If you see something interesting, most of the Strange Loop 2011 slides are on github.
Monday we get introduced to the massive main hall, for sessions involving all 900 attendees. Erik Meijer had some high level big data insight. We should probably try and get him down for WindyCityDb He stepped back from the SQL-NoSQL debate to look at how the two ideas are really similar and different. The basic difference comes down to which way the arrows point – relational has parts point at wholes, and ‘NoSQL’ has wholes point at (or contain) parts. The symmetry is comparable to a ‘duel’ in category theory. The duel of a thing is it’s ‘co’, so CoSQL might be a better term.
Heath Broders did an informative run-down of iOS testing frameworks. Unfortunately, I’m not actively doing iOS right now, but picking from such a large array of options is always daunting, so it’s nice to see good high level overviews like this.
Gerald Sussman started off by saying that everything here (at Strange Loop) is obsolete. One is often struck by how often some old CS paper seems relevant, or the latest fad is a reawakening of a trend from years ago. One wonders what unrecognized modern topic will be all the range in 30 years. I believe I’ve seen some of Sussman’s material before, but I have been known to read an academic paper from time to time. He did observer that memory and resources are no longer the problem – latency is. Meanwhile, evolving and maintaining programs is the primary cost. As an old Lisp (or rather Scheme) practitioner, he spoke highly of self modifying programs – certainly an out-of-favor technique. He then described a scheme of using stateless machines connecting statefull cells – which has the important property of monotonically increasing information – important for managing concurrency.