[MetaCRS] CS-Map: info about EUROP50-7P, a 7-parameter ED50-to-WGS84 datum shift?

[Mikael Rittri]

Hello,
sorry if this is the wrong mailing list for my question.

I found the following datum shift for ED50 to WGS84 in CS-Map\Dictionaries\datums.asc:

DT_NAME: EUROP50-7P
	DESC_NM: European 1950, mean Values (7 Parameter Transform)
# aka ED50, ED 50
# Used for Austria, Belgium, Denmark, Finland, France, W. Germany, Gibraltar, Greece
# Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland
	SOURCE: IPCC - Instituto Português de Cartografia e Cadastro
	ELLIPSOID: INTNL
     USE: 7PARAMETER
	DELTA_X: -66.863
	DELTA_Y: -134.888
 	DELTA_Z: -111.49
	ROT_X: 0.53
	ROT_Y: 0.14
	ROT_Z: -0.57
	BWSCALE: -3.4

which looked interesting, since I couldn't find it in the EPSG's database.
(The only EPSG shifts for ED50 over such a large area are 3-parameter 
 transforms taken from US DMA TR8350.2:
 - "ED50 to WGS 84 (1)" with the same area of use, and accuracy 10 m,
 - "ED50 to WGS 84 (2)", with a smaller area of use, and accuracy 6 m,
 the accuracy here is the one-sigma error.)

But I didn't find the accuracy of the 7-parameter transform in the CS-Map 
sources.  So I tried to find the Portuguese original source, without success.

However, I did find a similar transform in ArcGIS, called ED_1950_To_WGS_1984_PT7,
( http://gis.sfsu.edu/helpdesk/arcmap/projections/Map_Projection_Tables.pdf ), 
which is identical to the CS-Map version, except that 
    dX = -68.863 (not -66.863 as in CS-Map).

So, it seems that either ArcGIS or CS-Map has the wrong value for dX, but I don't 
know which one.  
   I realize that the result of using the wrong dX is probably less than the method 
accuracy, but I would still like to know which one is correct.  
   More important, I would really like to know the method accuracy (preferably 
both the one-sigma error and the maximal error, if the original source gives them). 

Best regards,
--
Mikael Rittri
Carmenta AB
SWEDEN
www.carmenta.com