Summary

Kevlin Henney takes a philosophical approach to encapsulation, polymorphism and inheritance, and explains what it means to write Java programs according to his view on OOP.

Kevlin is an independent consultant and trainer based in the UK. He has been a columnist for various magazines and web sites, including Better Software, The Register, Application Development Advisor, Java Report and the C/C++ Users Journal, co-author of A Pattern Language for Distributed Computing and On Patterns and Pattern Languages, and editor of the 97 Things Every Programmer Should Know.

About the conference

QCon is a conference that is organized by the community, for the community.The result is a high quality conference experience where a tremendous amount of attention and investment has gone into having the best content on the most important topics presented by the leaders in our community.QCon is designed with the technical depth and enterprise focus of interest to technical team leads, architects, and project managers.

Oct 05, 2011

• Wonderful presentation

  by Adam Nemeth / Oct 07, 2011 07:47

• Impl? No. Well... how about "not for long"?

  by J. B. Rainsberger / Oct 08, 2011 03:32

  • Re: Impl? No. Well... how about

    by bruce b / Oct 12, 2011 11:00

• Type of a presentation where...

  by Yuriy Zubarev / Oct 10, 2011 02:43

• Nothing

  by Allen Yu / Mar 04, 2015 03:43

• Wonderful presentation

  by Adam Nemeth / Oct 07, 2011 07:47

  I'm always surprised by Kevlin's broad literacy on programming subjects, and the ability to go from very-very high level concepts to actual code level consequences.
  
  This presentation shows that the SOLID principles of encapsulation, substitution and interface-dependency have deep inter-relatedness. It shows different aspects of these in virtually unrelated technologies, like C, C++, Java, COM and Javascript. It seems it recommends to shift the ADT taxonomy to interface inheritance in order to reach pure object-orientedness.
  
  What I miss is the expansion of the object/class concept from a perceptional view. I think it's a very interesting question of how the notion of classes sneaked in to the object-oriented paradigm, but perhaps this presentation simply wasn't about this.

  Back to top

• Impl? No. Well... how about "not for long"?

  by J. B. Rainsberger / Oct 08, 2011 03:32

  I prefer to interpret Kevlin's comment "Impl? No." a bit differently: if I don't yet know what makes a particular implementation of an interface different, then I will name the class XImpl; however this suffix "Impl" (like "Factory", "Singleton", "Manager", "Policy"...) signals a purely structural name which, while useful, restricts my understanding of the design. After some time, as I learn what makes this particular implementation significant within the context of the system, I rename XImpl to SignificantPropertyX, where "SignificantProperty" is an aspect of the implementation that matters in the current design. This corresponds quite nicely with Kevlin's example here. I don't know that "RandomAccessRecentlyUsedList" means more than "RecentlyUsedListImpl" when we don't yet know whether random access matters in the design.
  
  I have summarised an approach to improving names here: link.jbrains.ca/ovMMvz

  Back to top

• Type of a presentation where...

  by Yuriy Zubarev / Oct 10, 2011 02:43

  ... you learn couple of important things, and realize that you don't fully comprehend dozen of even more important ones.
  
  Commendable job Kevlin!

  Back to top

• Re: Impl? No. Well... how about

  by bruce b / Oct 12, 2011 11:00

  If the interface has one implementation, then you're going to be stuck with some generic naming convention: Impl, Default, Standard are all ones that I've seen. Besides if you follow his advice and the concrete implementation is only referenced with the new operator, then what's the big deal?

  Back to top

• Nothing

  by Allen Yu / Mar 04, 2015 03:43

  Nothing New in this talk. A liitle disappointed

  Back to top