Would it be possible/usefull/fast to implement expression overloading instead of operator overloading?

What I call "expression overloading" is like:

float add_mul_func(float a, float b, float c)
written 'a * b + c' or 'c + a * b'
{
version d
	{
	return a * b + c;
	}
version x86
	{
	asm
		{
		FMULADD c,a,b
		...
		}
	}
}

This group is obsolete. You need to post on the news://news.digitalmars.com/digitalmars.D group

"Bob" <bob.efrei@laposte.net> wrote in message news:cajkj1$30nq$1@digitaldaemon.com...
> Would it be possible/usefull/fast to implement expression overloading instead of operator overloading?
>
> What I call "expression overloading" is like:
>
> float add_mul_func(float a, float b, float c)
> written 'a * b + c' or 'c + a * b'
> {
> version d
> {
> return a * b + c;
> }
> version x86
> {
> asm
> {
> FMULADD c,a,b
> ...
> }
> }
> }