By Nate Anderson
July 03, 2008 - 01:15PM CT

As part of the discovery process in its $1 billion lawsuit against YouTube, Viacom asked for an astonishing array of information: the source code for the search functions that power Google and YouTube, the source code for YouTube's new "Video ID" program, a complete set of every video ever removed from the site, databases containing information on every video ever hosted at YouTube, and a copy of every private video.

The judge has now ruled, saying no, no, yes, yes, and no. Even with the limited discovery, some poor drone at YouTube can now look forward to weeks of copying information onto Viacom-provided hard drives. Search source code - Source code disclosure for search functionality was an easy call to make. Google insists that its algorithms are the product of more than "a thousand person-years of work" and remain a vital trade secret, and the judge agreed.

Viacom is trolling for evidence that Google/YouTube purposefully modified their algorithms to give preference to infringing work on YouTube, work that would apparently be more attractive to site visitors than videos of a two-year old snorting peas up his nose. The judge notes, though, that Viacom has no evidence for the belief and says that "YouTube and Google should not be made to place this vital asset in hazard merely to allay speculation." Google 1, Viacom 0.

Video ID source code - YouTube's recent Video ID system receives similar analysis. The code (the product of a mere 50,000 worker hours of engineering time) scans incoming videos against a database of video fingerprints; matches are flagged for human follow-up. Viacom wants the code so that it can show YouTube what it "could be doing" to stop infringement; in other words, Viacom will offer engineering advice to YouTube on how to design its systems. Ruling?