In the message dated: Sun, 15 Sep 2013 01:31:43 +1000,
The pithy ruminations from Francis Liu on 
<Re: [lopsa-tech] large scale storage - medium bandwidth> were:
=> I thought HIPAA would only apply to user-identifiable data, not any old
=> random big data dataset. Best keep the important legislation-constrained
=> data as small as you can.

Ha! As someone who works in medical imaging, I can assure you that there are
some big datasets with patient-identifiable information.

=> 
=> 
=> I think many of the current uses of "big data" actually have a short half
=> life. The bigger the feed you're getting now, the less useful it is in x
=> months, because the data will be aged or expired.

It depends on your field. We do processing for some long-terms studies
(15+ years and going). In addition to the simple cumulative effect of
"N patients * S scans/year * Y years", there's the effect of changes in
MRI technology, meaning that the scan done this year may have many times
the information (and file size) of the scans done at the beginning of
the study.

In some ways, the data gets more valuable as it gets older. Longitudinal
studies allow us to test theories and algorithms in ways that large
populations at a single instance do not.

=> 
=> As the OP indicates, maintenance isn't free, and "future-proofing" has
=> never ever been free.

Yep.

Mark


-- 
Mark Bergman 
_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to