Re: Group character...silly question

in article <[EMAIL PROTECTED]>,
james dolan <[EMAIL PROTECTED]> wrote:

|second there's the question: can you reconstruct the similarity class
|of the transformation from its distribution of eigenvalues?  the
|answer is "almost yes", probably, and it might be easier in "the
|present context" than in "the general case".

maybe one of the magic words i neglected to exploit here is "unitary".
maybe there's a good reason why the linear transformations you're
interested in here are unitary, and maybe that makes the problem of
determining the similarity class from the distribution of eigenvalues
rather easier.


-- 


[e-mail address [EMAIL PROTECTED]