Oracle vs Google - jury in

[phil swenson]

http://arstechnica.com/tech-policy/news/2012/05/jury-rules-google-violated-copyright-law-google-moves-for-mistrial.ars

[Cédric Beust ♔]

On Mon, May 7, 2012 at 12:56 PM, phil swenson <phil.s...@gmail.com> wrote:

http://arstechnica.com/tech-policy/news/2012/05/jury-rules-google-violated-copyright-law-google-moves-for-mistrial.ars

A bit disappointing but not really surprising.

The good news is that 1) fair use hasn't been settled and 2) Judge Alsup (who will make the final legal judgment) hasn't ruled yet.

-- 

Cédric

[Fabrizio Giudici]

On Mon, 07 May 2012 21:56:31 +0200, phil swenson <phil.s...@gmail.com>
wrote:

> http://arstechnica.com/tech-policy/news/2012/05/jury-rules-google-violated-copyright-law-google-moves-for-mistrial.ars

I'm living in a period in which I'm losing most of the consideration of
reputation of people, institutions, politicians, "experts", magazines,
etc. Perhaps it's because I live in Europe. When I see statements such as
the following one this trend speeds up (emphasis mine):


"... the end result is that Java may not be as OPEN SOURCE as it was
thought to be. Oracle may be able to essentially take Java out of the
PUBLIC DOMAIN, at least as far as its use in cell phones."


We know that FLOSS has got licenses that must be respected and the key of
the trial was just to understand whether or not Google violated it, not
whether FLOSS licenses have to be respected or not, which is obvious. The
author seems unaware of the difference between LICENSED OPEN SOURCE and
PUBLIC DOMAIN. Needless to say that Java has never been in the public
domain. What the hell is that "take Java out of..." about?


"If the verdict that Android infringed copyrights stands, it could put
programmers in a difficult situation. Java is an open source language, but
now it's not clear how free programmers are to use it, since Oracle has
said that anyone following the Java APIs—which are basically sets of
instructions about how to use Java—needs a license."

Using Java != releasing an alternate implementation. It seems that the
author of the article is well positioned to win the FUD prize of the year.



--
Fabrizio Giudici - Java Architect, Project Manager
Tidalwave s.a.s. - "We make Java work. Everywhere."
fabrizio...@tidalwave.it
http://tidalwave.it - http://fabriziogiudici.it

[Casper Bang]

> Using Java != releasing an alternate implementation. It seems that the
> author of the article is well positioned to win the FUD prize of the year.
>

That's a fine line though! Embrace + extend = use, and in all other
sciences, that's how progress is made. In many ways this is version 2
of Sun vs. MS, and even if Sun won the first one, they lost the war. I
also have a very hard time seeing how Oracle could win this one... the
battle, perhaps... but not the war.

[Casper Bang]

Let me try to be a bit more precise here. You can't very well write much Java code, without needing to implement an interface or an abstract class... or a lot of these actually. The AWT/Swing event system comes to mind as one of the most obvious parts of the API where this is true. So where would you draw the line... are we only allowed to implement interfaces? Extend abstract classes? Where is this stated/formulated?

How about if you think of the Java runtime as one giant SPI... except apparently you are only allowed to come with drop-in supplements, not replacements in a few places - even if you refrain from calling your work "Java". Are we only allowed to augment JDBC, JCE, JNDI, JAXP, JBI, Java Sound, ImageIO which have been designed specifically for third part drop-ins? Again, where is this stated/formulated?

The problem once again comes from the fact that Java is openly developed (which is nice) but *not* an open standard (which sucks). Had Java been standardized under ECMA/ISO, this whole copyright case would be moot from the get go. B.t.w any verdict against Google could be seen as having wide effect on the industry, i.e. Microsoft would be free to go after Wine developers etc. etc.

[Casper Bang]

Not so fast. Here are some more recent tidbits:

http://www.groklaw.net/article.php?story=20120507122749740#Update_7

Judge:  Google have proven that Sun and/or Oracle engaged in conduct Sun and/or Oracle knew or should have known would reasonably lead Google to believe that it would not need a license to use the structure, sequence, and organization of the copyrighted compilable code.

...

Judge: Do you want all their [Google's] profits?

Oracle: No, Your Honor. 

Judge: This borders on the ridiculous. Now you are changing your tune (adding to the agreed-upon statutory damages). Based on 9 lines of copying out of 15 million? That would ge a big, big stretch.

...

Oracle: If we were, as a matter of law, able to seek disgorgement on question 1.

Google mistrial briefing tomorrow...

Judge: Zero finding of liability so far.

On Monday, May 7, 2012 9:56:31 PM UTC+2, phil swenson wrote:

http://arstechnica.com/tech-policy/news/2012/05/jury-rules-google-violated-copyright-law-google-moves-for-mistrial.ars

[Kirk Pepperdine]

An API has two consumers, those that provide the implementation and those that use the implementation. How can a copyright only apply to one and not the other?

Kirk

--
You received this message because you are subscribed to the Google Groups "The Java Posse" group.
To post to this group, send email to java...@googlegroups.com.
To unsubscribe from this group, send email to javaposse+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/javaposse?hl=en.

[Fabrizio Giudici]

On Tue, 08 May 2012 06:19:10 +0200, Casper Bang <caspe...@gmail.com>
wrote:

> Let me try to be a bit more precise here. You can't very well write much
> Java code, without needing to implement an interface or an abstract
> class... or a lot of these actually. The AWT/Swing event system comes to
> mind as one of the most obvious parts of the API where this is true. So
> where would you draw the line... are we only allowed to implement
> interfaces? Extend abstract classes? Where is this stated/formulated?

In the GPL+CPE license. It's very clear. You can link without any problem
to everything that is in the Java runtime, and that's what 99% of people
do. You can even take OpenJDK, fork it, change the API and redistributed
it under a new name and the same GPL + CPE license. No problems with the
copyright, maybe problems with patents (the old discussion about whether
GPLv2 protects enough or not). For these things you don't need Java to be
a open spec, just GPL+CPE. Then, you can write a thing from scratch, as a
"clean room implementation", and decide to call it Java (Google didn't) or
under a different name. Here it seems that there are various kinds of
problem and an open spec would help. But it's clearly a very different
thing, programmers using Java aren't concerned about an alternate
implementation of Java.

[Kirk Pepperdine]

>
>
>> Let me try to be a bit more precise here. You can't very well write much
>> Java code, without needing to implement an interface or an abstract
>> class... or a lot of these actually. The AWT/Swing event system comes to
>> mind as one of the most obvious parts of the API where this is true. So
>> where would you draw the line... are we only allowed to implement
>> interfaces? Extend abstract classes? Where is this stated/formulated?

I didn't even consider this case but I think it's spot on.

>
>

> You can even take OpenJDK, fork it, change the API and redistributed it under a new name and the same GPL + CPE license. No problems with the copyright, maybe problems with patents (the old discussion about whether GPLv2 protects enough or not).

Isn't this essentially what google did?

-- Kirk

[Fabrizio Giudici]

On Wed, 09 May 2012 07:40:54 +0200, Kirk Pepperdine
<kirk.pe...@gmail.com> wrote:

+>> You can even take OpenJDK, fork it, change the API and redistributed it

>> under a new name and the same GPL + CPE license. No problems with the
>> copyright, maybe problems with patents (the old discussion about
>> whether GPLv2 protects enough or not).
>
> Isn't this essentially what google did?

Not. They claimed they did a clean room implementation based on Apache
Harmony, but it turned out at least some files weren't a clean room
implementation. Note that the allegedly copied files weren't in Harmony
itself, which as a matter of our knowledge is really a clean room
implementation, and in fact nobody sued Apache. Oracle killed Harmony in a
different way, by means of a deal with IBM.

[Cédric Beust ♔]

On Wed, May 9, 2012 at 12:02 AM, Fabrizio Giudici <Fabrizio...@tidalwave.it> wrote:

Not. They claimed they did a clean room implementation based on Apache Harmony, but it turned out at least some files weren't a clean room implementation.

From the findings so far:

"Meanwhile, Oracle prevailed only on 9 lines of code that Google admitted prior to trial to have included by mistake and then removed from current Android. Oracle's own expert, the judge pointed out in court, valued those 9 lines of code at zero. This is 9 lines out of millions."

 

-- 

Cédric

[Fabrizio Giudici]

On Wed, 09 May 2012 10:30:19 +0200, Cédric Beust ♔ <ced...@beust.com>
wrote:

> On Wed, May 9, 2012 at 12:02 AM, Fabrizio Giudici <
> Fabrizio...@tidalwave.it> wrote:
>
>>
>> Not. They claimed they did a clean room implementation based on Apache
>> Harmony, but it turned out at least some files weren't a clean room
>> implementation.
>
>
> From the findings so

> far<http://www.groklaw.net/article.php?story=20120507122749740>
> :

>
> "Meanwhile, Oracle prevailed only on 9 lines of code that Google admitted
> prior to trial to have included by mistake and then removed from current
> Android. Oracle's own expert, the judge pointed out in court, valued
> those
> 9 lines of code at zero. This is 9 lines out of millions."

Correct. My point is not related to the amount of money that Google should
or shouldn't pay, but to the rules that govern the bits and to the fact
that developers not replicating a JDK shouldn't really worry about the
thing.



Back to the 9 lines of code, I'm pretty curious because this coude be
an(other?) important precedent. The common word on GPL so far is that if
you put even a single line of GPL code in your stuff, everything you do
should be redistributed as GPL because of the virality. This is the main
argument that I hear from legal offices of corporates to justify internal
rules which caution against the use of GPL stuff in non open projects. I
seem to recall a possible precedent on this, maybe affecting the firmware
of a commercial network piece of equipment that used GPL code and for
which the manufacturer was put in trial, but I can't recall the whole
details. From an engineering point of view I do agree that 9 / (n * 1E6)
is neglibible, but laws aren't engineering.

As a third point, my personal opinion of the trial is that emails about
the value that Google engineering gave to Java, the fact that they tried
an agreement with Sun and then later decided to go on their own is more
relevant. Only the bona fide point is still to be understood. Of course,
I'm talking of my personal opinion, not the law evaluation.

[Casper Bang]

Not. They claimed they did a clean room implementation based on Apache  
Harmony, but it turned out at least some files weren't a clean room  
implementation.

If you go ahead and actually look at the files though, it's pretty hard to argue that "Android contains line-by-line copies of Oracle' JDK":

http://www.scribd.com/doc/40316099/orclgoogcode

There are differences in block scoping (with or without curly brace), argument variable names (logic english names vs. symbols), iteration constructs (for, do and while), constant folding etc. Even if this points to reverse engineering (it does smell of decompilation) of a binary class, it's still not an infringement of Oracle IP. 

Does anyone know which specific 9 lines are we talking about?

[Reinier Zwitserloot]

No; google did not release android under a GPL2 license; they released it under a BSD-style license. This is what the fuss is about:

Google took a (supposed) cleanroom implementation, namely, Apache Harmony which was already licensed BSD-style, tweaked it quite a bit, and released this, also under a BSD license. This is perfectly fine, as long as you steer clear of the following:

A) This is then not a java; java is a trademark, and sun/oracle has put forth the obligations for using java as a name for your alternate implementation: It must pass the TCK. The TCK is not open source and you need to be licensed to run the TCK on your alternate implementation. This was the central point of contention for Apache: They wanted harmony to be officially called 'a java', but sun was unwilling to license them the TCK to make it happen unless Apache included a clause in the license for harmony to restrict its use on cell phones (the so-called 'Field of Use' restriction). Apache was unwilling to do this and threw a fit. I'm inclined to think this fit was more or less fair (as that's not really how 'things are done' in open source land, and its very obvious apache would never agree to such a clause, so it might have been nice of sun did not let them on believing that harmony would ever be qualified as a true 'java'). Google, as far as I know, is NOT in hot water over this issue at all, as they never claimed that android/dalvik is a 'java' in a way that actually upsets the trademarking rules, but it is an important precursor to this courtcase, as it explains why Apache Harmony and Sun/Oracle were already not the best of friends.

B) You don't infringe on the copyright: You cannot rerelease GPL code as BSD code unless you are the copyright holder. Therefore, android, being BSD, cannot contain any code from oracle or it would violate the GPL license of that code, unless of course its public domain code (Which I believe the interface specs are, but either way the inclusion of i.e. java.util.InputStream was never the issue). Turns out there are 9 lines in the 15 million where the judge has conclusively granted Oracle the finding that these lines were stolen. As far as I know, the 'crime', so to speak, was on the part of an apache-allied programmer, not a google engineer. The judge evidently knows his stuff and basically laughed in Oracle's face when they demanded more than a pittance in damages for these 9 lines. These 9 lines (And a number of other suspect lines) have long since been rewritten.

C) You don't infringe on any patents. Releasing source under GPL2 does not give anyone a free ride on your patents. This is an area that has yet to be tackled by this court case, but the experts say even if this one swings wildly towards oracle, the potential damages are nothing in comparison to the copyright claim as in 'B'.

[Ricky Clarkson]

Google, as far as I know, is NOT in hot water over [calling Dalvik Java] at all, as they never claimed that android/dalvik is a 'java' in a way that actually upsets the trademarking rules, but it is an important precursor to this courtcase, as it explains why Apache Harmony and Sun/Oracle were already not the best of friends.

Although Oracle before the merger were in favour of Harmony being able to get the TCK and be certified.