On Thu, 24 Aug 2006, Peter W. Draper wrote:
> Hi Mark,
> thanks that seems to be working for me, except I don't seem to be able to
> re-send a VOTable back to TOPCAT. I can send one from SPLAT to TOPCAT to
> SPLAT, but then I get an IOException when sending to TOPCAT (I'd guess
> this might be because the backing file has been deleted?).
fair guess, but I do try to check for that. Turns out to be a
missing range check in TableSpecDataImpl. I've fixed that,
and also improved the error reporting in the plastic package so it's
easier to track that kind of thing down (plain IOException was a
bit unhelpful). You should now be able to play ping pong all day
between the two.
> > I have not yet written any documentation. I can add a suitable section
> > to sun243.tex if you like - tell me where in the document you'd like
> > it to go. In TOPCAT, the top-level Interop menu contains a
> > "Help on Interoperability" item (since there's no window dedicated to
> > to PLASTIC functionality, so no other context-sensitive help is going
> > to lead there - TOPCAT may one day control its PLASTICity from a
> > window, but not yet). If you want, I can arrange the same for SPLAT.
> Usually this would go in the "Help on window" page, but I've no objections
> to you just adding a section to SUN/243 and linking to that from the
> Interop menu, as in TOPCAT. Once you've done that I'll have a look at
> disseminating the information more widely in the docs.
Incidentally, I've just looked at SSAP more carefully and it looks
like the data access metadata is not very restrictive or specific.
The Access.Format field, which is the one that describes the
access reference (URL pointing to the spectrum) just has the form
of a MIME type; it it's a FITS file that means it has to be
plain "application/fits" which doesn't distinguish between a 1-d
(or N-d) FITS array and a FITS table. This is nastily orthogonal
to the way we are going with PLASTIC message IDs, which do specify
roughly what you can expect to find, at least syntactically,
within the FITS file. So I might have to backtrack and also
add support in SPLAT for a message that just says "this is a
FITS file". Happily, SPLAT is able to cope with that
and work out how to spectrum-ise it by looking at the content.
Mark Taylor Astronomical Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/