[Go-essp-tech] Quality Control (A Perspective)

[Gavin M Bell]

Hello Folks,

I want to make some comments on quality control.  The primary idea I
want to get across is the inclusion of the DOI number into the catalog.
 Furthermore the person who does this inclusion is the original
publisher.  As the system is designed it is a single editor system.  The
same party who published the original catalog must be the same person to
imbue the catalog with the earned DOI value.  This value must be
included as an immutable property of the catalog and thus triggers a
version change which subsequently triggers notification.  The semantics
of having a DOI (i.e. passing all the levels of quality control) will
inform other parts of the system, such as replication (and persistence).

So, essentially...
1) publishing is equivalent to passing level 1
2) All the other machinations though the different QC levels can be
indited in the meta data.
3) The final culmination of QC is a DOI value
4) That value is sent back to the *original publisher* who imbues that
value into the catalog and republishes. (version change triggered ->
notifications sent out).
5) Replication engine knows that all catalogs (and thus files they refer
to) can be safely replicated to archival sites for posterity.
6) Done.

One addition wrinkle, which kind of makes more work for me, would be...
When users are notified, the DOI reason for notification would be
handled differently.  The user would get the notification that no files
have changed but a DOI has not been granted.  Furthermore it would be
awesome if in that email I send the bib.tex for siting that dataset.
(cause I am a stickler for detail).  This bib.tex should be a part of
the meta data and should be downloadable document (IMHO).

In step #2 listed above, it would need to be worked out to decide, who
the quality control checkers (the people actually directly charged with
doing the reviews).  It would need to be worked out, what their
communication mechanism is (*out of band* from the ESG system).  Also
the entity who provides the DOI number/value...?  the end result of all
of that then goes to step #3 -> #4.

QC... done.

Did I miss anything?

-- 
Gavin M. Bell
Lawrence Livermore National Labs
--

 "Never mistake a clear view for a short distance."
       	       -Paul Saffo

(GPG Key - http://rainbow.llnl.gov/dist/keys/gavin.asc)

 A796 CE39 9C31 68A4 52A7  1F6B 66B7 B250 21D5 6D3E