[Dcmlib] Group Lenght and '0000'

[Mathieu Malaterre]

Jean-Pierre ROUX wrote:
> At 14:45 -0400 30/09/05, Mathieu Malaterre wrote:
> 
>> Hello,
>>
>>     I cam across this image recently
>>
>>
>> http://nova.nlm.nih.gov/WRAMC/DICOM/WRAMC%20VC-443/225017318462C331%2 
>> 0F2EA3F46FAE7321B/99B7F85B3D1343698B6D301746477DEF/2F612AF2451EEDF03AC 
>> 73C7D19AF8033/FFBE146B651B2A44C6ADC12DB614E3ED.dcm
>> from http://nova.nlm.nih.gov/wramc/data_index.html
>>
>> And I get (does not look like 'UL'):
> 
> 
> Mathieu,
> Before doing painfull modif, could you ask David Clunie whether element 
> 0x0000 of a shadow group is still considered as the 'group length' or not.
> 
> (I'm not in the Lab and I cannot download the image right now, but I'm 
> sure the modif will turn, for instance :
> 
> V 130d|0000 [UL] [Group Length] 
> [1099164365\1276292990\3804772504\555532589]
> 
> into
> 
> B 130d|0000 [  ] [gdcm::Unknown] [gdcm::Binary data loaded;length = 16]
> 
> And that shouldn't make a big difference for standard gdcm users.
> 
> JP
> 
>>
>> V 130d|0000 [UL] [Group Length] 
>> [1099164365\1276292990\3804772504\555532589] x(4183eacd)
>> V 140d|0000 [UL] [Group Length] 
>> [835531882\1219434392\151033222\3313626383] x(31cd346a)
>> V 150d|0000 [UL] [Group Length] 
>> [4259307444\1128414896\4113509537\2144277863] x(7fffffff)
>> V 160d|0000 [UL] [Group Length] 
>> [3824301659\1114577058\3573223579\2679727550] x(7fffffff)
>> B 160d|0001 [  ] [gdcm::Unknown] [gdcm::Binary data loaded;length = 36]
>> V 170d|0000 [UL] [Group Length] 
>> [3850819895\1116855751\1080659125\3543366330] x(7fffffff)
>>
>>
>>     Since I am responsible for the '0000' => group length, I would 
>> like to remove it. Hopefully nowbody is actually using it. There 
>> should not be any problem since the instances are still available in 
>> the dictionary.
>>
>> Anybody was using this 'feature'
>>
>> Mathieu
>> Ps: I'll do that ~tomorrow when my brain is up and running.

Ok I changed my mind. Since this is a private element, and there doesn't 
seems to be any interesting data in those fields. I'll just leave gdcm 
the way it is. If anybody know what are those 'things' anyway, just 
setup a different private dict.

Mathieu
Ps: According to some people on dicom newsgroup it might just be some 
post-process operation that went wrong...