Everyone knows that Oracle is suing Google over claims that Google used Sun's Java technology without appropriate licenses in the Android mobile OS. Now that Oracle owns Sun's technology, it wants to be paid for those licenses.
Google claims Android does not use actual Java intellectual property and that the claims are baseless. That's all fine, the stuff of usual Silicon Valley licensing battles. But what's not fine is what's in the small print of Oracle's actual lawsuit. What Oracle is saying and doing should scare everyone.
For example, one of Oracle's attorneys is David Boies. We know that name from the recent action that overturned the gay marriage ban in California and from the old DOJ antitrust actions against Microsoft. He also advised the Recording Industry Association of America in its file-sharing case against Napster and represented former vice president Al Gore in the disputed 2004 U.S. election results. Nobody hires David Boies to litigate unless they are serious.
There are the seven alleged counts of patent infringement, and all are about software process patents, a controversial type of patent that I believe never should been granted in the first place:
- Protection Domains to Provide Security in a Computer System (2000)
- Controlling Access to a Resource (2000)
- Method and Apparatus for Preprocessing and Packaging Class Files (1999)
- System and Method for Dynamic Preloading of Classes through Memory Space Cloning of a Master Runtime System Process (2008)
- Method and Apparatus for Resolving Data References in Generate Code (2003)
- Interpreting Functions Utilizing a Hybrid of Virtual and Native Machine Instructions (2005)
- Method and System for Performing Static Initialization (2000)
It was my hope that the recent Bilski ruling would have invalidated all software process patents at one stroke, but the U.S. Supreme Court chose not to go that far. But think about these claims: Does any one of them sound novel for their time? (Being nonobvious -- or novel -- is a basic requirement of gaining a patent, and one way of gauging that is whether there is prior art or previous products and services that work the same way.) Yes, I know that the U.S. Patent and Trademark Office approved them, but was anyone who knew about the prior art and what was obvious to an average practitioner in software engineering paying attention at the time?
There is also a count of copyright infringement. I think that is likely to be dismissed early in the proceedings, if the Android team did (as reported) a clean-room reverse-engineering of the Java virtual machine to create the Dalvik VM underlying Android -- and can offer proof. Phoenix Technologies managed that trick with the IBM PC BIOS, so it's quite possible that Google did as well with Java.