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

[Stephen R. Aylward]

Hi

This is great.

If you can find a cross-platform way of getting machine id, that is good 
idea.   My concern is that I thought some PC's have their machine ID 
disabled?   That was an old security concern when intel developed one of 
the pentiums.   I could be completely mistaken...

> As for how to use the root uid, I would encoded it in a header
> file.  Then I would suggest that we extend the DICOM writers such they will
> use the ITK_UID by default but allow an application to supply an alternative
> root id.  This may be
> hard to do because different root id's may be of different lengths but a
> complete UID has a size limit.  So our suffix generation routine would have
> to be written to generate suffices of various sizes.

Ugh - great point.  However, a really long UID is quite ITK-ish.

> 
> As Dr. Clunie pointed out, nobody cares what the UIDs are as long as they
> are unique.  So it may not be worth it to extend
> the writers to take an application supplied root uid.

I think some organizations will want to continue to use the root UID 
they have already received, instead of using ITK's.   Any day now 
Siemens will use ITK, and they might want to stick with their root UID.

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
===========================================================
Dr. Stephen R. Aylward
Associate Professor of Radiology
Adjunct Associate Professor of Computer Science and Surgery
http://caddlab.rad.unc.edu
aylward at unc.edu
(919) 966-9695