Snap! to WMR - WMR

WebMapReduce (WMR) is a simplified web-based interface for Hadoop that supports a choice of programming languages, implemented by Pat Garrity and Tim Yates. The idea for the software came from a Berkeley poster at SIGCSE 2008 (Portland) that reported on their work to integrate Hadoop into a Scheme interpreter used by student in their introductory course.

Milestones:
 * October 2008 - Dick Brown proposes a St. Olaf implementation of a Scheme interface to Hadoop to Pat and Tim, then first-year students in an accelerated introductory course.  The target platform was a [www.cs.stolaf.edu/wiki MediaWiki modified to accept Scheme] as a language for authoring pages, which was then our primary teaching environment for introductory CS (CS1).  This project used Hadoop's Streaming interface, which enables arbitrary executables to serve as mappers and reducers.  Linux scripts were wrapped around student-provided Scheme code to create those executables.
 * September 2009 - [csinparallel.org CSinParallel] grant received, which proposed creating WMR as a "standalone" (i.e., not integrated into a programming-language environment) web application enabling introductory students to submit Hadoop jobs using the programming languages taught in their courses. Students indicate (in a web form) their choice of programming language, provide code for mapper and reducer functions (or methods) in that language, provide source data, and submit;  results delivered via the web.
 * December 2009 - Dick class-tests an early Python version of WMR in his pilot CS for Scientists and Mathematicians introductory course (the first non-Scheme introductory course to be offered at St. Olaf).
 * March 2010 - Tim and Pat demonstrate WMR (with multiple programming languages) at the Intel vendor booth at SIGCSE 2010 to dozens of CS educators. Dick, Libby Shoop, Pat, and Tim offer a first CSinParallel workshop, featuring WMR -- a year ahead of the grant dissemination schedule.
 * March 2011 - Tim and Pat deliver the SIGCSE paper on WMR, and demonstrate a Django-based WMR.
 * Summer 2012 - Stephen Lee resumes testing and distribution for the Django-based WMR, after graduation of Tim and Pat. Dick, Stephen and other St. Olaf cluster managers collaborate with Giovanni and others to create a Snap! language interface for WMR.