Building and using MapReduce in Ruby
updated almost 5 years ago; latest suggestion almost 5 years ago
This proposal has been withdrawn...
I built an implementation of MapReduce in Ruby. I could talk a little bit about:
- What MapReduce is, and why it's useful
- Writing and testing software with dRB
- Writing a non-web-based project (a library, and a supporting command line tool) in an outside-in GOOS style
- Why you shouldn't use my cobbled-together-on-a-few-days-off implementation, why you should look at Apache Hadoop instead, and how you can talk to it with Ruby.
I tend to agree with all of James's points, it's a bit of a grab-bag of unrelated stuff at present. My original motivation for writing this was to (a) get to grips with writing distributed systems well, and (b) practice working in a strictly outside-in BDD style, so perhaps it would make sense to limit the scope to something like "Writing distributed systems the GOOS way" , coviering a mixture of points two and three over there <----<
I think there could be a good presentation here, but it's not clear at the moment to what extent you want to talk about your specific implementation, or if you'd prefer to use it as more of a simple framing device for some of the other topics.
Which part of what you've written in the proposal do you think is the most interesting/exciting/intriguing?