The good news (for those for whom that sort of news is good) is that, thanks largely to the efforts of Horst Meyerdierks at The University of Edinburgh, and following an intensive burst of work over Christmas on my part, SPECX should shortly be available on Starlink platforms. Large parts of the port are complete but it is not quite clear when there will be enough of a system to be worth releasing it to Starlink. At the date of writing (4th January) neither GSD nor FITS i/o has been implemented, so there is still no way to import data other than by direct conversion of VAX files in SPECX internal format. Horst has written a standard SPECX command to do this - for the present anyway it produces .SDF files, which use Starlink NDF and HDS libraries. Remo Tilanus has been working on a GSD reader, and converting the FITS stuff should not be too difficult (although tape i/o will probably not be supported under ), so with any luck it will not be too difficult to finish off V6.4.
Unfortunately a number of the nicest features of the VAX version - such as CTRL C handling and proper support for dual-screen alpha graphics terminals - are unlikely to be available in the first version (6.4), and users may notice some "debug" type statements which are unfortunately necessary to circumvent bugs in the f77 compiler, but for most purposes SPECX V6.4 should satisfy the need. With any luck it will be released to Starlink during February, and if not, then soon after. Whether the documentation will be available by then is another matter.
Horst is aware of the need to be able to export SPECX to non-Starlink sites without them needing to take on the whole of the Starlink environment, and I am assured that this will be possible. However in the first instance it is probable that software will be available only through ROE, or later through the Starlink Librarian -- in any case, probably not from me directly.
Meanwhile, for those who do have access to a VAX still, and while you are at the telescope, the improvements in V6.3 should make life easier. In fact the release of this version has been prompted mainly by the successful commissioning of the DAS, which in turn has required major changes in the internals of SPECX to accommodate extra header information required to define the frequency scales accurately. Several new commands have been added for dealing with DAS data, while others have been modified slightly.
Most importantly, the SPECX internal datafile and map format has changed. As always, SPECX 6.3 is able to read any existing files and maps you may have, but it is no longer able to write to them. Map files (with the .MAP extension) are converted to V3 the first time they are opened by the new program -- if you intend also to use earlier versions of SPECX then make a copy of the map before you use V6.3 for the first time. This is not true of ordinary datafiles; you just can't write spectra to them any more, but it is a simple matter to open a new file and then copy the spectra across one by one or in a DO loop.
The reason these file formats have had to change is to accommodate the LO and IF frequencies in the header, along with the velocity of the source in various frames. These were not previously available in the GSD file, but were derived within SPECX from the specified source-frame centre frequency of each sub-band. This is not however a natural way to do things, and proved prone to error with increasingly complex data such as those produced by the DAS. Hence the change. Since things were changing I have taken the opportunity to tidy up some other relics of bygone computer systems -- RA and DEC are now stored internally as R*8 degrees, and the map offsets are stored as reals. This has meant changes to some of the macros supplied with SPECX, and if you use these variables in your own macros then you will need to modify them accordingly.
SPECX --- A Millimetre Wave Spectral Reduction Package