[getpolarized]

Here's a good analogy for non-programers.

You're building a new car company.

You see all the existing cars, the roads, and the garages.

So you measure the width of the car and the garage and you decide that the car should be exactly 6' wide.

Then you ship your car. Enjoy 15 years of market success.

Then you find that Oracle created the first cars and says they have copyright on the width of your car.

That's how stupid this is...

[hashtagmarkup]

What if Google copied the name of Oracle's car and stamped it on the back? What if Google also copied the name of the in-car entertainment system and put all the buttons in the same place and stamped the same name on the dash? What if the shape of the car was exactly the same, and was only available in the exact same colors as Oracle's car? What if Google opened a car dealership next door to every single Oracle dealership with giant signs that said their cars are effectively interchangeable?

That's how stupid car analogies are...

[mehrdadn]

I don't think that's a good analogy at all. A car's precise width is not part of its interface nor a part of its design anyone particularly cares about. Its interface is more like, the design of the UX... how the steering wheel, dashboard, etc. look and feel. If a company copied those but changed everything underneath so that people could buy a cheaper alternative without seeing a difference on their usage... I don't know, is that actually legal?

[throwaway373438]

The car's width is absolutely its interface with respect to roadway interoperability.

We aren't talking about user interfaces here. This lawsuit does not involve UX. We are talking about interfaces between functional components. The analogy is spot on.

[mehrdadn]

The interoperability width is already set by the road... you're copying the road at that point, not the car.

The lawsuit involves UX for a software framework. Which its API constitutes. The user happens to be a developer here and their usage is software development.

[throwaway373438]

And this is why Google had to copy the interface -- otherwise their runtime would not operate with other existing products. As you can see given your example, the analogy is a perfect fit.

Your argument that the copying is once-removed is unfortunately irrelevant to copyright law.

[mehrdadn]

You're saying a car that's 1 inch narrower somehow wouldn't fit the same roads?

[throwaway373438]

What does "1 inch narrower" mean to you, in the context of this analogy? Are you suggesting the interfaces have different names? Different sized arguments?

None of those things would work.

Besides, creating minor differences do not solve this problem as they would still be derivative works.

[animationwill]

> A car's width is not its interface.

Let's improve the analogy!

My proposal => that you chose to make the door handles six inches because of UX, and Oracle showed up and said they made theirs six inches before you did.

[mehrdadn]

You can't seriously suggest copying the width of 1 door handle is equivalent to copying all of the Java API, right? I don't think if Google provided an alternative for 1 function it would be in this lawsuit. Your analogy is just a tiny portion of the alternative I already portrayed: they copied the entire structure of the framework. Copy 1 door handle, 1 steering wheel, 1 precise dashboard design, 1 glove compartment, etc. and I feel your analogy would result in a lawsuit just the same?

(For reference none of this is meant to reflect my position on what the law actually is, or should be. I'm just considering analogies here.)

[salawat]

The issue is that in engineering there are certain patterns of implementation that are simply converged on.

Upholding the API copyright would be like giving Craftsman grounds to sue Ryobi for manufacturing circular saws because Ryobi combined a power cord, an electrical motor, a toothed circular blade and a housing; a state of affairs that only seems even remotely plausible to endorse because source code is inherently a read product.

You look at it and see an authored work. I look at it and I see an n-dimensional turing-space relationship diagram as expressed through transform functions.

And last I checked, you could patent some math(RSA), but not not copyright it. Eventually it's just prior art.