Remove operator++ and operator-- for floating point representations?

[jansende]

Should operator++ and operator-- be disabled for floating point representations? Also, as for operator%, how should these operations be disabled?

[mpusz]

I agree that ++ and -- do not have much sense for floating-point representations but those operations are defined for a floating-point number. So if the following works:

double length = 3;
++length; 

I think it should work as well for a safe type:

quantity<metre, double> length(3);
++length;   // should also work

Also it has benefits from a generic point of view:

Quantity foo(Quantity auto q)
{
  return ++q;
}

The user should not be forced to check what is the Rep there.

Basically, if we do not provide those operators users will have to use += 1 in a generic code which basically is the same wrong for a floating-point number and we probably do not want to ban it too ;-)

[jansende]

I think generic programming is a fine argument for why these operator should be left.
Maybe we should add this to the paper in a section titled: Supported Operations and Representations?

[mpusz]

Sure, this a good idea. Do you volunteer to write one? ;-)

[mpusz]

More on this subject in #29