Are the coordinates of the moon wrong?

Hi,

I'm currently writing a software that needs accurate moon coordinates (in order to do Ozone measurements with a Dobson spectro-photometer).

I've written my software implementing the algorithms of Jean Meeus "Astronomical Algorithms", which are derived fromELP-2000/82, Stellarium is using ELP-2000/82B (http://www.stellarium.org/wiki/index.php/Precision http://bazaar.launchpad.net/~stellarium/stellarium/trunk/view/head:/src/core/planetsephems/elp82b.c).

So I used Stellarium as a reference software to test my program. And they do disagree. Of course, I first thought something was wrong with my code but:
*If I use Stellarium to solve the examples in the book, it doesn't find the correct answer, my program does.
*If I use the Fortran program provided by Chapront-Touze & J. Chapront (ftp://ftp.imcce.fr/pub/ephem/moon/elp82b) who published the paper ELP-2000/82B, my program find the good answer within +/-10", Stellarium is way off.

This is odd, because from Stellarium source code:
"I (Johannes Gajdosik) have just taken the Fortran code and data obtained from above and used it to create this piece of software."

Example chosen at random, position of the moon on the 16/12/2012 at 12:00 UT:
Stellarium gives : Ecliptic Geocentric: +308°58'33.0"/+5°49'31.0"
Chapront-Touze & J. Chapront Fortran program gives: Rectangular Geocentric Coordinates x=227572.1998 y=-287830.74822 z=32073.85743, which, if I'm not wrong, translates to: +308°19'53.4"/+4°59'44.3" which is also what my software finds.

I'm now very confused, I'm not an astronomer, and I can't understand where the difference comes from. Can somebody enlighten me? I might be a coordinate system conversion problem? A bug in Stellarium?

Best regards,
François.