The 3rd day of RubyKaigi 2009 (Afternoon)

All sessions of RubyKaigi 2009 is over.
Above photo is a novelty souvenir at this conference. This is a hand-held fan (in Japanese : 扇子).

In this afternoon, I attended these sessions:

• N-ways to distribute your Ruby desktop apps
  
  • speaker: Yutaka Hara
  • We use ruby, but almost as creating Web app
  • Why don't we use as creating desktop app?
  • Difficulty of distributing
    
    • It's not always true that user have installed Ruby interpreter
    • What packages do I make for?
    • How package do I make on?
  • Past libraries
    
    • RubyScript2Exe: not work on 1.8 heigher
    • Exerb: its author already doesn't use it
    • Ocra: for Windows only
    • Crate: for UNIX/Linux only
  • Why don't we become not to use as creating desktop app
    
    • GUI toolkit users are divided to there likings
    • No GUI toolkit users don't know existing GUI toolkit itself
  • Try to create and use Web app as Desktop app
    
    • Zipping application codes and local server process and distributing these
    • Data exist on local, so you can backup as you like
    • But many web apps consume local ports
  • I made RubyStation for this purpose
    
    • manage gems
    • manage local posts
    • manage lifecycles
  • If you use it, (n=1) Q.E.D. :-)
• Journey through a pointy forest: XML parsing in Ruby
  
  • speaker: Aaron Patterson
  • XML is from many nodes: elements, attributes and so on
  • parsers
    
    • SAX
      
      • REXML: cryptic APIs
      • libxml-ruby: cryptic APIs
      • nokogiri: lucid APIs
    • Push
      
      • nokogiri: only one
    • Pull
      
      • REXML: slowly
      • libxml-ruby: cryptic APIs
      • nokogiri: lucid APIs
    • DOM
      
      • REXML: slowly
      • libxml-ruby: cryptic APIs
      • Hpricot: CSS selector ready and lucid APIs
      • nokogiri: CSS selector ready and lucid APIs
  • Some syntaxes in CSS selector and XPath overlap. So nokogiti explicitly call which use CSS selector or XPath
  • nokogiri interprets XML namespace correctly, Hpricot somtimes not
  • When parsing HTML, nokogiri is closer to browser behavior than Hpricot
  • I hope that nokogiri will be merged into ruby-core in the future
• Re-introduction to Ruby
  
  • speaker: Shugo Maeda
  • Before the theme, Information of RubyWorld Conference: Please come over
  • Once upon a time, Ruby language system is the only one
  • Now, there are many ruby language systems
  • So we are trying to decide the specifications
    
    1. rubyspec, written in test specs
    2. standardization and documentation by IPA project
  • When Ruby 0.4.9:
    
    • Cannot omit the brackets
    • include is a reserved word
    • Documentation like specification exists!
  • Lexical scope in block is changed from Ruby 1.8 to 1.9
  • Block arguments behave like method arguments
  • But there are still unresolved problems
  • I try to suggest "explicit scope" as magic comment
• Keynote Address: Ruby, Nihon Ruby-no-Kai and me
  
  • speaker: Masayoshi Takahashi
  • This year's theme is "CHANGE"
  • Last year's one is "Diversity"
  • There are many Ruby language systems, and there are various Rubyists
  • Now is the time to change by / to change for Ruby and Rubyist
  • I had start relating to Ruby from 1997
    
    • version 1.1
    • only source distribution
  • My first post to mail list is in 1998
  • I'm not a commiter and I never send some patches, I'm just a user
  • Instead, I have written ruby books, have held ruby events, and have established Nihon Ruby-no-Kai
  • Problem has occurred recently: wishes of each other aren't accomplished at the same time
  • Accept changing if somethings are thrown away
  • At past time, Ruby was written by Matz only
    
    • Things are what Matz wrote
    • Speed of development is Matz' speed
  • Number of commiters and maintainers had grown up
    
    • Unification inside Ruby code is decreased
    • Delay as consensus delay
  • From Ruby 1.6, Ruby was known for Non-Japanese users
    
    • by "Pickaxe" book (authors: Dave Thomas and Andy Hunt)
    • But communities between Japanese and Non-Japanese had grown apart because each natural languages are different
  • From Ruby 1.8, Ruby was spread more by RoR
    
    • The level of requirement has been sophisticated
    • But cultures between Rubyist and RoR users have grown apart more
  • Should we hope not to change?
  • No, we don't be afraid of change
  • We should select type of changes and control changes
  • Nihon Ruby-no-Kai's activity : various
    
    • RubyKaigi
    • Sponsor of Regional RubuKaigi
    • Rubyist Magazine
    • Ruby Reference Manual
    • Hub of Japanese regional communities
  • Nihon Ruby-no-Kai's problems
    
    • Much less diversity
    • Activities are very slow
    • Lack reformation resources
    • Much less disclosure
  • Bottleneck is me becouse Nihon Ruby-no-Kai adopts "Benevolent Dictator" model
  • The solution isn't fixed yet
  • Granting right to other members e.g. representatives from each activities
  • Making or using support tools for each activities more
  • Reform Nihon Ruby-no-Kai acting without me
• Closing

N.B. These summaries may be imprecise.

The last keynote is hard to describe for me, it's very emotional. It's special experience for the people who were there.

Finally, I say thanks to all staffs, speakers and attendees.