Comments on: More Discussion On Operator Overloading

By: Aviad Ben Dov

Aviad Ben Dov — Mon, 10 Dec 2007 12:25:36 +0000

Oh okay – it makes sense now.

By: Daniel

Daniel — Sat, 08 Dec 2007 18:43:46 +0000

You’re concept works well with the “[]” and “[]=” operators, but it doesn’t suite algebraic operations. So, for what you suggested it for, it is a good approach, but for what I need operators for it doesn’t work.

Hmm, I can see my sentence could be misleading. I’ll re-word it.

By: Aviad Ben Dov

Aviad Ben Dov — Sat, 08 Dec 2007 08:54:35 +0000

Also, can you explain which use cases my suggestion breaks?

By: Daniel

Daniel — Wed, 05 Dec 2007 22:06:33 +0000

Sorry, I’ll correct the spelling of your name…

As for the reduction in abuse, someone is less likely to implement a method called “plus” that doesn’t fit the semantics of “plus”. There was an argument that someone is more likely to invent new semantics for a symbol (such as “+”), than for a word (such as “plus”).

By: Aviad Ben Dov

Aviad Ben Dov — Wed, 05 Dec 2007 20:03:57 +0000

First off, my name is Aviad. The I is before the A.

Second, I completely agree about the expressiveness required when implementing a solution for primitive-like types (Complex number and a Matrix are the easiest examples).

However, regarding your solution of always replacing the operators with method calls – how does that reduce abuse?