Not entirely correct, its slightly more complicated than that.

Say you write an short story. And you file a copyright for it with the copyright office.

Now, an anthology editor reads your story and wishes to include it in a collection of short stories that he is preparing. It contains, say 20 other short stories as well. This anthology is a "collective work". The anthology editor did not have copyright over the individual stories - he probably only has certain royalty arrangement with each individual author. But the editor can certainly copyright the layout of the anthology. Maybe he put together a brief intro for each story, maybe he ordered them in some artistic fashion based on progression of ideas in the stories etc. This would be a "collective work" copyright. Under a collective work copyright, the editor could sue another editor for producing a substantially duplicate anthology for copyright infringement. But if another magazine published just one of the stories by itself, the editor could not claim copyright infringement.

So in the case of oracle, oracle could have copyright the design of each API individally, and then also copyrighted the entire collection of APIs, thus copyrighting the "collective work" representing the entire language API. But they didn't. They only copyrighted the entire collection as a single work. Thus they (supposedly) cannot claim copyright infrigement when only a small part of it might have been copied by google. (I am not quite certain here).

What oracle is claiming instead - is that they copyrighted Java as a "collective work". which would allow them to claim copyright over the design of the entire API design, rather than just the source code of API implementation. They claim that this design is their creative input. But in this case, then they would also have needed to copyright individual APIs if they wanted to sue for infringement in those individual APIs. Otherwise they could only sue if someone someone copied the entire (or a very substantial part of) the design. They did not copyright those individual pieces.

Back to the analogy:

According to groklaw, Oracle actually copyrighted the entire anthology as an original single work. Oracle claims it copyrighted the anthology as a "collective work" - this is under question. Also, if it DID actually copyright the anthology as a collective work, it seems they did not file copyrights for the individual stories. As a collective work, thus, they can sue for infringement if another anthology is published with substantially the same layout, copying the creativity that went into putting the anthology together. But, if someone copied just one of the stories, they might not be able to sue for infringement as that would not duplicate the creative work of the anthology itself.

On the other hand, if they claim to have copyright the entire anthology as a single work, then if someone copies just the titles of each story and publishes another anthology with different content but same titles for each story, it might not count as infringement.

At least thats my understanding.

"Thus they (supposedly) cannot claim copyright infrigement when only a small part of it might have been copied by google. (I am not quite certain here)."

My understanding is that Oracle cannot sue Google over the small portions without either having registered copyrights for the individual APIs, or proving that those APIs originate with Oracle and were not contributed under license. Oracle apparently didn't do the former, and has not yet done the latter.

I'm probably misunderstanding this, but wouldn't this open the door to allowing someone to copy just portions of a copyrighted work if the work is only copyrighted as a "whole"? Would this require a company (or individual) to copyright every sub system in a larger system? I'm certain I am overlooking something as that seems an unreasonable burden.

The issue here is that Oracle can't assert copyright on the individual pieces because they didn't write them. They were contributed by others. Oracles only copyright is on the work as a whole.

To that end, you could go ahead and copy pieces of the copyrighted work, but you would still be subject to lawsuits by the people that wrote the original individual piece. You just wouldn't be subject to lawsuits by Oracle. It's still infringement, but Oracle isn't the infringed party.

I was unaware that the APIs were contributed by others. Do you have any idea who might be some of the (theoretical) copyright holders of the individual APIs?

There is significant code in the Java libraries that was first written by people outside of Sun/Oracle.

Back in the Java5 days some APIs provided by Java were simply javax.something wrapped around org.apache code.

But that would seem to imply that, for example, the RIAA can't sue someone for downloading half a song (or all but the last 5 seconds).

A single song is not a "collective work".

If anyone is tempted to start spinning the English meaning of the term "collected" to point out how in some sense a song is often a collection of a lot of other elements that may be further copyrighted, stop. The English meaning is not the relevant one, the legal meaning is. A single song is not, legally, a collected work, it is a single work.

If anyone is tempted to construct some exceptional song-like thing that is also a collected work, yes, very good, you've got your thesis topic for your copyright class all laid out, but it is an exception, not the rule.

Curious, would this case be different if the song was a Mashup? I suspect so, right?

A mashup would be a derivative work of several original works, not a collection of works.

Also, for the record, that wasn't the question I was trying to forestall with my last paragraph. I was more after someone doing something bizarre like creating a multi-hour "song" where each "instrument" is a full other song loaded as one big sample or something. Generally speaking the answer to this sort of trickery is "You're a programmer, and you're using thought processes designed to slip things past computers. Judges are human." There are certainly ways to "slip things past" a human judge as well, but they are very different. Human judges generally react poorly to outright sophistry, where computers don't mind it one bit.

This whole thing sounds like the quantum mechanics of copyright since it doesn't fit with common sense.

If copying individual elements (all the chapter titles) of a work is fine, how is it that it is unacceptable to make an original song that contains a 5 second sample of a single instrument from another song? Both are copying a very small excerpt from a larger work.

A single song is an individual work. If you wanted to make a "Best of 80's" CD that would be a collected work. If someone copies a single song from your best-of CD, well, that's not really your problem because you don't own that song. If someone copies your whole CD, then you've got a case if you registered that particular collection.

The list of songs on CD isn't copyrightable -- it's a fact, not a creative expression. Taking a 5 second sample is most likely to be considered fair use.

I don't think any of this is too far off from common sense.

Quantum mechanics is at its most mysterious when you haven't even taken the time to learn the basic terminology.

From my point of view, your question is odd because you've made up both parts of the if clause from whole cloth. Copying "individual elements" may or may not be fine, and five second samples are just plain irrelevant to the discussion being had.

Again, "collective work" is an important term here. It is not clear to me that you understand what that means, and generally trying to answer questions based on deep misconceptions gets nowhere fast. If you really care you're going to have to go pick up some basics from not-an-HN-comment.

I'm not sure how you got the idea that I was trying to argue about what copyright LAW says. I was simply making the case that, just as with quantum mechanics, nothing about copyright is based on common sense, so people need to be careful to be aware of this bias when they think about these issues, especially when a lot of people discussing this have moral / philosophical positions that confuse their view of the legal standing of the plaintiffs. "That doesn't seem right" is a question of legislation policy, not of interpreting existing law.

No, it makes more sense than you think. It's not total sense, but it makes more sense than you think. It's just that since you don't seem to know what the terms mean you're getting a greater sense of confusion than is justified, plus technical people just can't resist the lure of trying to create bizarre exceptions which don't help.

What constitutes a single work is basically what a person on the street would answer is a single thing. What constitutes a collective work is basically what a person on the street would call a collective work, after they answer the first question. (That is, if you start here they might say an album is one work, but if you get them to agree that a song is one thing once they probably won't say that.) What constitutes copying is basically what a person on the street would agree is copying. That certainly isn't good enough advice to go off of if you're going to try to build a business off of it or something, and edge cases abound, but the core concepts aren't anywhere near as screwed up as you might think. The problem isn't conceptual with copyright, it is how the concepts are used and abused.

Its not about pieces of a single copyrighted work its about numerous pieces of separately copyrighted works that can be part of an overarching copyrighted "collective" an anthology of short stories is a good example. Each short story's copyright can remain the copyright of the individual authors, but the anthology as a whole is the copyright of whoever put the anthology together. In this case Java is the collective work, the APIs are the short stories.

To the point below assuming a mashup is made by licensing samples from other works (avoiding fair use issues for a second) the Mashup would be a copyrightable work. But the authors of the individual samples still retain their own copyright.

I see, thank you.

Thank you - that makes sense. I must have glossed over the part that the 37 API's in question were not copyrighted by Oracle. Thanks!

Actually, 37 packages are in question. See https://twitter.com/#!/CalebGarling/status/19559638298225869...

> you cannot sue for infringement of a copyright work unless it is registered as a federal copyright with the Copyright Office.

I'm having a hard time understanding what it means to have automatic copyright when a work is created... but you can't sue for infringement. What kind of implicit protection does your work have if you can't take such action?

The issue is strictly a procedural one. For example, you get the copyright as soon as you author a work and put it in a tangible form of expression (let's say it is source code for a product you have developed). Then, you go about your business for several years using your copyrighted work internally and making sure it is subject to trade secret protections as well. If, after that time, someone lifts the code illegally and uses it to create a competitive product, you have a claim for copyright infringement, among other things. But you cannot file that claim in federal court unless and until you have registered the copyrighted work as a prerequisite to maintaining the suit. Many U.S. jurisdictions say you cannot even sue at all on the copyright claims until the copyright has finally been registered. In California, the Ninth Circuit holds that you can file the federal suit once you have applied for the copyright registration, even though it is not complete.

Therefore, once you create a qualifying work, you have a fully perfected copyright immediately. You just can't sue on that interest for infringement until you have taken the additional step of registering the copyright. You can, however, register at any time prior to filing an infringement action, even if that is many years after the copyright was created. For enforcement purposes, then, it is basically a 2-step process.

Oracle's problem is that it is at the tail end of a trial and has no ability at this point to do anything about registration. If Google is right concerning the motion, Oracle will be out of options on its copyright claims.

On the last point, if the claim is dismissed due to the lack of registration, can't they just register and re-file the claim? Afaik dismissals of that sort are without prejudice. Is the concern the expense/delay that having to file over again would cause?

I believe that while you cannot sue for monetary damages for infringement of a copyright work unless it is registered, you can sue to halt the infringement and have any infringing work seized/destroyed etc. Otherwise the first part where you say "copyright protection applies automatically as a work is created..." would have no real meaning.

Do you have to register it with the copyright office as soon as you create it, or could you create it, register it 10 years later, then immediately sue someone who copied it?

US law generally doesn't allow actions to be made illegal after the fact (ex post facto). So no, the act must be illegal when it was committed.

In this case, though, copyright infringement is already in principle illegal whether a work has been registered or not; the registration requirement is just a procedural requirement that bars collecting damages, not something that makes the action legal or illegal.

However U.S. copyright law does have a statutory limit on how retroactive registrations can be. If you file a registration within 90 days of first publication, the registration is retroactive to the date of publication; but if you file more than 90 days from publication, registration is only effective from the date of registration.

Fwiw, even with unregistered works you have some options available to you, just not damages. For example, you can file a DMCA takedown notice.

If you file a DMCA notice and the infringing party files a DMCA counter-notice, would you have to register the copyright before you could sue for infringement?

I believe the answer is yes, you need to first register before filing suit. If the offending work continues to be published, you can register and then sue.

An exception is foreign works, whose owners can file a suit without first registering, since treaty obligations don't allow the U.S. to subject foreign holders to "formalities". However foreign holders are limited to recovering "actual damages" in such suits, and need to register to be eligible for per-copy statutory damages.

You're supposed to register within 5 years, but if you put it off it's possible that an infringing party could register before you. I can't think of any cases where this actually happened, though.

not quite.

In general, copyright registration is a legal formality intended to make a public record of the basic facts of a particular copy­right. However, registration is not a condition of copyright protection. Even though registration is not a requirement for protection, the copyright law provides several inducements or advantages to encourage copyright owners to make regis­tration. Among these advantages are the following:

• Registration establishes a public record of the copyright claim.

• Before an infringement suit may be filed in court, regis­tration is necessary for works of U.S. origin.

• If made before or within five years of publication, regis­tration will establish prima facie evidence in court of the validity of the copyright and of the facts stated in the certificate.

• If registration is made within three months after publica­tion of the work or prior to an infringement of the work, statutory damages and attorney’s fees will be available to the copyright owner in court actions. Otherwise, only an award of actual damages and profits is available to the copyright owner.

• Registration allows the owner of the copyright to record the registration with the U.S. Customs Service for pro­tection against the importation of infringing copies. For additional information, go to the U.S. Customs and Border Protection website at www.cbp.gov/.

Registration may be made at any time within the life of the copyright. Unlike the law before 1978, when a work has been registered in unpublished form, it is not necessary to make another registration when the work becomes published, although the copyright owner may register the published edition, if desired.

Wait, so correct me if I'm misunderstanding here, but these two things:

  >However, registration is not a condition of copyright protection.

  >Before an infringement suit may be filed in court, regis­tration is necessary for works of U.S. origin.

..seem mutually exclusive. If you have copyright protection without registering, but can't sue without registering, what's the point of the pre-registration protection?

You can register after infringement but before suing. It is sort of like saying murder is illegal, but a suspect has to be arrested before he can be tried.