Rosetta Review - Thursday, 18 February 2016

Present in the room:

Anne Raugh (recording)
Tony Farnham
Dick Simpson
Mike A'Hearn
Ludmilla Kolokolova
Steve Joy
Emily Law
Cesare Grava 
Tilden Barnes


Joint Session with ESA


RSI
===

Reviewing RIDs:
  o The small amount of data was a decision on ESA's part to get something in
    for an initial review to avoid having to re-process large quantities of 
    data for all the likely corrections required of a first attempt.
  o The document marked as "out of scope" doesn't actually belong in this 
    data set.
  o It seems that there is a large document overhead in each of these small
    daily datasets.  Combining days would reduce a lot of redundancy there 
    and save space.  RSI will investigate what will work best for them in 
    this respect.

Reviewing Dick Simpson's issues:
  o Most of the "Conclusions" are covered by planned document updates.


CONSERT
=======

Roberto Orosei's review:
  o There is a question raised about whether it is appropriate to provide 
    any geometry values in the label, since they would be so variable over
    the observation.  This will be worked offline with the CONSERT team and
    SONC to find some reasonable solution for labels, and to provide 
    sufficient geometry to use the sounder data.

Reviewing Dick Simpson's comments:
  o Turns out the data are in a more primitive state than expected, so it is
    not possible to reproduce even a signal plot from the data as presented.
    This is an issue raised by both reviewers and will be addressed.


ROMAP
=====

Jan Soucek's review:
  o There is a large inconsistency in magnitudes for different calculations.
    The team (Philip Heinisch) reports that this is believed to be a 
    programming error in the pipeline. [AR: Doesn't that invalidate the 
    entire data set for archiving right there?!]

Reviewing Steve Joy's comments:
  o The small amount of level 5 MAG data is a reflection of the team still 
    attempting to find the correct solution and publish.
  o The preponderance of zero counts in the SPM data is an active area of 
    research.  It is possible the boom hit the comet surface and was 
    contaminated by comet material.


RPCMIP
======

Patrick Canu's review:
  o The quality factor of "-1" in the labels is likely a mistake.  The 
    ream have actually inspected the data, as indicated in the dataset.cat.
  o The team is considering changing the format of the data set. Both
    reviewers had issues trying to even read the data into analytical 
    software as presented.


RPCIES
======

Andrew Walsh's presentation:
  o Every single quality flag was set to "Good", which is suspect.  The
    ream reported the quality evaluation code isn't actually in place
    yet, but is expected for future deliveries.

[Formats for fields and particles experiments is going to be an ongoing 
 issue for PDS4.  The current experient with CDF makes the data 
 incomprehensible outside that software environment, which is a major
 issue for archiving.  The multi-file solution does not seem to be 
 satisfying for users.  Hm...]

Reviewing Steve Joy's issues:
  - No additional notes.


RPCLAP
======

Yuri Khotyaintsev's presentation:
  o He also found the data very difficult to use in its current state.
    This issue is being worked, but from comments made it's not clear to me
    that any solution would be applied to the present data as well as 
    higher level data, or data from a different phase.

Reviewing Steve Joy's comments:
  - No additional notes.


RPCICA
======

Colin Forsyth's presentation:
  o The data quality flag probably needs to be applied at a lower level
    (at the record level rather than the file level, for example).  There
    is a mention that bad data are identified by "scanning" (perhaps by a
    human?) and flagged when unpacked, but it looks like some bad data is
    not being flagged.  Ideally this should be an automated process.

Reviewing Steve Joy's comments:
  o The large number of zero values and small integers was discussed a bit.
    The team suggests this is because he was looking at areas outside the
    beam [?], but Steve noted he was actually looking for the beam and 
    couldn't find it.  Additional explanation in the documentation is 
    probably in order.
  o The level 3 product that appears to be rebinned is, in fact, a level
    4 product.  This was discussed earlier and will be addressed in keeping
    with whatever the standing requirements are for delivering level 3 and
    higher products.  The description of the associated data sets will be
    amended accordingly.
  o The comment on level 3 browse products is referencing the browse 
    products found in the corresponding raw (level 2) data.  The browse 
    plots seem to be of calibrated data rather than raw data, and so are 
    probably more useful in the level 3 data set, but this is up to the 
    team.  there may be reasons to have them in the raw data either in
    stead of or in addition to the level 3 data.


RPCMAG
======

Nicholas Achilleos' presentation:
  o Laundry list of additions/modifications to the EAICD, all of which 
    look like good actions to take.

Reviewing Steve Joy's comments:
  o The identical files in the calibration directory were also noted by
    another reviewer.  This file was originally expected to change, but it
    turns out it doesn't.  A bit of explanation in the label or the 
    calinfo.txt file would be helpful to puzzled users.

=====================

Next step is reports.  Mike A'H. and Larry will be working on this over 
the next couple weeks.