It certainly seemed to me that people were put off by the fact that Splat
etc were not pure Java and so could not be used in e.g. applets.
We could disregard this view. Or we could demonstrate that there is a very
simple installation mechanism (like Acrobat for example).
On the other hand given how much mileage things like Aladin make with just a
simple Java SLALIB implementation, I cannot help thinking that we should at
least be able to do that well - if we are careful about defining what
functionality we actually need.
...David
> -----Original Message-----
> From: Starlink development [mailto:[log in to unmask]] On Behalf Of
> David Berry
> Sent: 21 October 2003 12:01
> To: [log in to unmask]
> Subject: Re: ADASS and IVOA ideas
>
> > > We clearly DO have to read FITS WCS headers!
> >
> > So I'll happily wait for Java AST.
>
> Are we saying that the issue of a pure Java version of AST is back on
> the table? If so maybe we need to get a group of interested parties
> together to decide which bits of AST we really need now in Java, and which
> bits can be put off to later. I am aware that I have probably not inspired
> much confidence in the practicality of this project in the past, but maybe
> having a few extra heads on the job may help.
>
> What I have at the moment is a draft system which will read a FITS header
> containing celestial coords (not spectral), and extract something
> equivalent to a FrameSet. You can transform positions between the various
> Frames in the FrameSet, but it misses out most of the things which make
> AST really useful, such as:
>
> - the ability to find conversions from one Frame to another. This means
> for instance, that you can switch from galactic coords to RA/DEC, or from
> wavelength to frequency, etc.
>
> - there is no Plot class at all. In fact, will Java really be fast enough
> to do all the calculations involved in tracing out a grid line??
>
> - there is no way of cancelling out redundant steps within complex
> Mappings.
>
>
>
> David
|