Implementing the virtual method pattern in C#, Part One

re: Implementing the virtual method pattern in C#, Part One

Ben Voigt — Wed, 30 Mar 2011 17:33:04 GMT

@Peter, @Gabe: Why "C-flat" aka "B"? Why not "C"? It's a surprisingly good fit, all things considered.

re: Implementing the virtual method pattern in C#, Part One

Scott B. — Fri, 18 Mar 2011 19:43:28 GMT

This "explicit-this" stuff looks like "explicit-self" in Python :)

For the sake of completeness, with:

class Dog:

def make_noise(self):

return "yip"

Calling "s = dog.make_noise()" is exactly equivalent to "s = Dog.make_noise(dog)".

re: Implementing the virtual method pattern in C#, Part One

Alex Corrado — Fri, 18 Mar 2011 10:41:43 GMT

Every time I put "virtual" before a method declaration I cringe at all the indirection. Thanks for reminding us that nothing is really magic.

re: Implementing the virtual method pattern in C#, Part One

Peter — Thu, 17 Mar 2011 22:43:04 GMT

@Gabe: That's enharmonically equivalent to B, which is more than 40 years old (details are available at en.wikipedia.org/.../B_(programming_language) for those who are unfamiliar with it). More of a step backwards than I think Eric intended! (See also en.wikipedia.org/.../Enharmonic)

re: Implementing the virtual method pattern in C#, Part One

Neal Gafter — Thu, 17 Mar 2011 18:33:56 GMT

This implementation does not handle the Basenji.

re: Implementing the virtual method pattern in C#, Part One

Keith J. Farmer — Thu, 17 Mar 2011 18:20:24 GMT

The explicit-this implementation should be familiar to Pythonistas.

re: Implementing the virtual method pattern in C#, Part One

Brian Hartung — Thu, 17 Mar 2011 17:11:51 GMT

For anyone interested in an under-the-hood perspective, I think Chapter 5 of the Rotor Book (freely available on Joel Pobar's blog: callvirt.net/.../Shared%20Source%20CLI%202.0%20Internals.pdf) does a nice job of explaining (at least the SSCLI's implementation of) method dispatch all the way down to the level of the execution engine (stubs, generic resolvers, and slot maps, oh my!). It includes shiny pictures and everything...(ok, there's nothing shiny about the pictures, I was simply trying to appeal to the attention deficit crowd--plus I'm dumb as toast).

re: Implementing the virtual method pattern in C#, Part One

Joren — Thu, 17 Mar 2011 16:38:45 GMT

Here's hoping for a Part X: Interfaces somewhere in the future. :)

re: Implementing the virtual method pattern in C#, Part One

Aakash Mehendale — Thu, 17 Mar 2011 15:21:35 GMT

Framework Design Guidelines-inspired nitpick: SoreThroat should be called HasSoreThroat :)

re: Implementing the virtual method pattern in C#, Part One

NickLarsen — Thu, 17 Mar 2011 15:15:16 GMT

An aside from this, I just realized that extension methods are not as "broken" as I once thought, assuming they are guaranteed to make an unchecked call to a member of "_this" somewhere in their execution.