Minutes for tera-meeting of Friday 20th February 2004 Absent: Jean-Christophe (Victoria, Canada); Fred (Rio, Brazil); Emmanuel (Meudon, France)

DB-client2 (Gerard)

DBclient can now read a 20,000 object catalogue in around 3min consuming a few megabytes of memory (10 times less than before). There are still memory leak residuals that Gerard plan to fix over the next week. Unfortunately, during the loading process, no new image information can be loaded into the database. Idea is to run two instances of DB-client, one used to load catalogues in the DB, the other one to handle the loading of new images. Other issues:
 refragmentation of database not yet implemented, could be significant coding effort.
 Priorities for Gerard: DB-client2, memory leak, SExserver, panier, and refragmentation (lowest priority).

Defectix (Anthony)

Demonstration given. Image reconstructed from PCA shown. Description of defect detection algorithm. Next step: implantation of neural-network learning on PCA processed images to identify defects.

Spica (Laurent, demo by Yannick)

Seeing histogram by JC. Histograms by CCD and by image total for all the CCDs. Can see see how seeing varies CCD by CCD and exposure by exposure. Request: Can we add the total integration time per filter for the deep stack (Henry's suggestion). (Laurent's task) Request: Need to have the skyprobe digital data and the exact time of observation to accurately make extinction corrections (Yannick will ask Jean-Charles/CADC) Request: the skyprobe zero point seems wrong (never set to zero). Check with Jean-Charles (Yannick) Survey mode demonstrated, as well as the successful automatic production of a stacked image. A third mode will be added for wide area surveys, for sub-images. Still to be done: add quality assessment information for the final stacks.

QualityFITS (Fred, in absentia):

Problem with display with netscape still to be fixed. Would also like to add qualityfits documentation somewhere on the web pages. Still problem with QF netscape 4.5

SCamp (Emmanuel, in absentia):

How is it that SCAMP works for Laurent, but not Henry and Mireille? To be investigated. Laurent made the calculation with all the images and the pre-survey images in the d3-field.

Data processing of wide and deep images (Mireille, Henry)

D4 i* stack shown: 2hr integration with seeing of << 0.9'' for each individual image. Mireille has processed the 4 deg2 of the wide survey in gri. *note* some problems with the wide data in photometric homogeneity, and also in the FWHM variation across the field: r2 plots show a very large scatter. Characterisation and quality assessment of both wide and deep images images is on-going. Henry will communicate with Mireille all the procedures necessary to produce final catalogues. Henry will attempt to provide low-level documentation of the entire data-reduction chain. Yannick will ask Stephane Colombi permission to use his count in cell code in Terapix

ICD document (Yannick):

Software version and maintainer information should be should be added to header. Need to send CADC a detailed evaluation of amount of data Terapix will transfer to CADC (Yannick, Laurent, Jean-Christophe)

Data Sophie Maurogordato (Gilles):

in progress. Gilles and yannick will look at this next week.

Panomef (all):

Several problems identified. For instance: software doesn't work with large ( 1Gb) stacked images, whereas previous version worked. Doesn't work on samplix.

Hardware news:

Henry contacted Mr. Landat to discover the current status of the Opteron motherboards. Unfortunately, Mr. Landat claimed that the vendor was not convinced there was a problem with our motherboards and so was not sending us the replacements. Mr. Landat requested to talk to Emmanuel, who was not present on Friday. *action* on Monday, Emmanuel will call Mr. Landat, in the presence of Henry and Yannick, to discuss (again) replacement motherboards.

Other matters:

Cosmix system management (Gilles, Henry, Yannick):

A request arrived from off-site for the cosmix root password, this request will be refused. Gilles will attempt to install all necessary packages. Request for installation of IDL/mathematica. Update: IAP refuse to allow IDL installation on cosmix because it is not a 'public' machine.

Next tera-meeting: Friday 27th Feb, 2004 at 14:30 .