Teraminutes for January 4th 2007

Present: Yannick, Henry Joy, Patrick, Mathias M., Emmanuel, Mathias S. (video)

no synthesis for this first meeting of the year dedicated to the preparation of T05

T05 release

 The target date for the completion of the T05 release is set to May 1st. This means that the processing must be finished by April 15th. All CFHTLS data taken up to Jan 31, 2008 must be included in the release. Most recent data are delivered by Jean-Charles first. Older data will be reprocessed later on.
 T05 will be released even if Jean-Charles does not manage to provide his newly processed data on time. TERAPIX will decide around the 7th of February whether we should reprocess the old data or wait for the new data set.
 The 500 Elixir2 test images must have been ingested, QualityFITSed and quality controlled for January 15th. Colour-colour diagrams and other global tests must be done for the 21st.
 According to Gregory, the ingestion procedure should take an average of 6s per machine.
 Emmanuel estimates that QualityFITS run through Condor should take an average of 2 minutes per machine.

T05 Developments

 Fred should add the reading of the new PSFEx measurements (from the VOTable) in QualityFITS. The old FWHM reading based on FLUX_RADIUS should be kept to allow comparison with previous releases (provisory deadline: Jan 10).
 Emmanuel will take this opportunity to install the latest version of PSFEx for which some configuration parameters have changed name (deadline: Jan 7).
 Yannick and Emmanuel should provide relevant database parameters to Mathias M. and Gregory (deadline: Jan 4).
 Database ingestion should be ready for the 21st of January.
 The new SWarp now compensates for variable pixel size. Important: FSCALASTRO_TYPE must be set to VARIABLE for this feature to be activated (as it is not the default SWarp behaviour).
 Emmanuel shows an example of PSF homogenisation computed within PSFEx and applied to CFHTLS deep data. Work to implement this approach and test its relevance in the context of co-addition is in progress with the Dark Energy Survey team.

WIRCAM processing

 Henry will send an e-mail to CFHT complaining about the organisation of observations (the fact that only a small number of successive exposures are taken for each filter).
 Patrick expresses his concerns about the current organisation of WIRCAM data which needs to be streamlined to achieve a better efficiency. Henry and Patrick stress the necessity to automatise as much as possible the processing, and especially quality control, in order not to become swamped by the current and forthcoming WIRCAM PI data.

The next terameeting will take place on Thursday Jan 10 at 2pm (in the "conseil" room most likely)