Terameeting for the June 9th 2005

Present: Yannick, Emmanuel, Jean-Christophe, Fred, Mireille, Henry

This week's main topic: Status of release Yannick, Fred, Mireille

 Scamp d3-t02-i: there are still double images in the stack: Henry has removed the images from the wide in the list of images, and re-done the scamp with ugriz images from the deep. Unfortunately there are still double images in the final stack.

• Emmanuel will modify scamp to produce stats in image-by-image in scamp so we can try to isolate the images which have problems.

• In the meantime, Mireille will make some tests on the stacks and astrometric solutions.

• Another solution would consist in extracting a small region around a bright source, suffering from this double images problem, swarp them to a fixed WCS frame, make png images out of it, and then check them by eye to identify the faulty one

• Fred will prepare a table for the TERAPIX web pages which contains all the offsets which were applied to each stack in each filter (both wide and deep).

 Quality assessment

• It might be useful to have some method of computing the pixel autocorrelation function for the pixels, that might be a way in which to find the doubled images. Might imagine including this in a future version of qualityFITS.

• Laurent will write this program, and will test it on the stacks which have the doubled images.

• Yannick requests that the QFITS-in pages and database entries from t02 is saved so that people can examine these QFITS-in files after t03 is started. Some discussion with Fred about the best machine to put this on: maybe dbterapix or ftpix? Laurent propose to archive both QFITS-in and QFITS-out produced by stacks. His suggestion is adopted.

• We will not preserve the notes from the final stacks.

• June 14th we will define the commentary fields to be inserted in the qualityFITS tables.

 webservices: Jean-Christophe; will make the web service demo on June 14th; he was not able to run QF on his machine, which he needs to do in order to test the webservices utility. A problem related to NAXIS procudes failures during JC's test. new: Fred and JC found the bug. The CFITSIO lib used by JC needed a patch. Fred provided it to JC.

 Other matters

• Jean-Christophe has managed to install a version of USNO-B1 which works locally, so we can interrogate the USNO cataloge without passing by the CDS (and avoiding associated problems with the accessibility of the CDS by reason). Jean-Christophe will prepare a script which allow scamp to transparently access this mirror of the CDS.

• New version of QF: this will be ready by June 15th.

• Fred suggests we make some tests with bitorrent for the distribution of data. Need to be sure that this is OK with the IAP computer service, and with CADC.

• Yannick suggests to Emmanuel that Antony works on defectix one day each week. If not, then it is better to stop Defectix now and leave Anthony kept focus on EFIGI. Emmanuel will give his response to Yannick tomorrow. new: Emmanuel is ok to leave Anthony working on Defectix, provided he keep control of Anthony's activities. In the meantine , first demo has been done by Anthony and is very encouraging.

• Some discussion about the PI mode for spica. So far, PI data are processed outside the spica environment. Laurent will make some tests with 04BF03 data and will compare efficiency and results with the "hand-made" done by Mireille on these data.

 Hardware:

• Toulouse has not yet transferred the cash necessary for cosmix3.

• A purchase decision has to yet been made on the fibre channel storage array.

• To fix the date for moving computers around in the machine room we need to fix a date with Rene Verrières for the time when we can stop all the machines, to let the electrician connect the new fuses box.