There is a bit of a furor right now for the judge that ordered Google to hand over logging data, identifying usernames and ip addresses associated with video views.

I decided to take a look at the judge’s ruling, and there were some very interesting discoveries. Viacom asked for, in this order:

• The code that powers google.com to ensure that “defendants did not modify the search tool to more easily find infringing material.”
• The code that powers youtube.com, for the same reason
• As well they need the code to ensure that defendants did all that could be done to protect against infringement.
• The source code to the “Video ID” project, which compares incoming videos to “fingerprints” of videos provided by copyright holders. (Explanation a bit further on)
• Copies of all videos that were removed
• And finally, the IP Addresses and Usernames logging data associated with videos.

Take a look at that list. Viacom essentially fired a massive cannon at Google, asking for all their internal trade secrets, and as noted by the judge, they never promised they wouldn’t share those secrets. In addition, Viacom asked for the advertising database schema!

Pretty much, this ties in with what I said before about these broadcasts companies becoming scared of the many-to-many patterns we are seeing.

Viacom did whats called setting up the shot. They asked for extremely irrational things, low prices if you will, and if it had worked, they would have been extremely well off. Then, they sneak in a much less irrational request, one that Google’s lawyers fell for, which is what Viacom wanted all along. In comparison to the incredibly irrational requests, the logging data seemed eminently reasonable.

It was almost a win-win process, so long as Google fell for something or made a mistake. And they did.

Luckily the EFF are already reacting, citing the Video Privacy Act.

I sincerely Google provides the data in paper format.

Dolphins, in Kelowna, big dolphins statue, Google does not, and likely will not for a long time understand directions by landmark, indicates lack of cultural context.

Happy Belated Canada Day to everyone.

I discovered a major limitation to a service for everyone, like Google, yesterday(July 1st). I was asked to meet a group of friends at the “dolphins”. I asked where that was, and they gave a firmer location.

Before I asked, however, I had done several google searches, including local searches. None of the results were anywhere near where the event was supposed to take place. This information was nonsensical to me, as any form of local slang or landmarks is to someone new to the town.

As I approached the meeting point, I understood their meaning: they were in front of a very large, very hard to miss, dolphin statue.

But this illustrates a fundamental gap in automated services like Google: a lack of cultural context. There is no way to tell Google that this may be a locally relevant meaning, vs say, the Dolphin Hotel.

My gut feeling on the matter is that without active participation to tag and add cultural context to locations, that there will never be any way to add this information. As well, the people that know the contextual information are more likely to be people that don’t care, and already know what it means. Why would they add the information?

Then again, that is what seemingly confuses people about Wikipedia, about why would anyone contribute?

But in an increasingly network-centric world, this kind of data, this locally/culturally relevant data about locations and landmarks is important, and extremely valuable.