Google v. Oracle (Supreme Court Copyright Case)

[TwiceHorn 31440]

This case is coming up for oral argument at SCOTUS in October.  Not sure what the Ginsburg vacancy means for that, and she might have been an insightful justice, given that her daughter Jane is one of the nations's top copyright scholars.

The case is that Google copied aspects of Oracle's Java Application Programming Interfaces for use in Android, rather than licensing them.  The issue then is whether such APIs are entitled to copyright protection and need to be licensed at all.

Computer "source code," that is a program or module written in a computer language, is known to be copyrightable just like any other textual or literary work.  The copyrighted source code is compared to the infringing code for similarity in textual elements just as you might between a Stephen King novel and an alleged ripoff.

An exception to this is when the code is "logarithmic" or "mathematical."  A simplistic and lame but apt demonstration is e=mc^2.  That is surely text, and, if original, subject to copyright.  Except that that text expresses an idea and there's only one or very few ways of expressing that idea, as in the case of formulae or algorithms to solve problems, etc.  Ideas cannot be copyrighted, and when the expression is only one or one of a very few ways of expressing the idea, the idea and the expression are said to "merge" so that the expression is the idea and is not copyrightable.

Google contends that the Java APIs are ideas, rather than expressions, and, even if they are protectable expressions, the doctrine of fair use means they cannot be held liable for infringement.

Fair use is a limitation on copyright protection that historically arises from the First Amendment.  That is, people are free to copy portions of copyrighted works for the purpose of "commenting" on them, which can take various forms, to include parodies (see 2LiveCrew, Roy Orbison).  The idea is that copyright protection must yield to freedom of speech in certain circumstances to encourage the "marketplace of ideas."  Contrary to popular belief, fair use is not a doctrine of "de minimis" or "non-commercial," or "non-profit," or too small to matter infringement. 

The trial court held that the APIs are not copyrightable, and in any event Google's use of them was "fair use."

The court of appeals, which is the patent/IP-focused Federal Circuit, held that the APIs are copyrightable, and fair use does not apply, period, because Google's use of them does not implicate the First Amendment in any way shape or form.  The case has been ongoing for nearly a decade.

More deets at SCOTUSBlog. https://www.scotusblog.com/case-files/cases/google-llc-v-oracle-america-inc/

This is or may be one of the biggest IP cases to hit  the Supreme Court in decades.  Or maybe not.  The issues are interesting and it has attracted a lot of interest, especially among software anarchists, who seem to believe everything digital ought to be free.

Anyone give a shit?

Thoughts, comments, insights?

Edited September 25, 2020 by TwiceHorn

[Enchubben 5510]

fuck the brockermeyers Google. Go Oracle!!

[TwiceHorn 31440]

2 minutes ago, Enchubben said:

fuck the brockermeyers Google. Go Oracle!!

It is a battle of bigtech shitheads, for sure.  I tend to feel that Oracle is a little less shitheaded than Google/Alphabet.  And I'm am pretty sure I favor Oracle's legal position, but could be talked out of it.

[Captainant 27638]

Lol if the supreme court rules in favor of Oracle, the entire computing industry will be destroyed overnight. The concept of an abstraction layer for a service to be an interface to a program is fundamental to modern computing and scalable architectures.

I'm sure the supreme court will get it wrong thanks to the interest Oracle has had in other topics of "national interest".... 

[Viking 301]

Fuck Oracle. Google is in the right.

Some APIs don't even provide an implementation and it's up to 3rd parties to provide one by design. If I write an implementation for such an API am I infringing the copyright? I would love to know what percentage of an API I need to copy for me to infringe the copyright. Is it 1 method? All methods? Somewhere in between?

[TwiceHorn 31440]

2 minutes ago, Viking said:

Fuck Oracle. Google is in the right.

Some APIs don't even provide an implementation and it's up to 3rd parties to provide one by design. If I write an implementation for such an API am I infringing the copyright? I would love to know what percentage of an API I need to copy for me to infringe the copyright. Is it 1 method? All methods? Somewhere in between?

Well we aren't talking about "some APIs," we're talking about specific Java APIs, and not the ones without implementations, which wouldn't seem to involve copying anything.

And how much copying is enough to constitute infringement is not a question raised here.  The standard is "substantial similarity."  To constitute copying, something is pretty much going to have to be a verbatim copy.

It appears that your attitude is emblematic of "software anarchists," who believe they should be able to copy anything with impunity.

If you are willing to provide more concrete examples, that might be helpful.

[Dbeasy 15042]

I think the posters are correct.  An API has only very limited ways to be written.  You have data that must be shared, and in a format.  It's a lot like a formula. There are only a few ways to do it.  I side with Google.

[TwiceHorn 31440]

4 minutes ago, Dbeasy said:

I think the posters are correct.  An API has only very limited ways to be written.  You have data that must be shared, and in a format.  It's a lot like a formula. There are only a few ways to do it.  I side with Google.

I think the counter to that is, there are only a few ways to write an API to use Java, not to write software more generally.

[Captainant 27638]

10 minutes ago, TwiceHorn said:

Well we aren't talking about "some APIs," we're talking about specific Java APIs, and not the ones without implementations, which wouldn't seem to involve copying anything.

And how much copying is enough to constitute infringement is not a question raised here.  The standard is "substantial similarity."  To constitute copying, something is pretty much going to have to be a verbatim copy.

It appears that your attitude is emblematic of "software anarchists," who believe they should be able to copy anything with impunity.

If you are willing to provide more concrete examples, that might be helpful.

Well, and the ruling was that API's as a concept, not just implementation, are copyrightable. The supreme court can remand the appeals court decision on that basis. Because it's basically not just copyrighting your business's forms, but that you use forms at all for data processing

[JBJ 3654]

There's gotta be some technical mumbo that goes above my head in this case.

An API seems more similar to an industrial process than a creative work.  So I don't think this really falls in the realm of copyright.

It should probably have patent protection.  But that's a different discussion.

[TwiceHorn 31440]

34 minutes ago, Captainant said:

Lol if the supreme court rules in favor of Oracle, the entire computing industry will be destroyed overnight. The concept of an abstraction layer for a service to be an interface to a program is fundamental to modern computing and scalable architectures.

I'm sure the supreme court will get it wrong thanks to the interest Oracle has had in other topics of "national interest".... 

Well, the Federal Circuit, which is about as apolitical as a court can be, and has more technical expertise, by a very long shot, than the Supreme Court or any other court in the US, has already ruled against Google.

The one knock against the Federal Circuit is that it is generally pro-IP.

The Supreme Court seems to take extra pleasure in substituting its judgment for that of the Federal Circuit.

[TwiceHorn 31440]

9 minutes ago, Captainant said:

Well, and the ruling was that API's as a concept, not just implementation, are copyrightable. The supreme court can remand the appeals court decision on that basis. Because it's basically not just copyrighting your business's forms, but that you use forms at all for data processing

No, it is not a ruling that APIs as a concept are copyrightable.  It is necessarily taking into account the specific APIs in question.

Also, in delving into this, I'm kind of thinking that the copyrightability issue isn't properly before the Court.  Only the fair use issue.

Edited September 25, 2020 by TwiceHorn

[Rimbo 25913]

1 hour ago, Enchubben said:

fuck the brockermeyers Google. Go Oracle!!

 

1 hour ago, TwiceHorn said:

It is a battle of bigtech shitheads, for sure.  I tend to feel that Oracle is a little less shitheaded than Google/Alphabet.  And I'm am pretty sure I favor Oracle's legal position, but could be talked out of it.

Y'all clearly aren't familiar enough with Oracle.

 

[TwiceHorn 31440]

Glad I could spur a discussion, thanks for the comments.

As an IP attorney, I am generally speaking pro-IP.  But I'm also anti-bullshit.  Oracle could be full of shit, but so could Google.  A lot of the software industry's positions on IP are laughably bad.  There has been some bullshit with software patents when the PTO was ill-equipped to handle them in the 80s and until the late 90s.  There was some copyright bullshit in the late 70s and 80s (look and feel), when patent protection was unavailable.  But people like EFF are trying to throw the baby out with the bathwater and it's a libertarian/anarchical position that is no bueno.

It's an interesting factoid that Google and Sun were in negotiation to license Java, but the negotiations fell apart.  Sun took no further action, but Oracle did after acquisition.  That doesn't tell the whole story, but prompts the question why seek a license if you don't need one?

Edited September 25, 2020 by TwiceHorn

[Rimbo 25913]

31 minutes ago, TwiceHorn said:

I think the counter to that is, there are only a few ways to write an API to use Java, not to write software more generally.

You're talking utter nonsense.

There's essentially infinite ways to write a Java API, or any other. And it doesn't matter how many ways there are if the way Google did it originated at Google.

Now, if they hired Java devs away from Sun/Oracle to do the work... there might be a point here, but the implications are still scary. It's like saying a contractor can't hire a carpenter to build a doorway the same way he had for another company on another building.

[TwiceHorn 31440]

1 minute ago, Rimbo said:

You're talking utter nonsense.

There's essentially infinite ways to write a Java API, or any other. And it doesn't matter how many ways there are if the way Google did it originated at Google.

Now, if they hired Java devs away from Sun/Oracle to do the work... there might be a point here, but the implications are still scary. It's like saying a contractor can't hire a carpenter to build a doorway the same way he had for another company on another building.

I'm not sure who's talking nonsense here.

The Java APIs could have been written probably in an infinite number of ways by the Sun engineers who wrote them.  But they were written one way.  Google copied that one way.  One question is, could Google have written them another way to accomplish the same end.

[Rimbo 25913]

4 minutes ago, TwiceHorn said:

Glad I could spur a discussion, thanks for the comments.

As an IP attorney, I am generally speaking pro-IP.  But I'm also anti-bullshit.  Oracle could be full of shit, but so could Google.  A lot of the software industry's positions on IP are laughably bad.  There has been some bullshit with software patents when the PTO was ill-equipped to handle them in the 80s and until the late 90s.  There was some copyright bullshit in the late 70s and 80s (look and feel), when patent protection was unavailable.  But people like EFF are trying to throw the baby out with the bathwater and it's a libertarian/anarchical position that is no bueno.

It's an interesting factoid that Google and Sun were in negotiation to license Java, but the negotiations fell apart.  Sun took no further action, but Oracle did after acquisition.  That doesn't tell the whole story, but prompts the question why seek a license if you don't need one?

That's because the bulk of IP law -- even those intended to apply to software -- are like trying to apply automotive laws to fish. Software is much more like construction than it is authorship or similar creative works.

So yeah, we're "throwing the baby out with the bathwater" because it's not our baby, in fact that isn't even a baby; it's a skinned rabbit.

[Celery Man 20933]

7 minutes ago, Rimbo said:

 

Y'all clearly aren't familiar enough with Oracle.

 

yeah, I haven't waded into the specifics of what is being argued here at all (and holy shit I remember when this first popped off a billion years ago), and I'm not "Google is a kind benevolent force in the world here to simply make our lives better", but Oracle is the worst and i think it's a fair bet that them getting their way will lead to a general increase of misery in the world.

[Rimbo 25913]

Just now, TwiceHorn said:

I'm not sure who's talking nonsense here.

The Java APIs could have been written probably in an infinite number of ways by the Sun engineers who wrote them.  But they were written one way.  Google copied that one way.  One question is, could Google have written them another way to accomplish the same end.

You've responded to my accusation of nonsense with even more nonsense, and it's becoming patently obvious that you have no clue what you're talking about in the realm of software.

You're embarrassing yourself. And you don't even know how badly you're doing it.

[blacklab 11926]

 

1 hour ago, Enchubben said:

fuck the brockermeyers Google. Go Oracle!!

9 minutes ago, Rimbo said:

Y'all clearly aren't familiar enough with Oracle.

Rimbo is correct. This is more of a "aggy vs ou game" that you root for a meteor rather than someone actually winning.

[TwiceHorn 31440]

1 minute ago, blacklab said:

 

Rimbo is correct. This is more of a "aggy vs ou game" that you root for a meteor rather than someone actually winning.

My rooting for Oracle somewhat is based on being pro-IP.  And recency bias. Oracle are shitheads.  Google are shitheads.  Google has been more shitheaded more recently than Oracle. 

[TwiceHorn 31440]

14 minutes ago, Rimbo said:

You've responded to my accusation of nonsense with even more nonsense, and it's becoming patently obvious that you have no clue what you're talking about in the realm of software.

You're embarrassing yourself. And you don't even know how badly you're doing it.

Um ok.  Substance is welcome at any time.

And, do you even know what you're writing? This statement is dogshit:

That's because the bulk of IP law -- even those intended to apply to software -- are like trying to apply automotive laws to fish. Software is much more like construction than it is authorship or similar creative works.

Perhaps what you mean to say is that IP law as applied to software is like automotive laws to fish.  But that is not what you said.

And, I would agree that software is so different from other types of innovation that existing legal regimes of all types pose problems, including and especially the license/owner model.

Edited September 25, 2020 by TwiceHorn

[Dbeasy 15042]

The function of an api in general is fairly limited. It is the exchange of information in a format to enable an action by software using the API

However, there an an infinite number of ways to actually write out the code steps to transfer that information.

If Google copies almost verbatim those code steps, then maybe Oracle has an argument, although even then I’m not sure I buy even that.

The fact that it is java or any other language is irrelevant, except for the licensing structure for Java.

I agree with other that eff both of them. Go spend another $100m fighting it out.

[TwiceHorn 31440]

10 minutes ago, Dbeasy said:

If Google copies almost verbatim those code steps, then maybe Oracle has an argument, although even then I’m not sure I buy even that.

They did copy verbatim.  

Let's get something clear, because the software industry has a habit of promoting misinformation about the law:  this case DOES NOT stand for the proposition that ALL APIs ARE COPYRIGHTABLE.  It does stand for the proposition that SOME APIs MAY BE COPYRIGHTABLE.

I also might add that the impact of this on the software industry may be more limited than it seems, because language ownership in single private hands is comparatively rare.  This is a kind of weird interaction between open source or quasi-open-source languages/systems.

Edited September 25, 2020 by TwiceHorn

[Captainant 27638]

3 minutes ago, TwiceHorn said:

I'm not sure who's talking nonsense here.

The Java APIs could have been written probably in an infinite number of ways by the Sun engineers who wrote them.  But they were written one way.  Google copied that one way.  One question is, could Google have written them another way to accomplish the same end.

In this case, there are a few different factors that are subtle but make a difference

1. Google wrote their replacement API in a cleanroom, not referencing or looking at Oracle's code
2. The basis of Oracle's complaints is that that Google's recreations use the same declarative structures and interfaces as Oracle's IP.
3. Because of how computing works (at a logical, bits and gates level) these declarative structures and interfaces look very similar between Oracle and Google's implementations.
4. There's only so many ways to parse or package a message to be interpreted by an API. By its nature there is going to be an "optimal" implementation that is deterministic based on the underlying mathematics of how the data is parsed and processed.
5. Oracle's specific complaints against Google is that they mimicked the interface and model to ensure compatibility with existing software, but it is established case law that doing so in a clean room implementation is NOT copyright violation (NEC v Intel, 1990)

So in my (technically guided and based) view, the question being ruled on is: If Oracle builds an optimal implementation, does that mean that nobody else can independently develop the same optimal solution?

I get that how that maps into law may break that interpretation, but as someone who has a degree in computer science this is fucking looney tunes that we're at this point.

[washparkhorn 9939]

Google should win this.

Oracle should have patented their file structure if they wanted protection. 

I don't see this as a copyright case; it is a patent protection case without a patent. No patent - no protection.

Section 102(b) of the Copyright Act - "In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work." 

It sounds like the lower courts lost sight of the forest for the trees. But I could be deadass wrong, . . . 

[bernorange 8359]

I had to read the background section on wikipedia (dumbed down without lawyer lingo) to try and get a sense of what this case is about.  It says in part:

Quote

...
Google chose to develop a cleanroom version of the Java Standard Edition libraries, developing the libraries from a completely fresh start without any access to Sun's code. This became the engine behind Android's Dalvik virtual machine, a core part of the new system. Part of the virtual machine included 37 API calls and around 11,500 lines of code deemed central to Java, which were taken from Apache Harmony, an open-source cleanroom Java implementation developed by the Apache Software Foundation (ASF). Prior to this, the ASF had tried to obtain necessary licenses from Sun to support the Apache Harmony project as to call it an official Java implementation, but could not in part due to incompatible licensing with Java's GNU General Public License and ASF's Apache License, nor could it gain access to the Java TCKs to validate the Harmony project against Sun's implementation.[14][15] Though Google stated they used this code to ensure interoperability with the Java Standard Edition for other programmers,[5] during the second appeal hearing, Google stated that it had used this code for commercial reasons to rapidly complete Android and to avoid the "drudgery" of recreating the code.[12] ASF ceased maintaining the Apache Harmony in 2011, leading Google to take over maintenance of these libraries.
...

Based upon this understanding, I tend to agree with Captainant's previous post.

[TwiceHorn 31440]

11 minutes ago, Captainant said:

In this case, there are a few different factors that are subtle but make a difference

1. Google wrote their replacement API in a cleanroom, not referencing or looking at Oracle's code
2. The basis of Oracle's complaints is that that Google's recreations use the same declarative structures and interfaces as Oracle's IP.
3. Because of how computing works (at a logical, bits and gates level) these declarative structures and interfaces look very similar between Oracle and Google's implementations.
4. There's only so many ways to parse or package a message to be interpreted by an API. By its nature there is going to be an "optimal" implementation that is deterministic based on the underlying mathematics of how the data is parsed and processed.
5. Oracle's specific complaints against Google is that they mimicked the interface and model to ensure compatibility with existing software, but it is established case law that doing so in a clean room implementation is NOT copyright violation (NEC v Intel, 1990)

So in my (technically guided and based) view, the question being ruled on is: If Oracle builds an optimal implementation, does that mean that nobody else can independently develop the same optimal solution?

I get that how that maps into law may break that interpretation, but as someone who has a degree in computer science this is fucking looney tunes that we're at this point.

Thank you.  That is a very solid comment.

"Clean-rooming" can avoid infringement because it limits or eliminates "access" to the copyrighted work.  However, it is also established that copying may occur "from memory," without "direct access" to the copyrighted work.  Most "musical" copyright infringement occurs "from memory."

So "clean-rooming" is not a "kings x" to avoid infringement liability entirely.  And the verbatim nature of the copying belies a lack of access and the "independent development" that it implies.

Also, this statement may be telling:  "If Oracle builds an optimal implementation, does that mean that nobody else can independently develop the same optimal solution."  

If Google could have written a "sub-optimal solution" without verbatim copying, should the copyright law have forced them to do so or take a license from Sun/Oracle if they wanted to use the "optimal implementation."

Edited September 25, 2020 by TwiceHorn

[TwiceHorn 31440]

3 minutes ago, washparkhorn said:

Google should win this.

Oracle should have patented their file structure if they wanted protection. 

I don't see this as a copyright case; it is a patent protection case without a patent. No patent - no protection.

Section 102(b) of the Copyright Act - "In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work." 

It sounds like the lower courts lost sight of the forest for the trees. But I could be deadass wrong, . . . 

It is right at that intersection, yes.  But given that the Federal Circuit is THE patent court in this country, you would think it would be unsympathetic to an attempt to obtain a patent by other means, which is what you are saying this is.

[Viking 301]

1 hour ago, TwiceHorn said:

Well we aren't talking about "some APIs," we're talking about specific Java APIs, and not the ones without implementations, which wouldn't seem to involve copying anything.

And how much copying is enough to constitute infringement is not a question raised here.  The standard is "substantial similarity."  To constitute copying, something is pretty much going to have to be a verbatim copy.

It appears that your attitude is emblematic of "software anarchists," who believe they should be able to copy anything with impunity.

If you are willing to provide more concrete examples, that might be helpful.

Google didn't copy the implementation. If I recall they specifically did a clean room implementation.

Here's a simple example from java itself. The java.sql package is an API: https://docs.oracle.com/javase/8/docs/api/index.html?java/sql/package-summary.html

Database vendors implement this API so java apps can use their databases. Here is PostgreSQL's implementation https://jdbc.postgresql.org/documentation/publicapi/index.html

Is PostgreSQL infringing the copyright of this API? Can Oracle (also a database vendor) sue PostgreSQL for copyright infringement?

What PostgresSQL did is the exact thing that Google did, only Google implemented a much larger API.

[Enchubben 5510]

31 minutes ago, blacklab said:

 

Rimbo is correct. This is more of a "aggy vs ou game" that you root for a meteor rather than someone actually winning.

So the lawyers are the only winners, shocking.

[JBJ 3654]

 

2 minutes ago, Viking said:

Google didn't copy the implementation. If I recall they specifically did a clean room implementation.

Google wrote it's own implementation, but copied the declaration.  And absolutely copied not just built similarly.

[TwiceHorn 31440]

3 minutes ago, Viking said:

Google didn't copy the implementation. If I recall they specifically did a clean room implementation.

Here's a simple example from java itself. The java.sql package is an API: https://docs.oracle.com/javase/8/docs/api/index.html?java/sql/package-summary.html

Database vendors implement this API so java apps can use their databases. Here is PostgreSQL's implementation https://jdbc.postgresql.org/documentation/publicapi/index.html

Is PostgreSQL infringing the copyright of this API? Can Oracle (also a database vendor) sue PostgreSQL for copyright infringement?

What PostgresSQL did is the exact thing that Google did, only Google implemented a much larger API.

Again, we aren't talking a general case of APIs.  That issue remains open.  We are talking about a specific subset of Java APIs.  The clean room issue is discussed above.

[TwiceHorn 31440]

4 minutes ago, Enchubben said:

So the lawyers are the only winners, shocking.

Well both companies are doing it with their eyes open.  So fuck em.  Moar fees pls.

Edited September 25, 2020 by TwiceHorn

[washparkhorn 9939]

Is this the ultimate issue: whether copying the hierarchical structure of the library of implementation program and the declarations that reflect that hierarchical structure infringes on Oracle's copyright in that structure and those declarations.

[TwiceHorn 31440]

25 minutes ago, bernorange said:

I had to read the background section on wikipedia (dumbed down without lawyer lingo) to try and get a sense of what this case is about.  It says in part:

Based upon this understanding, I tend to agree with Captainant's previous post.

Oh fuck. If true, that's not a good clean room.  They copied it not from Sun/Oracle but from another unlicensed "infringer."  Fucking hell.

Of course, the clean rooming doesn't help at all when there are big chunks of verbatim similarity.

Note:  "copying" is the wrongful act prohibited by copyright.  Because we don't often have direct evidence of copying, it may be inferred from access to the copyrighted work and substantial similarity between the infringing and copyrighted works.  When the infringing work is not merely substantially similar,, but identical to the copyrighted work, the access prong tends to fall away and "clean rooming" becomes irrelevant.

Edited September 25, 2020 by TwiceHorn

[TwiceHorn 31440]

Just now, washparkhorn said:

Is this the ultimate issue: whether copying the hierarchical structure of the library of implementation program and the declarations that reflect that hierarchical structure infringes on Oracle's copyright in that structure and those declarations.

It's a good question, but not the way it has been framed to the courts.

[bernorange 8359]

Just now, washparkhorn said:

Is this the ultimate issue: whether copying the hierarchical structure of the library of implementation program and the declarations that reflect that hierarchical structure infringes on Oracle's copyright in that structure and those declarations.

I don't think so.

Oracle/Sun developed and published an API.  This is loosely a design for communicating with an underlying implementation (machine).  Oracle/Sun developed their machine (Java) that complies with the API specifications.  Google (et al apparently) developed their own machines that comply with the API specifications.  As CaptainAnt mentioned, there is really only one optimal way to implement a narrow algorithm for a specific API function/method.  Even developed independently, competent programmers are going to discern the same or essentially the same implementation code.

[deadshank 24652]

I used to shack up with a girl 10 years my younger that happened to work at Oracle.  She was a maniac.  She also never went to work.   They kept paying her and never fired her.  She quit because she was bored.  She was hot.  
 

That’s all I got.  
 

 

[TwiceHorn 31440]

1 minute ago, bernorange said:

I don't think so.

Oracle/Sun developed and published an API.  This is loosely a design for communicating with an underlying implementation (machine).  Oracle/Sun developed their machine (Java) that complies with the API specifications.  Google (et al apparently) developed their own machines that comply with the API specifications.  As CaptainAnt mentioned, there is really only one optimal way to implement a narrow algorithm for a specific API function/method.  Even developed independently, competent programmers are going to discern the same or essentially the same implementation code.

Again, as with captainant's use of the term "optimal,"  if Google could have used "essentially the same" rather than a verbatim copy, that would seem to undercut the notion that the "optimal implementation" is not copyrightable.

[JBJ 3654]

Going back to my industrial process analogy:

Let's say Oracle owns a machine that turns coal to diamonds.  It's real particular about the type of coal and has a series of automatic belts for putting the coal into the machine in very specific timings and quantities in order for it to work.  The coal miners have learned this method over time and have their coal preparation facilities built and workers trained specifically for Oracle's machine and method.

Google built a unique machine that can do the same thing.  But upon realizing a new process would be too complicated for coal miners to readily adapt to, that they adapted Oracle's system of automatic belts in order to make it easier for coal producers to use the machine. This could have and has been done any number of ways, but Google did it this way specifically so Oracle-using coal miners did not have to adapt to a new machine.

Edited September 25, 2020 by JBJ

[bernorange 8359]

Well, here you (TwiceHorn) get a "splitting hairs" type of situation.  Let's say Sun/Oracle develops their machine (Java) using the C programming language.  They write code to implement API function/method "Multiply" which accepts parameters X and Y and does some math (say X times Y) to calculate Z which it then returns.  Their code in C is straightforward:

Z = X * Y

When they compile their C code, the C language compiler turns the C language code into machine code using the C language libraries and such.

Now Google (or Apache or whoever) sets out to develop their own implementation for the API function/method "Multiply".  They develop their code in Assembly language (lower level language - closer to machine code).  Their code does the exact same thing, accepting parameters X and Y, multiplies them together and returns Z.  Their code follows the same procedural order, but maybe differs slightly in the use of registers or such taking advantage of efficiencies that can only be done when programming in Assembly.

At a high level (ie, if you were to "upscale"/translate the low level Assembly code into C code), it would be the same.  At the finished/compiled level, the code technically isn't exactly the same.  From a computer science perspective, there is really only one way to create an implementation for this very simple API function/method.

 

Edited September 25, 2020 by bernorange

[TwiceHorn 31440]

Just now, JBJ said:

Going back to my industrial process analogy:

Let's say Oracle owns a machine that turns coal to diamonds.  Its real particular about the type of coal and has a series of automatic belts for putting the coal into the machine in very specific timings and quantities in order for it to work.  The coal miners have learned this method over time and have their coal preparation facilities built and workers trained specifically for Oracle's machine and method.

Google built a unique machine that can do the same thing.  But upon realizing a new process would be too complicated for coal miners to readily adapt to, that they adapted Oracle's system of automatic belts in order to make it easier for coal producers to use the machine.

I totally get the urge to analogize, but this is really kind of it's own specific thing.  Sui generis if you will.  So analogies are going to be kind of flawed.

One question that your analogy prompts is "could Google have accomplished the same thing without copying the Sun/Oracle code?"  If the answer to that is "yes," then it undercuts Google's position substantially.  But the question still remains "should Google have had to accomplish it without copying the Sun/Oracle code."

 

[bschoolprof 6252]

I really want to understand what's going on here.  I tried.  I failed.  

[El Diablo 4232]

Fair use is getting blurrier and blurrier. Andy Warhol vs Goldsmith is another weird one, though not software. 

[Blotto 23490]

Lulz at some of the dinosaurs on the bench trying to wrap their minds around some of these concepts. 

[TwiceHorn 31440]

Oh yeah.  I forgot to add why I am hostile to Google in this, and generally.

It is because Google has blatantly attempted to fuck up the patent system for its own benefit while ignoring the damage it has done to other customers of the patent system.

I get that IP maybe doesn't work quite right for software.  And maybe for drugs too in some cases.  But it works well for a whole lot of American industry and shouldn't be fucked with generally for your own selfish ends.

Edited September 25, 2020 by TwiceHorn

[TwiceHorn 31440]

10 minutes ago, bernorange said:

Well, here you (TwiceHorn) get a "splitting hairs" type of situation.  Let's say Sun/Oracle develops their machine (Java) using the C programming language.  They write code to implement API function/method "Multiply" which accepts parameters X and Y and does some math (say X times Y) to calculate Z which it then returns.  Their code in C is straightforward:

Z = X * Y

When they compile their C code, the C language compiler turns the C language code into machine code using the C language libraries and such.

Now Google (or Apache or whoever) sets out to develop their own implementation for the API function/method "Multiply".  They develop their code in Assembly language (lower level language - closer to machine code).  Their code does the exact same thing, accepting parameters X and Y, multiplies them together and returns Z.  Their code follows the same procedural order, but maybe differs slightly in the use of registers or such taking advantage of efficiencies that can only be done when programming in Assembly.

At a high level (ie, if you were to "upscale"/translate the low level Assembly code into C code), it would be the same.  At the finished/compiled level, the code technically isn't exactly the same.  From a computer science perspective, there is really only one way to create an implementation for this very simple API function/method.

 

That's a lot like my e=mc^2 analogy.  That's "algorithmic" and is not going to get much or any copyright protection.  That's a pretty easy case.

This particular one is not so easy.

[bernorange 8359]

I tried to get a couple software (sorry, computer machine) patents.  The process has left me scarred and jaded with respect to the USPTO and the process/system.  The whole industry is a FYPM traveshamockery as far as I can tell.

[JBJ 3654]

11 minutes ago, TwiceHorn said:

I totally get the urge to analogize, but this is really kind of it's own specific thing.  Sui generis if you will.  So analogies are going to be kind of flawed.

One question that your analogy prompts is "could Google have accomplished the same thing without copying the Sun/Oracle code?"  If the answer to that is "yes," then it undercuts Google's position substantially.  But the question still remains "should Google have had to accomplish it without copying the Sun/Oracle code."

 

They could have and every other tech company has in fact.  Google has it's own machine, but it is used the exact same way Oracle's is.  And it's the only machine used the exact same way.