Current Status (30-jul-07)

Here is a non-exhaustive list of the tasks we are working on:

• We have merged the code specific to the WINDOWS OS in the main gildas source tree. We are now preparing monthly releases of the WINDOWS binaries.
• We are working on a bind between PYTHON and SIC with the long term goal that all the GILDAS possibilities be accessible from PYTHON.
• Several simultaneous efforts are done in CLASS90:
  • We are starting to improve the conversion from frequency to velocity as the current approximation becomes too crude with the increase of IF bandwidth of the heterodyne receivers.
  • After developing the basic tools to deal with a large quantity of spectra, we are now developing higher level interfaces to enable to quickly perform complex data reduction steps.
  • A beta version of the new FITS filler in and out the CLASS data format is available. It has been developed to be able to read and write Hershel/HIFI data.
• A major upgrade of ASTRO and CLIC happened to adapt to the new generation of receivers (at 3 and 1 mm) which were installed at PDBI during the autumn 2006. While ASTRO has been stable since September 2006, CLIC is being stabilized now. In particular, we are currently working to ensure backward compatibility with data acquired before this change of receivers. We are also preparing for the installation of the new 2 mm receivers.
• In the framework of the FP6 ALMA enhancement contract, we are working on the commissioning of the interferometric On-The-Fly observing mode at PDBI.
• At end of the summer, the PDBI on-line acquisition system will migrate from HPUX to Linux, which implies some work on GILDAS.
• As part of several of the previous efforts, several low level steps (mostly transparent for the users) are happening in the GILDAS kernel:
  • We have transferred most of the GILDAS kernel include files into FORTRAN90 modules. If this affects your programs compiled against GILDAS, please contact us by email (see the email address beside).
  • We are rewriting a significant part of the GILDAS building mechanism.
  • We will transfer all the GILDAS FORTRAN source code from fixed to free format.
• Last but not least, the holography of the ALMA antenna are officially done in CLIC. To do this, a filler in and out the ALMA data format is being maintained.

Here is a non-exhaustive list of recent major changes:

• We now fully support 64 bits processors, the Cygwin OS under Windows, MacOSX on Intel processors and the g95 fortran compiler.
• The GILDAS interprocess communication is now POSIX compliant. Although this should be transparent to the user, it is a major step towards future developments.
• The New Control System of the 30M telescope delegates the calibration of the telescope (pointing, focus, bandpass, etc...) to the data reduction packages. A library dedicated to this task and named TELCAL has been designed and is used by MIRA.
• MIRA, dedicated to the calibration of the raw format (IMBFITS) and the writing in the CLASS data format of 30M spectra obtained in 2006 and after, is now shipped with GILDAS.
• CLASS77 has been rewritten in FORTRAN 90 under the name CLASS90. CLASS90 is oriented towards the processing of a large quantity of spectra (see memo IRAM memo 2005-1). As a consequence, CLASS77 is obsolescent, i.e. not maintained anymore although still shipped as part of the GILDAS distribution.
• GRAPHIC and NIC are obsolete, i.e. they are not shipped anymore with GILDAS distributions.

The GILDAS architecture, distribution and building mechanisms have been fully overhauled in 2003/2004. This implied a few major changes:

• GRAPHIC is obsolescent; its features have been transferred to GREG and/or MAPPING; old GRAPHIC procedures should still work in GREG and/or MAPPING.
• Fortran-77 compilers (including G77) are not supported anymore.
• The old homemade fpre preprocessor have been replaced by the standard cpp preprocessor.
• GILDAS libraries and general include files have been renamed. Details may be found here.
• The gfits and cfits language have been replaced respectively by the vector\fits and class\fits commands.

In IRAM memo # 2003-2, we evaluated the GILDAS package for compliance with the ALMA off-line data processing requirements (ALMA-SW Memo 18). This evaluation follows exactly the same template that was used for the AIPS++ audit.

In ALMA memo # 464, the results i) of the audit of the AIPS++, and ii) of evaluations of GILDAS and MIRIAD packages for compliance with the ALMA Offline Data Processing Requirements are analyzed. These evaluations/audit show that about 2/3 of the Requirements are fulfilled by each data reduction package, and almost 90% are fulfilled if we use existing software from all three packages.

In ALMA memo # 465, it is thus argued that ALMA will benefit greatly by using software from the existing packages which were designed for millimeter arrays. Indeed ALMA would then benefit of the daily use and expertize for the current working millimeter arrays over the next 10 years of ALMA construction.