Lombardy region and its autistic CRS

Problem: our GPS readings on the field are always shifted, when projected and superimposed to Lombardy Region official cartography.
Problem: are we doing something wrong when processing (download-project-etc) GPS data or is Lombardi Region Cartography wrong?

The answer is (after several attempts involving Quantum GIS, ArcView GIS 3.x and ArcGIS, and official orthoimagery both from Lombardy Region and the Ministry of the Environment WMS) that Lombardy Region is wrong!

Problem (again): one should not reproject all the stuff coming from bogus Lombardy Region cartography services...
Solution: I defined a new proj4 string to fully support the EPSG:3003-on-acid foul CRS used in Lombardy Region Cartographic Service.
I appropriately named it Gauss-Buascia, after the name of tha Nationwide CRS for Italy (which is Gauss-Boaga) and the dialectal word buascia, which neatly is cow dung.
Here is the string:
+proj=tmerc +lat_0=0 +lon_0=9 +k=0.9996 +x_0=1500000 +y_0=0 +ellps=intl +towgs84=-215,-61,9,0,0,0,0 +units=m +no_defs
Problem (last!): who is Lombardy Region Cartographic Service official pusher, and what stuff do they smoke there?