> > One of the things on my FROG to-do list is to add VOTable support via
> > this module, and tighten the FROG<->TOPCAT integration (i.e. put up a
> > SOAP interface so that TOPCAT can call an already running FROG as well
> > as FROG spawning a TOPCAT process)...
>
> OK. Another possibility is TOPCAT starting FROG in the same JVM
> by just constructing an instance of FrogMain or whatever it is.
> This is easier to program (no SOAP) and more robust in the case that
> FROG is not already running, though it doesn't offer all the
> flexibility of doing it using SOAP.
That should be entirely doable...
It'd be quite nice to insert a time series directly into a running
instance of FROG for various reasons, not least because you might have
started TOPCAT from FROG in the first place, modified the time series and
are now trying to reinsert it into an "analysis in progress".
I guess we could do both? If it can't get anything on FROG SOAP socket it
falls back to spawning a new copy by calling FrogMain()....?
Al.
|