Comments on: Extension Methods, Nulls, Namespaces and Precedence in C#

By: Kaveh Shahbazian

Kaveh Shahbazian — Wed, 09 Jul 2008 22:08:47 +0000

I have found even crazier uses for extension methods (http://hcoding.blogspot.com/2008/07/foop.html)!
I think separating the namespace for extension methods is enough for working safely with them.
And the current design of extension methods is just for being usable in C# 2; that static class is totally un-necessary.

By: Jason Olson

Jason Olson — Wed, 09 Jul 2008 18:20:29 +0000

I think the concerns behind “writing bad code” being enabled by extension methods is the same concern that arrived with the introduction of partial classes. Yes, they can be abused, but they are also tremendously useful.

There are two current reasons why I love extension methods: 1) enabling of fluent interfaces built on top of a “normal” canonical API and 2) a “poor man’s” .NET version of monkey-patching from the Ruby world. So, for unit testing frameworks, instead of having code like “Assert.Equal(foo, bar);”, I could have “foo.ShouldEqual(bar);” instead due to an extension method on object.

By: admin

admin — Wed, 09 Jul 2008 14:19:52 +0000

@Will,

I’m with you. In my viewpoint, “sugar” is pretty important. Programmers spend a lot of our days doing little tasks again and again — out of the box APIs often mean that we need to use 5 lines of code to do one line worth of work the ‘obvious’ way, and that we need 15 to do it the correct way.

Extension methods are a great way to codify correct practices and put them on our fingertips — so it’s easier to do things right than to do things wrong.

By: superjason

superjason — Wed, 09 Jul 2008 12:18:31 +0000

I wasn’t aware of these. Thanks for putting this together.

By: Will

Will — Wed, 09 Jul 2008 11:51:54 +0000

I was reserved at first, because I was concerned that people would use extension methods for business logic rather than sugar. I haven’t seen any examples of this kind of abuse yet, but I’m sure somebody is doing it somewhere.

Extension methods: Good for sugar, bad for logic.