As widely reported by every tech site on the internet, last week Oracle (which recently acquired Sun Microsystems) sued Google for infringing upon a variety of software patents Sun obtained while developing the Java software platform.

For the facts, I can’t improve upon the fine commentary at Groklaw, CNet, and tech-specific sites like RedMonk. James Gosling, inventor of Java programming language, has even commented on it.

Two conclusions are inescapable:

  • Sun could have, but chose not to, sue over the same patents, likely (at least in part) to preserve goodwill with the developers who used the Java framework;
  • There’s a real chance that all of the patents are invalid under Bilski v. Kappos, since they represent, as most, mere ideas.

Thus, the lawsuit represents a substantial business risk not just for Google, but also for Oracle. Oracle runs not just a risk that they’ll lose the suit, but also a risk that they’ll be worse off than where they started by squandering much of Sun’s goodwill and invalidating patents so valuable that back in 2004 Microsoft paid Sun $900 million to settle an infringement suit over them.

The stakes are high.

So high that the case seems unlikely to settle, and thus may change patent law nationwide.

In some lawsuits, both sides generally agree from the onset that the case has merit. In the Deepwater Horizon spill, for example, there’s no doubt that BP is liable to someone for some amount — the questions are to whom and for how much, and in many ways the litigation is nothing more than a continuation of negotiation by other means.

In most cases, however, it takes some time before the parties come to similar understandings of the merits and value of the claim. That’s why so few cases settle before the close of discovery even though the vast majority eventually do settle.

The key point here, however, is that most cases eventually become negotiations by other means. In most cases, expert reports, summary judgments, motions in limine — sometimes even trials and appeals — are all just steps in that complicated dance, conceptually no different from a business negotiation.

A handful of cases, though, are anything but negotiation by other means; they are total war. The lawsuit will continue until all motions, trials, and appeals have been concluded and the sheriff has been called to enforce the judgment.

And so it seems to be the case with Oracle v. Google. Oracle isn’t stupid — even if they were, David Boies, their lead counsel on the case, isn’t stupid — and isn’t going to suddenly realize that the Bilski opinion effectively killed the precedent upholding the validity of software patents, leaving the question of their validity wide open. Oracle knew that going in. Oracle is also not going to suddenly realize that Sun had good reasons for not suing Google over these same patents. Oracle knew that going in.

In other words: Oracle has launched total war.

Same goes for Google: unless Oracle effectively abandons the case and offers to settle it for little more than the cost of defense — which is unlikely — then Google has too much to lose by settling. More importantly, Google has too much to gain by winning, namely the invalidation of Oracle’s (and potentially other companies’) patents.

Which is why this case — and the central question of whether or not Sun’s patents are valid — might make it all the way through up to the Court of Appeals for the Federal Circuit and possibly the Supreme Court. It’s no stretch to say that billions of dollars hang on the answer.

(Thankfully, we might get an answer sooner rather than later. The Northern District of California, where the suit was filed, pioneered the use of local patent rules to expedite patent suits, making patent infringement suits much quicker to litigate and to decide, even cases of this scale.)