Plus Sign Overloading

Plus-sign overloading is a problem in my opinion. It is difficult to tell by looking at the expression alone how it is going to handle various situations. And, each language that uses it seems to have different rules. I say dispense with overloading and have a dedicated string concatenation character.

As far as what the concatenation operation should be, I prefer "&" over "." because "." is best used for object/array dividers. But some languages use "&" for dereferencing.

I didn't mind the use of `++` in Haskell. In my domain I use string catenation far more often than arithmetic addition. Thus, I'd rather have a single character.

Ah. By that reasoning, I suggest you use `(`. That neatly avoids the PlusSignOverloading problem and simultaneously provides you with a one-character solution to your string(catenation(requirements.

String concatenation in Haskell is a grossly expensive operation, unfortunately. I learned that lesson really fast when writing the prototype for CUT 3. However, expressing strings as large trees of extents made things go much faster -- there, I'd flatten the tree only when absolutely necessary, and amortized the cost of concatenation or printing. Perhaps, with that being said, the best solution to string concatenation isn't concatenation, but rather, composition of extents in lists? --SamuelFalvo?

Even if you've got ropes or other forms of high-efficiency logical concatenation, you'd do well to have an operator for(indicating(intent(to(concatenate.
See also OperatorOverloading, OverloadingCommaOperator
CategoryLanguageFeature CategoryCpp CategoryRuby

View edit of May 19, 2008 or FindPage with title or text search