[Go-essp-tech] Out-of-bounds data and QC level 1 checks

[stephen.pascoe at stfc.ac.uk]

Martina,

I too had thought Metafor metadata compliance was going to be QCL2.  I
know there was a desire to make the metafor metadata compulsory for any
publication but there has always been disagreement on whether this was
practical -- how do we prevent datanodes from publishing?  A final
decision on this has been illusive.

The pragmatic approach is to allow publication (QCL1) without the
metadata since to enforce it would require tighter integration of the
whole federation.  It will be far easier to enforce during replication
and QCL2.

If we are really going to enforce this as QCL1 we need to have a
concrete plan in place for how beyond the "we'll make it work"
assertion's we've had so far.

S.


---
Stephen Pascoe  +44 (0)1235 445980
British Atmospheric Data Centre
Rutherford Appleton Laboratory

-----Original Message-----
From: go-essp-tech-bounces at ucar.edu
[mailto:go-essp-tech-bounces at ucar.edu] On Behalf Of Martina Stockhause
Sent: 28 June 2010 14:28
To: Bentley, Philip
Cc: GO-ESSP; doutriaux1 at llnl.gov
Subject: Re: [Go-essp-tech] Out-of-bounds data and QC level 1 checks

Hi, Philip,

Bentley, Philip wrote:
> My current (albeit hazy!) understanding is that the QCL1 checks are 
> equivalent to those implemented by CMOR.  So any dataset output by the

> CMOR library should pass QCL1, right?
>   
Actually, it is a bit more. On the data side there are some checks done
in the ESG publisher mainly to check if a temporal aggregation to form
an Atomic Dataset is possible and controls of the DRS and CF standard
names validity.

Apart from data checks the metadata checks performed in the metafor
questionnaire are part of QCL1. As long as there exist no valid metafor
metadata the QC level 1 is not reached.

We plan to finalize the Quality Control paper soon, after the open
question of accessibility (who may access data of which level?) is
answered. Then it should be made available on the CMIP5 website.

Best wishes,
Martina

_______________________________________________
GO-ESSP-TECH mailing list
GO-ESSP-TECH at ucar.edu
http://mailman.ucar.edu/mailman/listinfo/go-essp-tech
-- 
Scanned by iCritical.