(i) We should not replace monolithic packages with an unwieldy
interoperability layer.
(ii) Users don't really like C, and C++ is beyond the pale much though
programmers like it.
(iii) Portability IS important. Software should not be dependent on
UNIX/X.
(iv) Look at what they do like: IDL. I often here whinges from IRAF,
FIGARO, STARLINK users about the packages they use - rarely have I
heard that from IDL. (Perhaps this is merely because they paid $$$s
for it? :-))
(v) We should avoid reinventing wheels, but provide extra added
astronomical value to exisiting standards.
(vi) A wise friend told me "standards only work when they are
retrofitted to something that is already well-designed". The canonical
example of a horrid standard is Motif.
Here is an example of what I mean: if we were to invent a language
it would have to be IDL-like but free [ii]. It would have to run
on UNIX/Macs/WinTrash95 [iii] and be based on an already existent
well-designed language which runs on these platforms.
People who know my prejudices will already know what I have
in mind. :-) Those who don't can read my poster next week.
It would certainly be the best candidate for adding astronomy
libraries to upgrade to an IDL-like level.
see ya' all,
Karl Glazebrook
--- kgb@aaoepp.aao.gov.au [Anglo-Australian Observatory] ----> pubs: http://www.ast.cam.ac.uk/~kgb/papers.html ----> pgperl: http://www.ast.cam.ac.uk/~kgb/pgperl.html