all EU member states. If Poland fails to implement the directive by such an action a EU violation case would start ultimately leading to consequences.

OTOH this is a tested way to get stuff in laws that are hard to find national support for. Make it an EU directive and explain to your people that "Sorry, but even if we don't like it, we must implement this now." - see EU data retention directive etc.


What if the law in question has existed before Poland joined the EU?


Part of the process of joining the EU is to make your legislation compatible with the EU legislation.


National governments in the EU have a duty to implement EU Directives. If a government implemented a law inconsistent with a directive, the EU would take enforcement action for failure to implement the directive properly.


Since this is Union Law, and according to the treaties, a party in Poland could escalate such a copyright case to the ECJ, which could then overrule any Polish higher court. The thing is that the ECJ requires (or advices) Poland courts to comply to its decisions, but cannot actually enforce them, which is totally up to the country. So if Poland decided to go on its own on this, theoretically it could since the EU states are sovereign, but in practical terms, and this is a really fuzzy divide, Poland would not do it if it wants to remain a politically "healthy" member of the EU.


I am trying hard to keep the site up, but you guys and gals are hammering the sh*t out of my little vServer ;-)


Throw CloudFlare in front of it.


What are the ramifications, if Oracle win in the USA but isn't protected by EU legislation?


I'm not fully conversant with international copyright treaties, but the Berne Convention to which both the US and various EU countries are signatories to, provides that a work that is protected by copyright in one signatory country will be protected in another country, so the ramifications in terms of creating a divide between the US and non-US countries may not be overly significant.


Oracle (and Sun) knew this - that's why these are patents.

The law allowing reverse engineering of APIs and interconnects is well established in the US and Europe and has been for decades. In the US it goes back to Amdhal making plug compatible peripherals for IBMs and the right to reverse engineer is enshrined in an EU directive.

That's why the Oracle case isn't about copyright. They claim to have patents on various concepts of the language. So you are perfectly at liberty to implement a 'while' command with the same function template - you just aren't allowed to use it to do a loop !




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: