On Tue, 21 Apr 1998, Guy Cox wrote:

> >And to all this, I would ask if anyone can provide me with an example of
> >someone wanting (or should I say, really needing) to examine raw data from
> >even five years ago?  Perhaps some legal types can make a case for this
> >"need" (i.e., my client gets off if they can't produce the original data),
> >but I challange those from the ivory towers to document going back to data
> >acquired more than five years ago and really making something out of a
> >reanalysis.  My philosophy is that data is NOT like wine - you use it
> >within a relatively short time frame or you forget about it.  It does not
> >become better or more informative with age - in fact, just the opposite.  I
> >don't save string or rubber bands either...
> >Rob Palmer
> >CEB/UT
>
>
>
> Well, not being American, I don't get every bit of research I do published
> the next day .... it's quite often that data acquired 5+ years ago ends
> up in a published paper.  Then, later on, comes the possibility that one

        The other point that Guy did not mention (and I am surprised given
he is Australian) is that there is a statutory requirement in Australia
for everyone receiving funding under NHMRC (==USA NIH) and I think ARC (==
USA NSF) to maintain all "primary data" for 5 years. This ruling stems
from a scientific fraud case some years ago. "Primary data" has not been
legally tested as a definition as far as I know yet, but this institution
covers itself in the IT sense by trying to keep backups of all computer
based data and operating environments going back five years or more. This
is done by permenently archiving total site backups every 12 weeks - there
is an implied assumption of any useful data staying live for at least 12
weeks before a user deleted it. In some cases where data flow is high and
constant (Confocal, FACS, Sequencers etc), all weekly backups of the data
are archived. In a case of intentional fraud, we admit that we might be
beat. It is a matter of making a best attempt at "keeping the law".

        In a more pratical sense though, this Institute receives 5 year
block funding with a "Spanish Inquisition" type review every five years or
so.  It is amazing the amount of data retrieval that we have to pull from
these archives in the lead up to the review each five or six years.

regards, Tony Kyne
 ===============================================================================
Dr. Tony Kyne, Head, Computer and Bioinformatics Services,
               The Walter and Eliza Hall Institute of Medical Research,
               P.O. Royal Melbourne Hospital, Victoria, 3050, Australia.
Phone: International +61-3-9345-2586   FAX: International +61-3-9347-0852
            National 03-9345-2586                    National 03-9347-0852
Email:  [log in to unmask]
 ==============================================================================