Minutes for the terameeting of April 30th 2007

Present: Chiara, Yannick, Henry, Emmanuel, Mathias.

Emmanuel's summary document

CFHTLS T04 processing

 Yannick explains that the "digit" problem was a round-off problem in the database handling.
 Emmanuel claims that the extreme slowdown experienced on mix6 last week might be due to XWIN measurements which almost freeze from time to time.
 Yannick mentions that in SPICA, images might be copied or linked up to three times at once; this behaviour generates a lot of confusion and should be avoided in future versions of the pipeline.
 According to Yannick, the rate of interruptions caused by "convert" and CDS connexion problems sets the quantum of uninterruptible processing time for a future pipeline: that is, a full day or a few days at most.
 Yannick insists on finishing the Wides first (before the Deeps).
 Emmanuel and Chiara discuss on the detectability of saturation for T05. An action has to be set before summer. There are two possibilities: either do it automatically from the pixel values within SExtractor (which may not be possible for stacks of frames with many differents exposure times), or by propagating the lowest saturation level metadata through SExtractor, SCAMP and SWarp.
 Dedoubling of traces in QFITSout: Chiara suggests that it might be an effect of a low number of exposures in some bands (note: this has been confirmed by Yannick since).
 The best explanation for the u-band shift in W4 is a CFHT calibration pb in u during some period of time.
 The z-band shift seems the same in all bands.
 Emmanuel suggests that Mathias adds galaxies to his photometric statistics in order to check for possible finite aperture effects.

Hardware/System
 Because of the missing support for the SATA interface, terapix.iap.fr's system may have to be heavily upgraded. In order to avoid downtime, a new machine should be bought (while the old one becomes a back-up server).
 Action: Fred should install forkmanager on mix3 prior to using it for production.
 TERAPIX' credit lines are currently stuck: Yannick, Emmanuel, Delphine and possibly Valerie will discuss about TERAPIX and EFIGI money on Wednesday afternoon.
 Chiara will check whether she can make use of Laurent's PC.

Miscellaneous
 Henry asked his collaborators to make transparencies about preliminary WIRCAM results for the CFHT user's meeting.
 Chiara will do a summary of the WIRCAM processing of P.I. programs at the CFHT user's meeting. She should mention the lack of feedback she gets from P.I. and propose to set up a "feedback form". She should also say a few words about the MEGACAM processing of P.I. programs.

WIRCAM
 Henry will do the MAG_APER version of Herve's study this week.