Teraminutes for the meeting of 12th December 15: Present: Everyone, except Henry.

- Panorapix: Fred.

- Still in progress. The MEF handling part demands much more work than expected. - The WCS is not yet done, but it should not be critical, though Emmanuel raised some points that may accumulate to make the WCS working on all images. - Fred will spend the whole coming week to focus ONLY on Panorapix and provide a release for December 19, SHARP.

- QualityFITS: Fred.

- Fred implemented all requests from Laurent for running SPICA. The new version is currently is test phase by Laurent.

- In order to control duplication of images in the database, we agreed that a full file checksum will be automatically done on all input Megacam images by QualityFITS. The checksum parameter will then be stored in the database. Fred can do this for next week since it does not demand a huge amount of work.

- The FILENAME problem : Jean-Christophe, Gerard

- The implementation of a checksum control in QualityFITS will provide a control parameter to control duplication of images.

- In order to decouple Terapix from unexpected parameter modifications at CFHT, it is decided that we will use the file name as the FILENAME keyword. Gerard and Jean- Christophe checked this can be easily done in DBclient, with minor changes. It will force the data base to use the file name, regardless the value of the FILENAME keyword.

- SCAMP: Emmanuel

- Some progress done, still under development. Emmanuel will focus on SCAMP as a top priority.

- ICD: Jean-Christophe

- Jean-Christophe delivered to Laurent new XML files that describe the processing history.

- Since SPICA keeps track of all details that permit to reproduce exactly the processing done in the past, with the software version used at that time, all configuration files and software releases must be stored somewhere. The CVS seems the most efficient way to keep all versions indenfitely SPICA should then check automatically that the version is may need at any time is indeed in the CVS repository. If not, a warning email should be sent to teralist.

- Jean-Christophe modified SNOOPPIX to avoid the output swarped and weight map images to be loaded on the ouput disks to CADC.

- VO-tables : Gerard

- Currently under developement. Gerad will present a prototype on December 19.

- Mailserver: Gilles

- Apparently the teralist and swarp mailing list are no longer working. Gilles will check this point and will add all mails missing in these lists.

- Web pages: Gerard

- Gerard improved the protection of Terapix intranet pages and homogenised the access to back office, public and intranet page access.

- More generally, it is requested that all Terapix people get used to the CVS and to use it systematically.

- Zero Point AF19 images: Mireille

- Mireille has checked the ZP and color equations of CFHT standard stars. Several points are still unclear in the CFHT procedure.

- Yannick will ask Jean-Charles for a detailled description of the photometric calibration process used by CFHT. In particular we do not understand why BD+26-206 is saturated and not observed with telescope defocused.

- Yannick will provide color equations to trans from the SDSS to CFHT filters. This has been done by Schulteiss and Annie Robin.

- DBclient: Gerard. Jean-Christophe

- Still there. Emmanuel recommends to use mem usage on DBclient to check the memory allocation.

- Data Estelle Moreuax: Gilles, Mireille, Yannick

- The astrometric solution provided by Gilles shows unexplained variations. Gilles will use "getrms" to get more detailled and quantitative quality assessements. If "getrms" cannot work on XP1000, Gilles will use Astrometrix on Mandrake platforms. Mireille recommends to specify a value of the "map" qualifier in the getrms command.

- Yannick will send to Mireille the image names that have bad astrometry. Mireille will run Astrometrix on her own on these images.

- Harware: Emmanuel

- The new BIOS seem to solve all problems the Opteron has had before. Mandrake can then be installed. The 3-ware boards work, but only disk partitions up to one 1To can be created.

- First tests done with SWarp show the opterons are 1.5 faster than node5 (2x2Ghz Athlon MP)

- It is decided that the Opteron will be installed in the Terapix production cluster on December 19.

- Next Tera-meeting: December 19.