MS “Gem­ini:” I had heard about Gem­ini before. At the BI Con­fer­ence Keynote last year Gem­ini was demoed — a user with a lap­top run­ning Excel with 100 mil­lion rows of data. My mis­taken under­stand­ing at the time was that this clever par­lor trick had a large, server-based in-memory data­base sys­tem net­work con­nected. Au con­traire Mon Ami! This was a more or less stan­dard 4GB lap­top run­ning Vista and Excel with an inter­est­ing Excel plug in (“Gem­ini”)! The magic is colum­nar com­pres­sion — the cat­e­gory of tech­nol­ogy also used by the log file tool ven­dors use for fast log file analy­sis. When a data set is imported into Gem­ini this colum­nar com­pres­sion occurs, but after that, all the pro­cess­ing is on the lap­top (MS also demon­strated Gem­ini on a Net­book, admit­ting you prob­a­bly couldn’t get 100M rows on a 1GB machine, but you could get 20M!). Gem­ini also remem­bers where all the data came from with XML meta­data, solv­ing the tra­di­tional prob­lem of ad hoc spread­sheets when the cre­ator is no longer around (“where the hell did this data come from?!!”). The Gem­ini team is in the under-commit mode for now (smart) but I think this kind of tool could have a pro­found impact on busi­ness ana­lyt­ics over time.