Everyone is talking about the Oracle vs. Google lawsuit; everything from compatibility tests to misusing open source technology has been spoken about in the forums, but really, that's not what this lawsuit is about. The reality is, this lawsuit is simply about copyright and patents - five of them to be exact. In the most glib of manners, this article will help you put the patent part of this patent lawsuit into perspective.

The Patent Part of the Oracle vs. Google Lawsuit Demystified

Per Ed Burnette: Your best hope seems to be to waste a lot of time filing as many patents as you can of your own, and hoping that anybody who sues you is unknowingly infringing on one of yours. That gives you a bargaining chip so you can do some kind of cross licensing deal. http://www.zdnet.com/blog/burnette/oracle-uses-james-gosling-patent-to-attack-google-and-android-developers/2035 Below article may explain more on the oracle patents: http://blog.headius.com/2010/08/my-thoughts-on-oracle-v-google.html

Indeed. We don't see huge patent fights between Apple and Microsoft because each could pull so many out of their vaults that the billions they make in revenues could never cover the legal bill.

The problem Google has here is that they don't have a similar nest egg of patents that they could use against Sun. Google is too young, and simply hasn't been in this space long enough. The patent infringement allegations are a great way for Oracle to twist Google's arm. And I think that's what it is. I do think it will actually lead to some positive changes, as Oracle works to make Java more compatible and accepted in the mobile space.

This lawsuit isn't necessarily doom and gloom!

Cameron,

Thanks so much for lessening a little bit of the rampant negativity on this. I think with the current economic climate, everyone is jumping on the doom and gloom bandwagon way too quickly...

In the "worst case scenario", it would not be that terrible if Google indeed did come up with something like G# that competes with Java, is "pure" open source and is not backed by either patents, copyrights or standards (and I am being quite serious, not sarcastic)...

Cheers,

Reza

In the "worst case scenario", it would not be that terrible if Google indeed did come up with something like G# that competes with Java, is "pure" open source and is not backed by either patents, copyrights or standards (and I am being quite serious, not sarcastic)...

Interesting idea.  Were that to happen, I would be drooling at the mouth!  Especially if it worked the same in a browser as it did in a server.  I'm real sick of this completely separate web-tech and server-tech great divide.  Seems the more successful you are at one side the lower the odds are that you'll be able to penetrate the other because of the time involved being utterly unproductive and what that means to you in terms of opportunity cost. What an incredible boon to the industry it would be if these distinctions were more of coding style and not of languages.

Obviously Java aimed there and missed; shoot for the stars, reach the moon.  They landed somewhere short of the vision, but way way further than anyone had gotten before and advanced the industry significantly (thank you!).  If Sun had had a strong cell phone and web browser offering in addition to their server hardware, they might have made it all the way.  Google seems in a unique position to do that right now.  How often do things line up like this.  Once every 10 or 15 years?  Might be worth another try.

On a side side, note; I like standards.  I like them better when I can contribute to them under an OSI approved license with all the protections that offers; forkable, distributable, etc.  Standards are supposed to open a market and protect against vendor lock in.  But what happens when vendors themselves get locked into to a growing standard....

 

I do think it will actually lead to some positive changes, as Oracle works to make Java more compatible and accepted in the mobile space.

Wow! What a load of Bart Simpson!

1) Mobile Java is already widely accepted. Android sales are now greater than iPhone sales, and growing fast. Or do you think mobile phone vendors will flock to Android if they are required to pay for Java license to Oracle?

2) Compatible with what? Dalvik is pretty much perfectly compatible with core Java libraries. There's nothing else to be compatible with. Nobody cares about the J2ME anymore and Sun/Oracle doesn't have anything else to offer.

I'm not sure if we're on opposite sides of this debate.

>>Android sales are now greater than iPhone sales...

Exactly. And more to the point, the ability to write Android apps makes Java 'kewl' again with the young college kids who want to write apps for their phones. Android has been a Godsend for Java, and it's not in Oracle's best interest to mess that up. If anything, it's in Oracle's best interest to make things better. So, pulling in the reigns a bit, and getting Android closer to the write-once, run anywhere mantra, and getting the Google team to work on cross platform compatible code would be a good thing, not a bad thing.

>>Dalvik is pretty much perfectly compatible with core Java libraries...

Indeed, that's what we love about it. It's the code that's generated that's not. Things would greatly improve if Oracle could push Google to concentrate more on that area.

In the end, these topics don't have anything to do with the lawsuit though. This is really just wishful thinking, that somehow Sun might be able to put pressure on Google to make Android even better and more likely to be adopted by cell phone makers. As it stands right now, none of this is in the lawsuit. The lawsuit is simply about five patents and copyright, and that's it.

>>Android sales are now greater than iPhone sales...

So, pulling in the reigns a bit, and getting Android closer to the write-once, run anywhere mantra, and getting the Google team to work on cross platform compatible code would be a good thing, not a bad thing.

Cross-platform with, exactly, what? Android apps would look ugly on laptop screen. Mobile apps don't look good even in xPad format.

Or do you suggest porting (shudder) SWING to Android?

>>Dalvik is pretty much perfectly compatible with core Java libraries...

Indeed, that's what we love about it. It's the code that's generated that's not. Things would greatly improve if Oracle could push Google to concentrate more on that area.

Which generated code? .dex-files? They are just an implementation detail, it's even possible to de-compile them to Java. What would we gain if Android suddenly switched to plain .jar-files?

I'm serious, what could we (as developers) gain if Android is switched to standard JVM (say, OpenJDK)? Given the constraints of mobile devices, of course.

In the end, these topics don't have anything to do with the lawsuit though. This is really just wishful thinking, that somehow Sun might be able to put pressure on Google to make Android even better and more likely to be adopted by cell phone makers. As it stands right now, none of this is in the lawsuit. The lawsuit is simply about five patents and copyright, and that's it.

Yes. This lawsuit is not about making Android better. This lawsuit is about patent trolling for pure monetary gain.

Java ME is dead ..

As we all know, bytecode is slow, and native code is fast. Similarly, gasoline is slow to ignite, while nitrous oxide is fast.

I know this is a bit off topic but for your safety and the safety of those reading this, gasoline is extremely volatile and its fumes are extremely quick to ignite.  Nitrous oxide, on the other hand is a non-flammable gas.

Nitrous oxide can be used as a fuel oxidizer.  It is not explosive or flammable by itself.

Interpreting functions utilizing a hybrid of virtual and native machine instructions.  For this one, I think you are unduly glib.  It's basically a patent for the idea of partial native compilation.  I could be wrong but it's basically a patent for a core part of JIT compilation.

For Method and apparatus for pre-processing and packaging class files  I think you may have missed the point of this one.  The main point seems to be taking redundant information from a number of classes and storing them in one spot that is shared across all the classes e.g. the String constant pool.

Method and apparatus for resolving data references in generated code The patent isn't saying that direct is faster than direct.  It's saying that the VM can replace indirect references with direct references at runtime.  Think inlining.

System and method for dynamic preloading of classes through memory space cloning of a master runtime system process The patent isn't for using memory vs. disc.  As I read it the patent is for sharing class definitions across different VMs.  I believe this is something that was implemented in Hotspot fairly recently.

I'm not saying this makes these valid or novel but we should at least discuss the real meaning of these patents.   I don't think oversimplification helps.

I think you've made some great points, Mr. Watson.

Indeed, James' comments won't sink Oracle, but there's no doubts that his comments will be used to their fullest by the defense lawyers if this thing ever goes to trial, which I highly doubt it will.

I'm not saying this makes these valid or novel but we should at least discuss the real meaning of these patents.   I don't think oversimplification helps.

I agree, the article sounds more like a monologue from a radio talk show host than any serious analysis and simply dismisses the patents out of hand.

Everything looks obvious in hindsight. Many patents look obvious to domain experts when they put their thought to the problem, especially when lead by the patent. "Well of course that's how you could do it!"  Yes, of course. Now. Now that you've read the patent.

That doesn't mean it wasn't "novel, and non-obvious" at the time.

Let's take the JIT patent. It's from 2005. 2005?? That's pretty damn late in the VM game. Post .NET (You'd think that maybe MS might have thought about the JIT problem a bit back then), and it's ancient in terms of Smalltalk VM development. 2005? So, perhaps, maybe, assuming reasonable searches at prior art, nobody else who are "domain experts", practicing in the industry, seems to have implemented this concept as described. Or if they did, they certainly didn't tell anyone about it.

Perhaps Apples 68K emulation on PPC is prior art to this, or, arguably, it's backwards (that would be "byte code" (i.e. 68K machine instructions) embedded in "native" (i.e. PPC code)). But clearly Apple is aware of this technique and patent, since they build and support their own JVM for the Mac. Maybe they have a patent sharing agreement with Sun/Oracle, so the conflict never came up.

And we won't even mention IBM that patents pretty much anything someone goes "What about...". They've been doing VM work for, well...forever I think.

Simply, JITs aren't new, and Sun wasn't the only company working on them, and this is a pretty late patent, all things considered. So, perhaps, there actually may be something novel being done here.

 

Simply, JITs aren't new, and Sun wasn't the only company working on them, and this is a pretty late patent, all things considered. So, perhaps, there actually may be something novel being done here.

It seems rational to assume that because a patent was issued that the patent was novel and non-obvious.  But the reality is that the patent system is broken.  There are far too many patents being issued for anyone (including the patent office) to determine whether they are novel.

If you haven't seen this, it's worth a look.  Note: the following patent was reexamined so you need not worry of being subject to triple damages if you were you take advantage of this method after having read the patent.

http://www.newscientist.com/article/dn2178-boy-takes-swing-at-us-patents.html

http://www.freepatentsonline.com/6368227.html

The original test for a patent was that it had to be a machine (or machine assembly) that was not described in literature nor would an expert mechanic be able to create it by looking at existing literature.

Obviously, not just physical machines are patentable now.  But if we compare the kinds of patents that are being issued now at a rate of 20 per minute or something crazy like that, they don't meet the prior art test above.  They don't even come close.

It seems rational to assume that because a patent was issued that the patent was novel and non-obvious.  But the reality is that the patent system is broken.  There are far too many patents being issued for anyone (including the patent office) to determine whether they are novel.

Then where are the conflicting patents in this case? I would expect Google to simply cough them up, and show not just that the specific patent is voided, but that the entire system is in question.

The point is that that VM space is very competitive, and it is dominated by companies who wield patents with a heavy hand (such as IBM and MS). Whether people like software patents or not is irrelevant, as they are accepted law currently in the US. As for the Patent system, of course it's flawed. Name one public institution that isn't flawed, each with egregious examples.

But those are outliers and edge cases. Any suggestions to update and modify the patent system are tweaks and clarifications versus outright destruction and do over. That pretty much suggests that it MOSTLY works, MOST of the time. 

This VM patent is recent, it's not some continuation. It's in a very competitive space, where "companies patent anything and everything", yet this passed a novelty and prior art test in order to be issued. I haven't seen a flurry of "here's the prior art" posts scattered about the tech websites. Just the hand wavy "patents suck, they're broken, there's never any basis" posts.

But neither Sun/Oracle, Google or any of the other large tech companies are patent trolls. This case wasn't filed in East Texas. It's not Big Bad Corp against stay at home hacker. It's deeeeep pockets and big guns pointing at each other. Ships of the line in the open sea. Battle of Jutland stuff here.

is it Litigation over Competition? Of course it is, but if Google is infringing, why should they get a free ride. Litigation is a component of competition, just hard truth. Will it likely be settled out of court? Yes.

But that doesn't make the case baseless.

 

The point is that that VM space is very competitive, and it is dominated by companies who wield patents with a heavy hand (such as IBM and MS). Whether people like software patents or not is irrelevant, as they are accepted law currently in the US. As for the Patent system, of course it's flawed. Name one public institution that isn't flawed, each with egregious examples.

But those are outliers and edge cases. Any suggestions to update and modify the patent system are tweaks and clarifications versus outright destruction and do over. That pretty much suggests that it MOSTLY works, MOST of the time.

What you are ignoring here is that before the mid-90's, software was not considered patentable.  Before that time, there were no such patents.  At that time the interpretation of patent law was changed (no laws changed) and ever since then the number of patents issued has increased very rapidly.

We aren't talking about the entire history of patents in the US or even 50 years.  We are talking about the last decade-and-a-half or so.  And people are making legitimate arguments that since that change, things have gone completely haywire.

The reality is that today, if you write software, you are violating patents.  And I don't suggest you try to see if you are because if you do so, you greatly increase your exposure to damages.

That you are unlikely to be sued is really beisde the point.  Today, everyone validates patents and enforcement of those patents is effectively random.  The fail to serve the intended purpose and have created a tax on innovation.  The only people really benefiting in this system are the patent lawyers.

>>The reality is that today, if you write software, you are violating patents.  And I don't suggest you try to see if you are because if you do so, you greatly increase your exposure to damages

+1

That you are unlikely to be sued is really beisde the point.  Today, everyone validates patents and enforcement of those patents is effectively random.  The fail to serve the intended purpose and have created a tax on innovation.  The only people really benefiting in this system are the patent lawyers.

correction: "Today, everyone violates patents..."

It’s going to be an obstacle for Oracle to get a court of law to take a serious look at a their lawsuit when one of the most respected names in the Java world has described his own patent filings as "goofy."

I doubt that blogs are admissable evidence and Oracle has these patents.  It's going to take more than a geek's opinion (regardless of how famous that geek may be) on the web to invalidate those patents.

Thank you! The problem as you indicated is that people granting patents and people arguing these cases (probably both lawyers) are not smart enough to understand what you have written. I think there is a lot of sleaze going on at Oracle.

This article is completely biased. I am about to read the patents referenced in the article, which all but one have been summarized in a most biased manner.  All of the patents seem to encapsulate creative performance enahncements to the JVM which Oracle spent the time end effort to both implement and Patent. Surely these valuable "ideas" are potentially Patentable, more so than a screw top to a soda bottle, which is quite a bit more obvious and was worth man millions in the 60s.

Overall, the article stinks of bias and is of little use to someone trying to weight the merits and potential outcome of the case, which could have significance to both Oracle and Google. Try and be a little more professional in your "opinions"

Surely these valuable "ideas" are potentially Patentable, more so than a screw top to a soda bottle, which is quite a bit more obvious and was worth man millions in the 60s.

Nope. Every one of these patents is either obvious or documents a particular solution. Hell, one of them is essentially a patent on the classic Unix 'fork' system call!

Stephen wrote, 

All of the patents seem to encapsulate creative performance enahncements to the JVM which Oracle spent the time end effort to both implement and Patent.

All of the patents arguably seem to encapsulate creative performance enhancements to the JVM which Oracle spent the money and effort to buy and defend in court.

Unless I'm grossly mistaken, whether these patents actually encapsulate creative enhancements, is, in fact, the entire question under debate. (Well, half the question. The other half is whether or not Google used these enhancements.) 

Blithely stating that these patents seem to be creative, without providing any argument why, strikes me as being no less biased than the article.

As of now GNU own the licensing terms for java.  Its interesting now that GNU has to decide whether to sue Oracle for violating GPL license that allows free distribution.

Stallman Sr.

Folks,

It looks like Oracle is probably going to talk about their perspectives on this at JavaOne: http://www.oracle.com/us/javaonedevelop/062264.html. Not saying it will quell the debates on open source, intellectual property, patents, standardization any but at least it is another crucial perspective to consider. Given Ellison's persona, I personally expect a say-it-like-it-is, take-it-or-leave it talk...

Cheers,

Reza

Shucks to Oracle - they always were a bit late in the game! The following were in Sybase V3 c1988

1) c Function Pointers to assembler instructions sound remarkably similar

2) Sybase stored procs which resolved the first time you ran them.

3) Again Sybase stored procs did just this cloning the proc if you ran it concurrently

4) By only storing the unique bits, rather than the shared bits Sybase got the user footprint down to 25k, rather than Oracle 5's 1M

5) Stored procedures yet again - a near perfect description of their security mechanism.

6) Sounds very much like the Sybase record structure with varchar handling

I guess Oracle never did understand the nuts and bolts of what they were doing - just very good a selling it and making money from it.

 

 

 

 

 

Oracle was hugely benefitted by their success with Relational Databases in the late 1980s. Though IBM did the System-R Research, it was Oracle who gave the first working product (on VM/CMS operating system of IBM). Later Oracle got the benefit when IBM put the stamp of approval on Relational Databases when DB2 was introduced. Alongwith client/server and mini machines (VAX; Unix to follow soon) here also Oracle was the beneficiary, not IBM! In reality, IBM speeded up the 'death of mainframe' by removing all third-party databases from the mainframe market place.

What happened in the 1990s and later is now history.  Oracle was able to play the relational database 'syndrome' better than anyone else - Ingres, Sybase, CA-DB/VAX or Informix. Ingres and Sybase were always superior products, technically.

Oracle's success in manipulating the database world may not work in the Javaworld. Here Oracle is fighting against 'Open Source' product line. Doubts about Java's openness will speed up the 'After Java what?' debate.  If IBM had filed a lawsuit against Oracle in the 1980s regarding the use of 'SQL' what would have happened?

The technical 'JavaWorld' is least interested in the laws, patents etc. Many doen't even understand all these legal jargons.  I think continuation of this lawsuit will definitely give rise to a new language G# as many have pointed out already.

PS. Caldera which inherited DR-DOS challenged Microsoft in 1996 and got substantial  damages. Something which legendary Gary Kildall, as a gentleman, never did!

"preloading ... through memory space cloning of a master runtime" ... gee, that sounds a lot like Unix fork() to me.

"Unfortunately, as a defensive strategy, Oracle has been forced to protect itself by selectively applying for patents which will present the best opportunities for cross-licensing between Oracle and other companies who may allege patent infringement.
Oracle Corporation Patent Policy

http://www.nosoftwarepatents.com/en/m/basics/coldwar.html

 

Do you mean result?