张仲奇AdvancingDeclarativeQueryintheLongTailofScience.ppt

Advancing￿Declarative￿Query￿in￿the￿Long￿Tail￿of￿ScienceBulletin of IEEE Computer Society Technical Committee on Data EngineerAbstract•Long tail•Relational database•Delivery vector•SQLShare•Non-programmerLong tailP. Murray-Rust and J. Downing. Big science and long-tail science. http://blogs.ch.cam.ac.uk/pmr/2008/01/29/big-science-and-long-tail-science/, term attributed to Jim Downing.Abstract•Long tail•Relational database•Delivery vector•SQLShare•Non-programmerSummary•Compete with other lightweight languages•Increase collaborative data sharing•Reduce 9-to-1 ratio of time•Even non-programmer responding positively•A simplified interfaceIntroduction•9-to-1•SQLShare•Starter kit•Allow me to do science againMetagenomics￿as￿Set￿ManipulationIllustration of the steps in an algorithm to (a) identify the surface of proteins, (b) calculate various statistics, and (c) synthesize “stealth” molecules that could mimic the protein surfaces. The use of SQLShare provided opportunities to exchange “a 10 minute 100 line script for 1 line of SQL.”S. Tringe, et al. Comparative metagenomics of microbial communities.Science, 308(5721):554–7, 2005 Apr 22SQLShare￿System￿Details•No schema•Unifying Views and Tables•Incremental Upload•Tolerance for Structural Inconsistency•Metadata and Tagging•Append-only, Copy-on-Write•Simplified ViewsThanks。