Complexity nomenclature

[Andrei Alexandrescu via Digitalmars-d]

On 12/04/2015 01:05 AM, Tofu Ninja wrote:
> Also maybe a simpler idea would just be to annotate the the operations
> with there complexity with UDAs. That way things that really care about
> the complexity can get it, and those who don't can ignore it. It has the
> benefit of being self documenting as well.

Well look at what the cat dragged in:

http://dpaste.dzfl.pl/711aecacc450

That's quite promising. The code is very preliminary and uses strings 
for typical complexity values (e.g. "constant", "linear", and later 
"loglinear" etc). I need to see how to integrate this whole idea.

Also an unpleasant problem is overloading - when present, user code 
needs to specify which overload they're referring to.

Anyhow, this is really interesting. Thanks Tofu.


Andrei