Check out Phil Potter's somewhat more complete notes, at least for day 2/3 of the conj at: https://github.com/ppotter/clojure-conj-2011-notes.git

Also focus's blog post on the conj:

http://clojure.com/blog/2011/11/17/second-conj.html

Categories Key:

-Of Immediate (U)se

-(E)soteric

-(I)nteresting Implementation

Rating of topic/speaker: 1 (Meh) to 5 (Wow!)

Learning Clojure - Next Steps

STUART SIERRA

Categories: U,E

Rating: T4, S3

-Mostly a survey of some advanced Clojure concepts.

-Of particular interest:

--clojure.test.generative creates sets of test data, validates function against those sets given assertions provided (data type, etc), attempts to use all cores for a defined amount of time

Concurrent Stream Processing

DAVID MCNEIL

Categories: U

Rating: T4, S3

-revelytix, not yet available to the public.

-run large sets of data concurrently and in parallel via "streams"

-pipes are protocols that enqueue and dequeue the data streams, nodes are the functions operating on the data

-can split or join streams

-pipes chunk the data for performance

-input to a pipe provided by push, output is pulled by consumer

-fork/join thread pool is used for execution

Clojail

ANTHONY GRIMES

Categories: U

Rating: T5, S3

-Sandboxed Clojure environment, prevents dangerous side effects, i.e.:

--I/O (interaction with file system or internet)

--infinite loops

--rebinding

--execution of arbitrary programs

--destruction of jvm

-Hardest part of Clojure sandboxing is the state

-Looks at the various parts of the code

-Prevents def abuse (rebinding, memory usage)

-Creates timeouts for loops

-Also:

--Customized eval contexts

--Allows users to block anything (selective blacklisting)

ClojureScript

CHRIS HOUSER

Categories: U,E

Rating: T4, S4

-Deep dive into cljs's guts

-Top to bottom:

--Data

--Protocols

--Macros

--Compiler

Striving to Make Things Simple and Fast

PHIL BAGWELL

Categories: E,I

Rating: T4, S5

-Great talk about a lot of stuff I didn't fully understand about new data structures

Introduction to Logic Programming with Clojure

AMBROSE BONNAIRE-SERGEANT

Categories: E,I

Rating: T4, S3

-core.logic - (mathematical) logic programming through clojure

-speaker was fairly new to the concepts and wasn't able to answer a lot of the questions, but slides still had good information/examples of the non-deterministic nature of logic programming

-mentions The Reasoned Schemer for anyone interested in logic programming

Hacking the Human Genome using Clojure and Similarity Search

ARNOLDO JOSE MULLER-MOLINA

Categories: E,I

Rating: T5, S4

-extremely interesting topic, mapping/defining transcription factor generation when applied to genes.

-not a lot of code, but a very cool field where clojure is being used to predict how genes will react to different transcription factors and where those transcription factors will attach

Ousterhout's Dichotomy Isn't (Part 2 of the Simplicity/Power/Focus Series)

STUART HALLOWAY

Categories: E

Rating: T2, S4

-attempt to define power like Rich defined simple, but fell short in my opinion

(Neal's) Master Plan for Clojure Enterprise Mindshare Domination

NEAL FORD

Categories: E

Rating: T3, S4

-how to get clojure into corporations, not a problem at hotelicopter

Predicate Dispatch

DAVID NOLEN

Categories: E,I

Rating: T4, S4

-core.match and core.logic - pattern matching

-also mentions The Reasoned Schemer, the authors were in the audience

Extending JavaScript Libraries from ClojureScript

KEVIN LYNAGH

Categories: U,E

Rating: T3, S3

-some arguments for using clojurescript over javascript.

-would be very interested to hear T2's opinion on this talk

-makes a comparison of javascript vs clojurescript as the imperial system vs metric system

-namespaces and immutabiilty is safer and more extensible than globals and 'this'

From Linear to Incremental

CHRISTOPHE GRAND

Categories: E,I

Rating: T4, S3

-incremental computation of parse tree after edits are made

-lower cost

-use associativity to only compute what has changed and what comes after in the tree

-parsley is the implementation

Logs as Data

MARK MCGRANAGHAN

Categories: U,I

Rating: T4, S3

-everything as data

-"data via simple abstractions with a general library"

-generating events as a stream of what is happening, able to capture and record these events

-aggregate the streams of events which are just data structures rather than strings, easier to manipluate, destructure rather than parse, can apply normal data functions to them

-can alert or drop events that error during processing

-dates need a datatype?

-github.com/heroku/pulse

Modeling the world probabilistically using Bayesian networks in Clojure

CHAS EMERICK

Categories: E,I

Rating: T4, S3

-used for document analysis

-"sometimes logic isn't enough"

-handle multiple variables to generate a joint probability distribution, but they don't scale

-applications:

--process control

--decision systems

--prediction

--classification

-implementation: Raposo

Clojure and Android

DANIEL SOLANO GOMEZ

Categories: E,I

Rating: T3, S2

-not easy, not really supported

-conversion to correct package type necessary

-benefits to having clojure on android: dynamic development

-clojure.core is a space hog

-Book: "Decaffeinated Android"

Keynote

RICH HICKEY

Categories: U,E

Rating: T5, S4

-Language next steps, suggest watching this when it becomes available

-phil potter has extensive notes on this talk

Cascalog

NATHAN MARZ

Categories: U,E,I

Rating: T5, S3

-high level abstraction for Hadoop map-reduce

-Hadoop:

--high latency batch processing

--fault tolerant

--petabyte scale

-abstraction and composition

-examples were tweets and authors at twitter

-used for implicit joining of different datasets

-github.com/nathanmarz/cascalog-conj

Functional Data Structures

DANIEL SPIEWAK

Categories: E

Rating: T4, S5

-interesting talk on functional data structures, specifically in Scala

-sequential, associative

-looking for immutable, asymptotic performance and structural sharing

-a look at trees and performance

--2-3 finger tree

---constant: append, prepend, first, last

---log: insert

---linear: concat

---non-locality of reference, slower

--red-black tree (clojure's sorted-map)

---log: get, insert, update

---linear: intersect, union

--bitmapped vector trie (hickey/bagwell)

---constant: append, first, last, nth, update

---linear: concat, insert, prepend

---locality of reference, much quicker

---max depth in 32 bit is 7!

Macronomicon

MICHAEL FOGUS

Categories: U,E

Rating: T5, S3

-main point was avoid if at all possible, same as in training

-use cases:

--create binding forms

--control flow

--abstraction

--transformation

--optimization

Performance in the Wild

CRAIG ANDERA

Categories: U,E,I

Rating: T5, S3

-a look at performance

-performance as a continuous concern, not a last minute effort

-the traditional requests vs throughput graph misses out on latency

-a more useful model:

--distribution of latencies vs throughput for a given transaction mix

-Optimization loop:

--Benchmarking

--Analysis

--Recommend

--Optimize

-case study:

--clojure wasn't the problem

--database and logging were the main issues

-recommend watching this video

Programming Music with Overtone

SAM AARON

Categories: E,I

Rating: T5, S5

-communication with SuperCollider sound synthesization

-create way for musicians to interact simply with the program

-allow for markup of through symbols/patterns similar to music notation

-able to tell that the musician is badass by their performance, not just a guy with a computer

-monotone