[Esip-preserve] Error discovered in Antarctic sea-ice record : Nature News & Comment

[Parsons, Mark via Esip-preserve]

Indeed. The irony here though is that this was a human error in a human correction. This data set is already know to be irreproducible because it involves undocumented corrections based on human judgment not mat, stats, or FORTRAN.

cheers,

-m.
On Jul 23, 2014, at 6:21 PM, Bruce Barkstrom <brbarkstrom at gmail.com<mailto:brbarkstrom at gmail.com>> wrote:

We're probably hit by public expectations that the process of producing high-quality,
homogeneous data can be done instantaneously.  After all, if we can get weather
reports with data less than 6 hours old, how come it takes a couple of decades
to get corrections to the old errors?

Another point is that the scientific review process is not well equipped to handle
long, detailed technical analysis of algorithms that come with as many as hundreds
of pages of derivations and assorted streams of technical data such as calibration
coefficients, spectral responses, and perturbations to the operational procedures.

I guess we may need to train some folks in physics and math so that they can
understand the full picture, not just the semantic web or database issues.  There
may even be some necessity to teach IT folks FORTRAN (even 1977 FORTRAN)
and such numerical analysis details as roundoff and truncation errors that were
part of routine physical science training forty years ago.

Or, to put it a bit more broadly, there's a need for a much deeper education into
more than the mathematical graphs represented in provenance, there's a need
to understand the logic that gets designed into data production scenarios - and
that's not sound bite stuff.

Bruce B.


On Wed, Jul 23, 2014 at 4:46 PM, Parsons, Mark via Esip-preserve <esip-preserve at lists.esipfed.org<mailto:esip-preserve at lists.esipfed.org>> wrote:
provenance, baby!
http://www.nature.com/news/error-discovered-in-antarctic-sea-ice-record-1.15605


_______________________________________________
Esip-preserve mailing list
Esip-preserve at lists.esipfed.org<mailto:Esip-preserve at lists.esipfed.org>
http://www.lists.esipfed.org/mailman/listinfo/esip-preserve



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.esipfed.org/pipermail/esip-preserve/attachments/20140724/2190076f/attachment.html>