Google had pushed for a fall trial date at the earliest but Oracle argued vigorously (and successfully) for an immediate date on grounds that the case was ready with Google only playing for time. To induce the judge to give it an early trial date, Oracle offered to dismiss with prejudice all patents that had been rejected in a final office action by the USPTO subject to reinstatement in the event the PTO reversed itself prior to the start of trial. Trial started on April 16 and, lo and behold, the PTO did reverse itself concerning one significant patent (the '702 patent) such that Oracle would normally have been able to pursue its claims for infringement based on that patent. However, it did so a few days after the trial had started. As a result, that claim wound up being finally dismissed for purposes of this lawsuit. Oracle tried to renege and pushed the judge to reinstate the claim but the judge said no (decision here: (http://www.scribd.com/doc/91307218/Oracle-denied-using-reviv...).
Thus, Oracle screwed itself as a result of having charged full speed with tomahawk swinging wildly in the air. In pushing aggressively for tactical advantage, it essentially threw away the one patent claim that had any potential for viability (the two patent claims on which it lost at trial were in fact relatively weak, with one of the two patents about to go down before the USPTO and set in any event to expire in December, among other things).
Oracle now finds itself in a deep hole. It won a few crumbs in the copyright phase and got skunked on the patent phase. More copyright drama to follow as the judge rules on API copyrightability. Oracle might win on that issue, as it is a tricky one under Ninth Circuit law (see http://news.ycombinator.com/item?id=3980642), but it then faces a retrial in which the judge would give much more refined instructions on what it means to infringe the SSO of the Java APIs than he gave in the first trial - between that and the (I think, formidable) fair use defense that Google has, it is strictly an uphill fight for Oracle from here on out. (By the way, Google has done a masterful job of arguing the issue of why APIs should not be copyrightable and, for those inclined to read through a superb legal brief on the issues, here it is: http://www.groklaw.net/pdf3/OraGoogle-1137.pdf).
Oracle may yet rally if a lot of things go its way. I wouldn't bet on it though. Much more likely, in my view, is that the case becomes a testament to what happens when a party makes a high-stakes opportunistic legal grab that goes badly awry. What Larry Ellison set out to justify as Oracle's vindication has instead become Oracle's folly.