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.