Re: Overloading arithmetic

[Per Bothner <per at bothner dot com>]

Frederick Ross wrote:
> Can kawa handle overloading *, +, etc.?  For instance, if I take the
> 2d-vector class from the docs, then
>
> (define (+ (a :: <2d-vector>) (b :: <2d-vector>)) (a:add b))
>
> Then using it (or (a:add b) as well) promptly fails with:
> Argument #1 (1.0) to '+' has wrong type (gnu.math.DFloNum) 
> (gnu.math.DFloNum)
>
> So the simple minded approach isn't right.  Is there a way to make
> this work properly?

Yes and no.  There is no general mechanism for operator overloading
at this point.  There is some support for generic/overloaded functions,
but that isn't currently used for the operators like +.

However, you could define <2d-vector> to extend gnu.math.Numeric
or Quantity.  Numeric has lots of abstract methods which you would
have to  implement.  For example the add method implements addition.

You may want to read this carefully:
http://www.gnu.org/software/kawa/internals/numbers.html
(Not fully up-to-date, I suspect, but close enough.)

I don't know how easy this would be, and whether core parts of
Kawa would need to be changed.  Worth an experiment, I guess.
--
	--Per Bothner
per@bothner.com   http://per.bothner.com/