expression templates

[Dmitry Olshansky]

On 31.03.2011 0:39, Daniel Gibson wrote:
> Am 30.03.2011 22:34, schrieb Dmitry Olshansky:
>> On 30.03.2011 2:14, so wrote:
>>> On Tue, 29 Mar 2011 22:56:10 +0300, dsimcha <dsimcha at yahoo.com> wrote:
>>>
>>> Occasionally i encounter this argument that operator overloading is
>>> bad thing when it is abused.
>>> I don't overload operators offensively myself, i use dot(vec, vec)
>>> cross(vec, vec) for example because there is not a suitable operator
>>> and these names suits much better.
>>>
>>> On the other hand i am not against languages being flexible, quite
>>> contrary i don't call it a language if it is not.
>>> OO is an impressive tool and we need tools like this for better
>>> libraries. You can't change a language for a long time but you can
>>> update a library in very short period of time.
>>>
>>> vec add(a, b) { return a.x - b.x }
>>> vec operator+(a, b) { return a.x - b.x }
>>>
>>> Is there a difference?
>>
>> The second one plays havoc with parsing :) It's not that it's such a big
>> problem but ..
>> BTW when does + denotes a difference?
>>
>
> (Probably) Never..
> I think his point is that using a bad operator-overload (overload the 
> "+" operator for subtracting) isn't worse than using a bad function 
> name (like "add" when you're really subtracting).
>

Thanks, I see now. Next time I'd better read the whole thread first...

-- 
Dmitry Olshansky