All,
I would say that having the STARJAVA shared libs for MacOSX and Windows,
(built a fresh from the code in CVS) would be a good thing. We could then
have (give out) instructions on how to build STARJAVA for any platform from
our CVS repository.
Steve.
-----Original Message-----
From: Giaretta, DL (David) [mailto:[log in to unmask]]
Sent: 26 February 2004 10:10
To: [log in to unmask]
Subject: Priorities and NAM preparations
Dear all
It seems to me that we don't want to fall between stools and have several
half finished (non-working) things for NAM. On the other hand we don't need
perfect things - they just need to be demonstratable. Things DO have to be
ready for the release but that's after NAM and we have more time for that.
The new build system is, not particularly surprisingly, more difficult than
we had hoped - well that's a pity but not a show stopper. Given this it
looks as if we should not expect to have anything like a complete CVS load
of Starlink classic in time for NAM. Remember that the plan was for Norman
to put a full app into CVS so it could build with the new build system. Then
others would pile in their own apps. Clearly we are not at this stage yet so
there is no real conflict in priorities for people in terms of classic into
CVS vs Java.
Concerning the question of FROG vs ORAC-DR for NAM: we need FROG to be at
least demonstratable - and it looks as if it is right now, although you are
in the best position to know if there is some really essential functionality
missing (before you ask, complete documentation is not an essential for a
demonstration). It may not be ready for the FULL release, but that's after
NAM and it should be ready for that.
Next question: can distributed ORAC-DR be demonstratable for NAM? Is it a
realistic aim? From what you say, given that it would be difficult to have
both FROG and ORAC-DR ready, then it should be a doddle to get just ORAC-DR
ready! Ready for a demonstration - again, it may not be fully tested and
perfect. Is this correct?
Coming back to the new build system - Norman will continue to slave away in
the mines of libtool and autoconf, although I seem to remember that
TimeFrame is loitering in the background. I suspect that if we could have at
NAM even ONE classic app building and running under MacOSX, and of course
also under CYGWIN, then we could justifiably blow our trumpet about that,
and that may be a reasonable aim.
It seems to me that more important than any single app - since we don't have
a "killer app" - is the ability to run in a distributed way (Web Services
with ORAC-DR and other things and also JavaSpaces) and the tie in to the VO
using Web Services. This will SHOW that Starlink is not just more of the
same - although we should keep reminding people of all the world-beating
things like AST and NDF/X - but that it is an essential part of the future.
I emphasise SHOW - hence the need for things to be demonstrable but not
perfect (for NAM).
By the way, we need to put together posters and handouts for NAM therefore
people should keep their eyes open for good screenshots. If in the course of
work you think you have something really eye-catching then please take a
screen shot and send it to me with a suggested caption.
This is a fairly pragmatic approach and I hope it's clear what the
priorities should be.
Any comments/disagreements?
...David
> -----Original Message-----
> From: Alasdair Allan [mailto:[log in to unmask]]
> Sent: 25 February 2004 20:07
> To: [log in to unmask]
> Subject: Re: CVS: now try it out
>
> > Talking of remote data access, can STIL process a catalogue coming back
> > from the SkyCat servers (TST format)? Or should TOPCAT have a means of
> > accessing the SkyCat catalogue access classes that come in JSky (as a
> > means of getting catalogues into TOPCAT)?
>
> On a related note, one of the things I'm doing out here is embrace and
> extend the JNI and SOAP wrappers for Starlink Classic, and insert SOAP
> messaging into ORAC-DR, which will initally co-exist and eventually
> replace ADAM mesaging for the pipeline.
>
> As a testbed for this we're thinking about replacing CATSELECT inside
> ORAC-DR with a pure Java application with the same SOAP top level as
> the JNI wrapped classic application. Since this is the only CURSA
> applicaiton in ORAC-DR we can remove it as a dependancy.
>
> Mark, it's pretty obvious that most (all?) of the CATSELECT functionality
> is present in your tables stuff, but is it in STIL or TOPCAT?
>
> See http://www.starlink.rl.ac.uk/star/docs/sun190.htx/node50.html for
> a brief description of CATSELECT if you're not familiar with it.
>
> > To get MacOS working there has to be some kind of indication that this
> is
> > a priority. [it's a clear priority from the Mac users in the UK - there
> > will be many at NAM] Is anyone going to try Cygwin? Are there clear
> > priorities for everyone as to what the targets are for the NAM
> > pre-release? How soon before we really can start adding all the apps to
> > CVS? Is it a higher priority to put Classic into CVS than finish the
> next
> > bit of Java code?
>
> Dave, what is my priority here? Should I be doing the SOAP/ORAC-DR stuff
> or concentrating on finishing FROG? Might be hard to complete both in you
> want both ready by NAM. Should I actually be ignoring both of these and
> concentrate of trying to get PHOTOM and the rest of my classic apps into
> shape to be put into CVS?
>
> Al.
|