I am a big fan of open software. Meaning, Open Source and the thought and philosophy behind it. And I also think that in the long run, things based on open standards with open discussion and open algorithms are the infrastructure that allows us to move forward. Private, proprietary companies might advance the state of the art at any given time, but in the long run, it’s the open, publicly documented things that hang around.
So today’s talk by Jimbo Wales at the NYU Free Culture Club (talking openly about Wikia for the first time) begs the question, how can open algorithms square with the “better” results gained from having full search histories? How do you anonymize? Can you, in an open space?
Howard Greenstein was there and paraphrases:
Question on the open corpus of search info from Wikia – even though it will be somewhat anonymized, even some sets of data that were supposedly anonymous have been shown to be able to be analyzed (implying the AOL leak).
They want to be careful on this – very tough question for Jimmy.
Can we deal with the identity and privacy of the end users while providing the ‘best possible’ search engine? Are those two things actually opposites?
Another quote explaining Jimbo’s vision…
Wikia – all the algorithms will be open source. Free search – as in transparent, testable, researchable. All the good research has been behind the walls of commercial companies (like Google, Microsoft, etc.) No place for computer scientists to go to do such research.
Want it to be:
– Participatory – bring best elements of Wikipedia to the problem of search
– Open
– Democratic
Tags: greenstein - jimbowales - nyu - wikia
View blog reactions
Subscribe