I would like to have an applet to display Principal Component
Analysis. The applet should have the following specifications:

1.Read a tab delimited tex file that contains data. The format of  the
file is like this:

	Col1	Col2	Col3	Col4	etc?
Row1	123	456	789	239	etc?
Row2	234	567	890	123	etc..
Row3	345	678	901	234	etc?
etc?

2.A panel that displays a 3D scatterplot that can be rotated, printed
(via Tomcat for example), and when mouse over the spot, display the
name of the corresponding row. Obviously the scatterplot should
display the data in the datafile.

3.Three configurable list boxes that display the names of the columns
in the data file. Based on the example above:
a. The initial state of the listbox1 should be ?Col1?. The initial
state of the listbox2 should be Col2. The initial state of the
listbox3 should be Col3.
b. When there is a change in any of the list boxes (for example
changing Col3 to Col4), the applet should update the scatterplot.

4.A method to modify the color, the shape and the size of the nodes.
This can be done via javascript or through another data file with the
following characterirstics:

	Cls1	Cls2	Cls3	Cls4	etc?
Col1	A	Pos	H	2.0	etc?
Col2	B	Pos	L	1.2	etc..
Col3	A	Neg	M	3.0	etc?

The rows in this file matches the columns in the first datafile above,
and the columns in this file represent classifications for the rows.
For example, I would like to color based on Cls1, shape by Cls2 and
size by Cls3.

In the example above the spots in the graph will be like this:

	Cls1	Cls2		Cls3
Col1	Red	Circle	10 pixels
Col2	Blue	Circle	12 pixels
Col3	Red	Square	14 pixels

5. Panels that display the labels for the color, shape and size.

6. The applet must be contributed as open source and very fast. A
typical data set may have 500 rows by 20 columns but they may could be
much bigger.