Google's Opening Statement - Language is free | Hacker News Reader

Google's Opening Statement - Language is free

wired.com

67 pointshluska29 comments

First off, I am clearly in complete agreement with the argument presented.

That said, now that this case is live and in the media, it should be noted that there's a certain deep irony in it. Andy Rubin directed all GPL code (all that was practical, anyway) to be removed from the Android userspace out of fear of IP purity concerns. And in the process of all that forking he did deep harm to the open source community. Not just code was forked, but infrastructure was abandoned: compare the level of robustness and transparency of Fedora/Debian/Ubuntu/OpenSUSE packaging to that of the "I cooked a ROM, download it here!" culture in the android world.

Yet Rubin allowed the use of the developer tools of a clearly proprietary product from another company without batting an eyelash. And no one called him on it. And now they have to have this ridiculous fight. So what was the point of forking Linux again?

I personally don't care that a company is looking out for it's "best interests"* internally. I am happy that Google is effectively defending the Java community from Oracle (since Oracle doesn't wish to be it's steward.)

* If this move is actually in their best interest, and what best interest IS I have no idea. But it's their perception.

You could check out http://wiki.openmoko.org/wiki/Main_Page Not sure if they are still running these days, but they have the goals you seek in mind.

I fundamentally disagree that Android did deep harm to open source. That's just absurd.

I said it hurt the community. Your code is open. And much of it is great. But it's not something that can be meaningfully shared.

Again: compare the community work available at Launchpad or rpmfusion or build.opensuse.org (compatible, signed, versioned packages with automated dependency tracking) to "I cooked a ROM doodz!". The far and away best example of community work in the Android world (Cyanogen) is, what, five months behind the core team at even shipping the software that's already in the field? It's depressing.

And what's absolutely infuriating to me (and the point of this rant) is that you started with all that good stuff and threw it out the window chasing some ridiculous notion of IP purity that -- surprise! -- was basically worthless as you're being sued over IP anyway.

Do you really think the legal picture would be any different if Android were under copyleft?

Of course not, that wasn't my point at all. I think the community picture would be very different if Android had inherited the infrastructure and community of desktop linux instead of charging ahead with needless and inferior reimplementations of glibc, udev, init/upstart/systemd, rpm/dpkg, etc... in a way that cut them off from the rest of the linux community.

Ubuntu and Fedora users can package stuff for others in a robust way, including modifications to the system libraries. How nice would it be to fix the awful accelerometer polling on the Galaxy Nexus and point users to a RPM repo? Try out some new gesture tuning? See the point? That kind of work exists, sort of, in the Android world, but it sucks rocks.

And the reason for all this breakage was, ostensibly, "IP purity" and the desire to avoid law suits and ambiguity about the software licensing. Which, as it happens, was a fools errand anyway as the platform is being sued regardless.

Is it clear now what I'm saying?

You clearly lack the experience necessary to understand the dynamics at play here. The phenomenon you bizarrely, insultingly refer to as "I cooked a ROM doodz!" is not arrested by copyleft. Small-scale non-commercial GPL violations by individuals are absolutely rampant in a number of communities. Android would not be any different -- nor is it, in fact, since the kernel is still under the GPL.

>compare the level of robustness and transparency of Fedora/Debian/Ubuntu/OpenSUSE packaging to that of the "I cooked a ROM, download it here!" culture in the android world.

I don't really understand what your point is here -- a ROM is an image of the operating system, and as far as I know the situation is the same with mainstream Linux. You typically install a distro by downloading the iso from the web and then running it.

There was a tribe in Chile that tried to sue Microsoft for supporting their language, not sure what came of it.

http://en.wikipedia.org/wiki/Mapuche_language#Microsoft_laws...

Are they arguing that all languages are free or that the Java language specifically was free? If it's the former it seems a little strange. If I tomorrow invented some programming language that for instance described some novel way to handle concurrency, why wouldn't that be protectable? I'm not pro-patents but languages being public domain per default doesn't seem right either

If you are arguing for monopolies on ideas, that makes you pro-patent.

Aren't all patents ideas? Surely the difference between a patent or not can't be whether the ideas is a physical object or not.

If all patents were abolished tomorrow I wouldn't mind, but as so long as they aren't it seems silly that some things that really demand a high degree of creativity and innovation isn't protected. That kinda seems to be the point of patents, not like coming up with the idea of buying something with 1-click

Creativity and innovation have been happening for much of human history. Perceived unfairness that the progress of technology doesn't delineate individual's contributions isn't really something that has held back society.

On the other hand, the patent litigation game is negative sum (it destroys wealth) and mostly enriches lawyers: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1930272 http://techcrunch.com/2011/08/19/terrible-cost-patents/

In addition to the litigation costs, the prohibition to build incrementally or to reuse or rediscover ideas makes R&D more expensive for lesser results.

So making ideas exclusive has not rewarded R&D and has made society poorer overall, despite rhetoric about incentives, propriety, and fairness.

I'm not arguing for patents, only that a programming language should probably be patentable if other similar creative efforts are, I don't see why they should be exempt. I agree with you that patents are bad overall (if that's your stance) and should be revamped or abolished.

Isn't this about the implementation of the virtual machine rather than the language itself?

Funny enough, thinking about a project over the last few weeks I found myself wondering if there were any legal restrictions to implementing your own JVM with commercial purposes. I couldn't find any guidance on this at the Oracle site. Maybe the answer to this question will be borne out of this trial?

I believe all the implementation patents have been dismissed by the USPTO bar 1/2 of one which is still undergoing review. What is left is Oracle's claim of language copyright which will hopefully get the ruling it deserves.

If that is true, does that mean anyone will be free to implement any programming language developed by anyone, regardless of the designers' wishes?

Looks like there be will no monetary incentive in inventing a very innovative languages when the big guys can just come in and re-implement it and not pay a cent.

Someone should ask Wolfram what he thinks about this. ;)

There is already a trademark-based restriction on the name Java. You can't call your system Java unless it passes the TCK. You can't pass the TCK without signing some sort of license about what your implementation can and can't do.

TeX has the same concept with the TRIP test.

Ignoring the first-to-market incentive (which is substantial), is there any monetary incentive now? Has anyone else ever tried to make a business out of somehow "licensing" a general-purpose programming language minus implementation?

The concept doesn't even make sense. Do you pay someone to let you speak English? Why would you pay someone to let you "speak" any other language?

What if I ship a language with implementation but would like to prevent others from implementing it?

Sometimes language implementations are the product - things like R, Matlab and Mathematica are close examples. What if I figured out a much better programming language for Finance software and would like to prevent every other company from implementing it?

This debate is far more complex than what people think.

R's source code is released under the GPL. MATLAB has a GNU-licensed equivalent (GNU Octave), which has been around since 1988.

How could you patent a language? What are you patenting? The compiler/interpreter? The context-free grammar (assuming it has one)? The language specification?

If the specification, then you really mean a copyright, not a patent, because a language specification is just a description of a language, and you can't patent that. But just because you've placed a copyright on your description doesn't mean you hold a copyright on anything that anybody creates after having read that description - that would be ludicrous!

Sure, we can talk about the compiler/interpreter itself, but that doesn't prevent someone else from going ahead and creating a completely separate compiler that implements the same exact specification (look at the GNU Compiler Collection, more commonly known as gcc).

What you want and what the law allows may be two different things. Ideas are not subject to copyright. Specific implementations might be.

The value of Mathematica is not that "Integrate[Sin[x^2], x]" is especially brilliant syntax.

Wolfram Research is infamous for claiming copyright to the language of Mathematica threatening to sue the University of California. They have also have a long history of causing automatic translators and shims for converting Mathematica syntax code into other systems to disappear (supposedly based on the language copyright).

Do you have any meaningful citation for these events?

The only mention I can find of the incident you seem to be alluding to is a brief, vague, and un-sourced mention in a text file dating to 1991[1]. I'd be entirely unsurprised to learn that there is far more nuance to the story than what you and said file state, or that the incident was largely the result of a game of telephone.

[1] http://progfree.org/Copyright/against-user-interface-copyrig...

It's not terribly complex. The answer to your question is no.

If you ship a language with implementation and would like to prevent others from implementing it, you should not be able to and fortunately even under US law it seems you cannot.

The motivation for a state-provided monopoly on "inventions" was to foster innovation (and we should be clear this institution is a state intervention in the market and has nothing to do with "natural rights" and such). The ability of patents to provide society with this benefit is debatable but fortunately there's clear consensus that patenting/copyrighting/owning ideas is bad. And a programming language is essentially an idea rather than an artifice and it certainly seems intuitively obvious that giving people ownership of language would get really nasty, really quickly.

> does that mean anyone will be free to implement any programming language developed by anyone, regardless of the designers' wishes?

Not completely regardless of the designers' wishes. If the language designer never publishes, then he or she won't have to worry about re-implementations. Publishing is by definition (and should be) a relinquishment of some amount of control.

"monetary incentive in inventing a very innovative languages"

I will be astonished if you can cite a single example of a programming language whose inventor was inspired by the possibility of language licensing revenue.

Note that implementation licensing revenue does not count, as implementations are covered by copyright.

Reply on news.ycombinator.com