[R] qestion on cmd scale

[G Ilhamto]

Dear list,

I am starting a new project in cmdscale, and I have a question
regarding distance matrix and covariance matrix. Can anyone help me in
this? I sent this message so many times but always bounced with no
further explanation.

(1) Can I use covariance as my distance matrix?
(2) Any good reference in this matter?
(3) is the approach that I wrote below valid?

Thank you,
ilham

Below is what I did:

> y1
     x y  z
[1,] 1 4  0
[2,] 2 5 -3
[3,] 3 6 -2
> y4 <- scale (y1)
> y5<- cmdscale(dist(t(y4)))
> plot (y5.x, y5.y, type="n", xlab="", ylab="", main="cmdscale(example)")
> text(y5.x, y5.y, rownames(y5), cex=0.8)
and I got the plot of x, y and z in 2-d graph.

Q: Is this legitimate way to say re. the distance between one variable
to another ?