undefined | Better HN

0 pointsmikeash13y ago0 comments

It seems to me that it's a cultural problem more than anything. People know that they shouldn't take an existing, well-known method name and repurpose it for something completely different. But somehow a lot of people think it's OK to repurpose an operator to mean something completely different just because it's convenient. C++ being the classic example. << means bitshift. Just because it looks like arrows doesn't mean it's sane to repurpose it into a completely different "output this" operator. You wouldn't override "leftShift" to output stuff, so why do people do it with <<?

I think it depends a lot on the languages. Operator overloading is easy and common in Python, but in my (limited) experience, it's used to create new classes which respond to existing operators with existing semantics, e.g. to make new number classes. For this, it's great. There's no particular reason the types built-in to the language should get special treatment for operators, as long as programmers can resist the temptation to be idiots.

0 comments

pcote13y ago

>Operator overloading is easy and common in Python, but in my (limited) experience, it's used to create new classes which respond to existing operators with existing semantics, e.g. to make new number classes.

In Python, it's called "special method attributes" and they can do a lot more than just create classes. The __eq__ attribute returns a straight boolean for example in response to the "==" operator. And yes, by the way, you actually can accidentally alter attributes in the calling and callee objects if you aren't careful.

mikeashOP13y ago

Sorry if I was unclear, but I meant that this facility lets you create classes whose instances work like built-in numbers, not that the facility itself is used to create clases somehow.

stelonix13y ago

"But somehow a lot of people think it's OK to repurpose an operator to mean something completely different just because it's convenient. C++ being the classic example."

Do you have any data to prove such assertions? A C++ STL operator chosen more than a decade ago is not enough statistical data for one to have such a narrow-minded vision and say these bold statements. Maybe you need to check all the nice C++ libraries that make use of operator overloading.

tragomaskhalos13y ago

There's basic language bigotry going on here: an aged internet meme declares that C++ operator overloading is bad because you can indeed, in theory, do idiotic counter-intuitive things with it. Yet the identical capability in Ruby, Scala and Haskell (inter alia) is deemed A Good Thing by the consensus.

ScottBurson13y ago

I think it's an interesting phenomenon in the sociology of programming languages.

A powerful new feature is introduced in a niche language, and novices abuse it. The feature gets a bad reputation. Yet eventually, as new languages adopt it, the community learns how to use it and how not to use it, and people become accustomed to it, and come to agree that it's not such a bad thing.

I'd say that's happened, to some degree, with garbage collection and lambda expressions. We seem to be in the middle of the process with operator overloading, and still in the early stages with Lisp-style macros (I can hope, anyway :-).

ethzero13y ago

It's not an identical capability. Scala doesn't technically have operator overloading, it has symbolic method names and operator syntax for methods. This means you can use any valid name as an operator.

With C++, you can only overload the built-in symbolic operators. This means instead of choosing the best symbolic name to use as an operator you're forced to re-purpose some built-in operator like "<<". In Scala you can pick any name or symbol you want.

2 more replies

kzrdude13y ago

You can't overload operators in Haskell. You can define new operators though. You're not going to change the meaning of + unless you remove the old.

mikeashOP13y ago

For me, and perhaps others, this is because C++ enshrines idiotic counter-intuitive operator overloading in its standard library.

1 more reply

pjmlp13y ago

The common example people pull off the hat with C++ is the use of << operator for streams and some UI toolkits.

Which I personally never had a problem with.

mikeashOP13y ago

You seem to have confused the phrase "classic example" with "the worst thing out there, with data to prove it".

It's an example. Do you dispute that? It's classic. That's subjective, and I claim it by raw assertion. That's it.

j / k navigate · click thread line to collapse

0 pointsmikeash13y ago0 comments

0 comments

pcote13y ago

mikeashOP13y ago

Sorry if I was unclear, but I meant that this facility lets you create classes whose instances work like built-in numbers, not that the facility itself is used to create clases somehow.

stelonix13y ago

"But somehow a lot of people think it's OK to repurpose an operator to mean something completely different just because it's convenient. C++ being the classic example."

tragomaskhalos13y ago

ScottBurson13y ago

I think it's an interesting phenomenon in the sociology of programming languages.

ethzero13y ago

2 more replies

kzrdude13y ago

You can't overload operators in Haskell. You can define new operators though. You're not going to change the meaning of + unless you remove the old.

mikeashOP13y ago

For me, and perhaps others, this is because C++ enshrines idiotic counter-intuitive operator overloading in its standard library.

1 more reply

pjmlp13y ago

The common example people pull off the hat with C++ is the use of << operator for streams and some UI toolkits.

Which I personally never had a problem with.

mikeashOP13y ago

You seem to have confused the phrase "classic example" with "the worst thing out there, with data to prove it".

It's an example. Do you dispute that? It's classic. That's subjective, and I claim it by raw assertion. That's it.

j / k navigate · click thread line to collapse