refine GGraph/GPL syntax - please help very close

[webicky]

I am trying to get a a 2-d chart with multiple categorical layers
using a scatter in GGRAPH. What
I'd like is to get a shape and color for group 1 (e.g. blue ovals).
a different shape and
color for group two (e.g Orange squares). This is easy - its x-y by
color variable. The next part. different fill
pattern based on a fourth variable variable (E.G. fill in only
certain
points) - blue with blue -- orange with orange). Every time I try
this GGRAPH fills in all the variables or changes the color of the
some other pattern.

Any suggestions? The syntax I have is very close - but I'd liked the
loc points to be solid with the color of gen rather than totally
clear.

For example :

y=Salary
x= years of service
gen = men/women
loc = U.S/overseas

GGRAPH
/GRAPHDATASET NAME="graphdataset" VARIABLES=middle_sci
mean_scaled_score_8_9 sci_band p_03 MISSING=LISTWISE REPORTMISSING=NO
/GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
SOURCE: s=userSource(id("graphdataset"))
DATA: middle_sci=col(source(s), name("middle_sci"))
DATA: mean_scaled_score_8_9=col(source(s), name
("mean_scaled_score_8_9"))
DATA: sci_band=col(source(s), name("sci_band"), unit.category())
DATA: p_03=col(source(s), name("p_03"), unit.category())
GUIDE: axis(dim(1), label("recalculated middle school sci (08b)"))
GUIDE: axis(dim(2), label("8th GradeCST Algebra I "))
ELEMENT: point(position(middle_sci*mean_scaled_score_8_9),
color.exterior(sci_band))
ELEMENT: point(position(middle_sci*mean_scaled_score_8_9), shape
(p_03*sci_band))
ELEMENT: point(position(middle_sci*mean_scaled_score_8_9),
transparency.exterior(sci_band*p_03))
END GPL.