Dear ALPS developers,

Some background on the problem I have:
---------------------------------------------------
I wish to use the exact diag. application
for the FQHE effect on a thin-torus.
This can be mapped to a 1D problem on a ring which has both
total momenta (P) and total center-of-mass conservation (X).
In a ring the position (j), is actually ill defined,
therefore I need to define X as a non-Hermitian operator
$X = \sum_j e^{i 2\pi j/L} \psi^{\dagger}_x \psi_x$ instead.

My questions are:
----------------------
1. Could a non-Hermitian operator such as X, be given as a conserved
quantum number to the exact/sparse diagonalization applications ?

2. For a ring geometry and when within an <SITEOPERATOR name="X" site="j"> tag:
may I use the site's name ("j") as though it is a variable which holds an integer
denoting the position (I need this for the last definition of X).

Thanks in advance,
Zohar.