> The current parameter system with VPATH, PPATH, indirection and backups of
> values in HDS files has, however, always seems more than was needed in most
> cases!
Brian's argument reminds me of certain professors who because they and
their students didn't use Starlink software, PPARC should cut Starlink.
While you may not use a particular feature, it may be important for
someone else.
Speaking both as a programmer and as an astronomer reducing data, I did
use these facilities. Yes some made only guest appearances. Features
like VPATH and PPATH, however, were used extensively in KAPPA. The only
feature I can recall having doubts about, is the tie to HDS in parameter
files (they're not backups and record current values). The HDS
parameter files too often get corrupted.
Yes the full scope of the parameter system can be daunting if you only
programme occasionally, and a simple interface is better if that's all
we ever did. For the beginner, wWe did offer overview programming
documentation like SUN/101 and SC/12, and there is a user tutorial in
SUN/95. Keywords ACCEPT, PROMPT, and RESET came from Figaro, and were
added by *user demand*. Some users do read the documentation! Novice
astronomer-programmers don't have to use the full power of the system.
Yet IMHO it would be silly to restrict ourselves to a language that is
popular with beginners. It's move to the lowest common denominator.
I still want a full-featured parameter system.
You're never going to reach all the users. One example: after years of
pushing findme and showme, some respondents to the 1996 Questionnaire
claimed that the Starlink documentation wasn't online. We do emphasize
the good habits of removing AGI files and resetting parameters, but it's
hard to reach some people.
There may be an issue here for new postgrads and recent postdocs
unfamiliar with Starlink, now that we don't issue paper manuals+, and
we're in a limbo between Classic coding and Java. We could have a piece
in a future Bulletin, reminding folks which documents are available if
you want to programme in Classic Starlink. I'd expect we'd have an
article on programming in Java, which references to introductory, lucid,
and friendly documentation. (-:
Malcolm
+ I think we should send one copy of printed copies of new, largely
rewritten or massively expanded manuals to the major Starlink sites.
|