Good bye GPL ?

http://arstechnica.com/tech-policy/2016/05/google-wins-trial-against-oracle-as-jury-finds-android-is-fair-use/
http://arstechnica.com/tech-policy/2016/05/op-ed-oracle-attorney-says-googles-court-victory-might-kill-the-gpl/

That seems a little hysterical to me. As I understand it the dispute was about the Java API not about Oracle’s software. That being the case shouldn’t Oracle be careful what they wish for seeing as SQL originated from IBM?

@Steve: I disagree. According to this verdict you can simply disregard licencing restrictions and use a “free for open source development” tool for commercial apps by claiming “fair use”. For example Microsoft, Qt, etc must now seriously consider their business model. But it goes further than that. For example the Xojo license (like others) forbids its use to develop an IDE for other languages. But that ruling (if it stands) throws serious doubt on the enforceability of that provision if you can claim “fair use” where “unauthorized, commercial, competitive, harmful use of software” is permitted.

The case was simply about whether or not Google’s use of the Java API constituted ‘fair use’. It’s par for the course in the open source world: pypy offers the same api as Python 2.x, LuaJIT is a drop in replacement for Lua, browsers all have their own implementation of JavaScript with a common api and the Apache Harmony project that Google used for Android had the same api as Sun’s Java. Harmony and Android existed before Oracle bought Sun. This court decision just maintains the status quo.

As for Microsoft, Mono already reimplements .NET’s API.

This has been brought up before, and the EULA has changed its language a few times concerning this…
currently it says

This is not a huge issue, unless you want to disavow the tools you used, and then I can see XOJO having an issue with you :slight_smile:

but I think the major statement is this

This is VERY gray, at least in my opinion as it relates soley to the IDE. At what point would a 3rd party IDE be considered to be a “derivative” and at what point does it not? The fact that it is a code editor? What about the first quote, some level of IDE is required for ANY language… (even if just a code editor that shells to a compiler)… Is the Visual Studio IDE a derivate of the XOJO one? (or vice-versa)… or XCODE? etc… etc…

Rulings are RARELY so narrowly interpreted that they relate solely to the case brought before the court and ONLY that issue.

While the dispute might have been narrowly about the use of API’s the ramifications of claiming “fair use” are much wider than the narrow original issue.
If an API can be cloned based on “fair use” then dual licensing may be impacted - whether that was intended or not.

How so? The API is not the implementation.

It seems to me that licensing for commercial use of the code the API calls would not be affected and so the API cloner would still need to write their own implementation.

While coming up with a good API is not trivial, it is only a small part of the work.

Cloning an API of popular product can however help sales of a product as it can then serve a drop in replacement as mentioned above.

  • Karen

Read the opinion piece at http://arstechnica.com/tech-policy/2016/05/op-ed-oracle-attorney-says-googles-court-victory-might-kill-the-gpl/
It’s an interesting scenario they present and you can bet that someone somewhere will put this to the test to see how enforceable the provisions of the GPL actually are given this case as precedent.
Until that happens its all just speculation .

I think the problem with the way this has turned out is now a person/company using an API is potentially a lawsuit target that would have to defend themselves with a fair use argument.

Hopefully this will get sorted out eventually, but I would be very surprised if there was anything other than $$$$ motivating Oracle to pursue this. In fact, I would bet that if audited, Oracle would be violating the “principle” behind their suit and “stealing” from some other companies APIs.

Some interesting facts in this case.

  • Sun were licensing Java to many hardware vendors, including Samsung, HTC and so on.
  • Google went to great lengths to make their OS Java compatible, to avoid licensing costs (see the court documentation).
  • Sun didn’t pursue Google when Android first came out, at that time they thought it was a good thing.
  • As Google courted more and more phone vendors, they stopped renewing their Java licenses and Sun lost out in a big way.

The trouble is the company that created Java, is now dead. Partly because of bad management and partly because Google took their technology and rather than pay the license fee like everyone else, they cloned it.

Google profited from Sun’s investment and Sun didn’t. That said Oracle’s case really is simply trying to get some money, it should have been Sun suing them, Sun may have won.

On the flip side it means anybody can blatantly clone software and claim fair use.

Define “clone.” I was under the impression that APIs, interfaces, and “methods of operation” that were not under patent were never well protected by law, if they were protected at all. WINE and React OS come to mind. So does the first reverse engineered BIOS for the first PC clone.

Did Google steal actual code? Or write their own code which takes A and B as inputs to produce output C, just like Oracle’s would?

If it’s the latter then Google is late to the party. That has been done countless times.

Patents protect ideas
Copyright protect “written” material

Software “could” fall under both.

Dave - I don’t think an interface between two software components falls under anything unless it wins a patent. I could absolutely be wrong on that. I’m not a lawyer and I did not sleep at a Holiday Inn Express last night. But I can think of many examples where someone cloned a piece of software such that their clone used none of the actual code, but was a drop in replacement in terms of the API.

perhaps we define clone differently

to me “CLONE” means to create an identical copy
independently created code that does the “same thing”, but in a different manner is not a “clone”
therefore it “might” violate a patent as it implements a protected “idea”, but “might not” violate a copyright because it was designed differently… of course either/both distinctions would be for the lawyers to hash out (hence my use of the word “might”)

A Patent must be granted by an authorizing organization, where as a “copyright” can simply be claimed, although in most cases it may be in the parties best interest to have it registered to prove a timeline.

An API that is public should always be accessible. The code that sits behind the API is copyrighted and therefore protected, but the API itself tends to be public. Looking at how an API works without looking at the supporting code, by clean room reverse engineering (see https://en.wikipedia.org/wiki/Clean_room_design) has been done on multiple occasions by multiple companies. An API has an expected input, so you can always test what result a certain input gives, and then develop your own code that for the same input gives the same output. If we are not allowed to copy api’s (not the code) without restrictions, we probably would have just one webserver software (no landscape as it is now, with IIS, Apache, nginx, lighthttpd, xitami, or even Realbasic’s webserver in 100 lines of code (http://www.macdevcenter.com/pub/a/mac/2004/11/19/realbasic.html), one database (probably oracle, one relational database (probably INGRES), no Wine for Linux, no Android … The list could be very long.

Another point is that keeping your API’s secret is most definitely frowned upon, as Microsoft found out when they refused to hand out the specs to SMB. They lost the court case and were forced to document it.

From the case timeline (see https://fsfe.org/activities/ms-vs-eu/timeline.en.html)

[quote]11/09/2003 FSFE’s submission to the EC, regarding, specifically, Microsoft’s settlement in the United States and its ineffectiveness at bringing back competition into the market.
The Samba Team proposes that the European Union should require Microsoft to make its proprietary protocol information and its proprietary application data formats (the Microsoft Office file formats) available in a royalty-free, non-discriminatory fashion, in exactly the same way as the TCP/IP protocol itself is available.[/quote]

and then

[quote]
17/09/2007 Microsoft loses its appeal against the merit case. The EUR 497 million fine is upheld, as are the requirements regarding server interoperability information and bundling of Windows Media Player. However, the Court rejects the Commission ruling that an independent monitoring trustee should have unlimited access to internal company organization in the future.
10/2007 Microsoft announces its compliance with the Court decision: it will publish protocols and will demand 0.4% of the revenue in patent-licensing royalties, only from commercial vendors of interoperable software, and promises not to seek patent royalties from individual open source developers. The interoperability information alone is available for a one-time fee of EUR 10,000.
10/2007 The EC holds that Microsoft complied with the obligations under the 2004 Decision.
12/2007 The Protocol Freedom Information Foundation (“PFIF”) is formed and becomes licensee of the no-patent license to the interoperability information, after further negotiation with Microsoft that removes the remaining GPL incompatibilities. The Samba Team becomes a “contractor” of the PFIF and access all the documentation.
27/02/2008 The EC fines Microsoft an additional EUR 899 million fine for non-compliance with the 2004 Decision (namely, the obligation to authorise the use of the interoperability information on “reasonable and non-discriminatory” terms).
09/05/2008 Microsoft appeals to the Court seeking to overturn the EUR 899 million fine.
16/08/2008 FSFE and Samba Team seek leave to intervene as third parties.
Inter alia, the Software & Information Industry Association, ECIS, IBM Corp, Red Hat Inc, Oracle Corporation seek leave to intervene in support of the Commission, the Computing Technology Industry Association, the Association for Competitive Technology – in support of Microsoft.
20/11/2008 All of the applications to intervene have been granted.
24/05/2011 Oral hearing at the Court on Microsoft’s appeal.
27/06/2012 The general Court upheld the 2008 fine reducing it to €860 millions because of a “miscalculation” done by the European Commission[/quote]

Isn’t that ironic? Oracle was part of the team that said the European Commision was right in forcing Microsoft to open up interoperability since not opening up was an abuse of its monopoly, but apparently they want a monopoly on the Java API’s.

Needless to say this is a complex case, but to me it has little to do with code (and copyright) and a lot with interoperability/openness and not allowing monopolies to abuse their position in a fast moving market.

Phoenix created this by having one team analyze the bios & write a spec which was then implemented by a team who had no experience with the existing BIOS (hence the term clean room) and the two teams never interacted beyond this spec.
The spec was a functional spec.

I guess one of the questions in this case could, or should have been, whether Google did know the internal implementations when they created their “clone”. That could have colored the outcome - but appears not to have.

[quote=268922:@Daniel Taylor]Did Google steal actual code? Or write their own code which takes A and B as inputs to produce output C, just like Oracle’s would?
[/quote]
There appeared to be no contention that Google did steal code.
But they DID have an ulterior motive to supplant Oracle’s Java with a license free compatible version - and that loss of license revenue seems to be what motivated the suit.

It has. And done correctly like Phoenix did (complete separation of teams, the implementors have no prior knowledge of the existing implementation, have never seen the originals implementations etc etc etc - which is much harder to prove today) it’s already an established means of legally making a new product that is compatible.

However, if Google USED actual Oracle code then it seems it should be on a lot shakier ground since that WOULD be a copyright violation at the very least.

How many of us would have enough resources in case of litigation ? I wish I was Google with billions of dollars in my pocket to fend off legal annoyances. Maybe APIs are not copyrightable, but yet, Oracle tried to get at Google. Not once, but twice, until it finally lost. And BTW it intends to appeal. Sure, there are countless examples of APIs that have been emulated without problems. But one suffices to down in flames any small company.

Exactly, and that’s why in this case I firmly side with Google. If there are any questions at all as to the fact whether we are allowed to freely use public API’s, yes or no, then we all have a very big problems as developers.