[idvusers] Lambert Conic Projection query

I am not sure of the correct home for the ucar.unidata.geoloc.* package, is it 
IDV, Netcdf, or some other jar?

Anyway, in idv 2.2 there exists ucar.unidata.geoloc.projection.LambertConformal. I am copying/transcribing this class into a GWT (Java -> Javascript) class. I note this snippet in the projToLatLon method:

  double yd    = (rhop - fromY);
        double theta = Math.atan2(fromX, yd);
        double r     = Math.sqrt(fromX * fromX + yd * yd);
        if (n < 0.0) {


In the call to Math.atan2, we clearly have some x distance as first parameter, 
with y second.

Yet the api for java.lang.Math.atan2 suggests that the parameter order is y, x. Can anyone explain why the apparent mismatch? I have used this class in apps before and it seems to work fine.

Thanks

Stuart Maclean