LA RubyConf 2013

gistfile1.txt

Refactoring Fat Models with Patterns: (http://www.larubyconf.com/proposals/43)
==============================================================================
@brynary
  @codeclimate
  codeclimate.com
http://blog.codeclimate.com/blog/2012/10/17/7-ways-to-decompose-fat-activerecord-models/
github.com/codeclimate/refactoring-fat-models
https://github.com/solnic/virtus

1 value objects     # data clumps, attribute logic. if lot of methods have the same prefix 
2 service objects   # short lifecycle, coordinating multiple models, external service interaction.
3 form objects      # one form, multiple models. create and update flows. use the Virtus gem.
4 query objects     # first class objects encourage refactoring
5 view objects      # .. (read about the two-step view pattern)
6 policy objects    # use if you have complex reads, ancillary reads. whatever that means.
7 decorators        # yeah.


Impressive Ruby Productivity with Vim and Tmux: (http://www.larubyconf.com/proposals/24)
========================================================================================
brew install tmux
https://github.com/spastorino/vinimum  # vim plugins
http://goo.gl/98m56


Dude where is my scalable API? (http://www.larubyconf.com/proposals/62)
=======================================================================
@eljuanchosf
http://kr.github.com/beanstalkd       # Fast job queueing - Beanstalk
Varnish-link                          # request caching
Ilya Grigorik (www.igvita.com/)       # Highly recommended blog


How to make the most of a conference (http://www.larubyconf.com/proposals/84)
=============================================================================
Corey Haines, Corey Haines. @coreyhaines
  http://www.MercuryApp.com
  http://cleancoders.com
  http://hugboard.com
  http://coderetreat.org

The Pragmatic Programmer by Dave Thomas
Smalltalk Dave Thomas
Jim Weirick


Python for Ruby (http://www.larubyconf.com/proposals/23)
=============================================================================
@panopticdev
@facebiff
panopticdev.com
leone.panopticdev.com

Arrays are Lists
Hashes are Dicts
map(function,list)

how python is different:
------------------------
no blocks. but has function decorators which are similar to blocks.
first-class functions - functions as variables.
one-line lambdas by default.
[1,2,3] # a list
(1,2,3) # a tuple - immutable
from urllib2 import urlopen   # select what you import from a lib, 'explicit is better than implicit'
false, None, [],(),{},'',0 all evaluate to false
doesn't have enumerables, has built-in functions # not as convenient as Enumerable
no case statements
no module mixins.
can't open a class and extend it. can't extend built-in types.
numpy, scipy # widely used scientific calculation libs, that ruby doesn't really have
python on windows is generally easier
community is bigger

"harder to write code that pisses of other developers in python"


Backbone.js, Jasmine and Rails: The Lust Story (http://www.larubyconf.com/proposals/8)
======================================================================================
.


Why I like JRuby (http://www.larubyconf.com/proposals/80)
=========================================================
@msfionatay, pivotal

http://github.com/GreenplumChorus/chorus 
around since 2002, traction in prods since 2005

good:
java libraries: NLP
actionmailer doesn't provide fine grained control. JavaMail allegedly better
directly call java code from ruby
other languages built on jvm: scala, clojure
might be ruby gems that wrap the java libs if you are lucky
leverage existing java deployment infrastructure

bad:
gem availability. gems with c-extensions don't work with j-ruby
long app startup time. inhibits TDD. difficult to set up Spork with jruby

ugly:
concurrency. jruby has real threads. mri has threads with global interpreter lock.
- non-thread safe libraries results in bugs - rails, airbrake.


It's not your test framework, it's you (http://www.larubyconf.com/proposals/74)
===============================================================================
jbehave              # first BDD tool out there
rbehave              # ruby
story runner /rspec  
gherkin              # state diagrams
  given (start state)
  when (event)
  then (state transition)

summary
  fight pain
  reveal intent


Recommendation Systems Using Ruby (http://www.larubyconf.com/proposals/54)
==========================================================================
@rweald
@sharethrough
ryan@sharethrough.com

two main categories:
  1. collaborative Filtering (nearest neighbor)
    memory based
      similarity functions
        Pearson Correlation Coefficient
        Cosine Similarity
    model based
  2. content based - classification
    k-means clustering
  
  3. hybrid of 1. and 2.
    either combine them, or feed one into each other

evaluate recommendation quality
  precision vs recall           # textbook
  clicks and click through rate # actual data
  thumbs up thumbs down         # user feedback

apache mahout
jruby mahout gem
sciruby
recommenderlab for R

recommender systes: an introduction
:amazon.com recommendation: item to item collaborative filtering"
grouplens an open architecture..


Keynote, Ron Evans
==================
@hybrid_group

https://github.com/hybridgroup/artoo
http://www.gosphero.com/