[Insight-developers] [Fwd: Your UID prefix from Medical Conne ctions]

[Miller, James V (Research)]

> Regarding using study ID, etc - my only thought is that such > >
constraints 
> begin to limit the potential uniqueness of the number and add > unecessary

> complexity without adding value.  However, 11^20 instead of 
> 11^36 is 
> perhaps not that big of a difference.

Stephen,

I wasn't suggesting encoding the study id into the instance uid. 
What I was saying that there are lost of circumstances where
a uid needs to be generated.  For instance, the output of 
an algorithm cannot reuse the series uid of the input dataset.
It needs to put its results into a new series with its own 
seried uid.  However, the algorithm can reuse the study uid
since you are allowed to add a series to an existing study.

In DICOM, there can be Study ID, Series ID, Image ID, but there
are always Study UID, Series UID, and Image (instance) UID's.
There are also UID's associated with contours and frames of 
reference.

My original point was that there are a number of items for which
will need to generate UID's, not just for an image.  These
other items will consume UID's.  However, 11^36 sounds like 
a lot of uids. But just one of CT scan can chew up 1000 to 2000 uids (N
slices at K recon kernels at M different slice thicknesses).

Jim